diff --git a/README.md b/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..479e3fc83d2cd89bd1801f454e6c1b6e7628d7f7
--- /dev/null
+++ b/README.md
@@ -0,0 +1,88 @@
+---
+tags:
+- espnet
+- audio
+- automatic-speech-recognition
+- speech-translation
+language: multilingual
+datasets:
+- owsm_v3.1_lowrestriction
+license: cc-by-4.0
+---
+
+## OWSM: Open Whisper-style Speech Model
+
+[OWSM](https://arxiv.org/abs/2309.13876) is an Open Whisper-style Speech Model from [CMU WAVLab](https://www.wavlab.org/). It reproduces Whisper-style training using publicly available data and an open-source toolkit [ESPnet](https://github.com/espnet/espnet).
+
+Our demo is available [here](https://huggingface.co/spaces/pyf98/OWSM_v3_demo). The [project page](https://www.wavlab.org/activities/2024/owsm/) contains various resources.
+
+[OWSM v3.1](https://arxiv.org/abs/2401.16658) is an improved version of OWSM v3. It significantly outperforms OWSM v3 in almost all evaluation benchmarks.
+We do not include any new training data. Instead, we utilize a state-of-the-art speech encoder, [E-Branchformer](https://arxiv.org/abs/2210.00077).
+
+This is a small size model with 367M parameters and is trained on 70k hours of public speech data with lower restrictions (compared to the full OWSM data). Please check our [project page](https://www.wavlab.org/activities/2024/owsm/) for more information.
+Specifically, it supports the following speech-to-text tasks:
+- Speech recognition
+- Utterance-level alignment
+- Long-form transcription
+- Language identification
+
+
+### Citing OWSM, Branchformers and ESPnet
+
+```BibTex
+@misc{peng2024owsm,
+      title={OWSM v3.1: Better and Faster Open Whisper-Style Speech Models based on E-Branchformer}, 
+      author={Yifan Peng and Jinchuan Tian and William Chen and Siddhant Arora and Brian Yan and Yui Sudo and Muhammad Shakeel and Kwanghee Choi and Jiatong Shi and Xuankai Chang and Jee-weon Jung and Shinji Watanabe},
+      year={2024},
+      eprint={2401.16658},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL}
+}
+@INPROCEEDINGS{owsm-asru23,
+  author={Peng, Yifan and Tian, Jinchuan and Yan, Brian and Berrebbi, Dan and Chang, Xuankai and Li, Xinjian and Shi, Jiatong and Arora, Siddhant and Chen, William and Sharma, Roshan and Zhang, Wangyou and Sudo, Yui and Shakeel, Muhammad and Jung, Jee-Weon and Maiti, Soumi and Watanabe, Shinji},
+  booktitle={2023 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)}, 
+  title={Reproducing Whisper-Style Training Using An Open-Source Toolkit And Publicly Available Data}, 
+  year={2023},
+  doi={10.1109/ASRU57964.2023.10389676}
+}
+@inproceedings{peng23b_interspeech,
+  author={Yifan Peng and Kwangyoun Kim and Felix Wu and Brian Yan and Siddhant Arora and William Chen and Jiyang Tang and Suwon Shon and Prashant Sridhar and Shinji Watanabe},
+  title={{A Comparative Study on E-Branchformer vs Conformer in Speech Recognition, Translation, and Understanding Tasks}},
+  year=2023,
+  booktitle={Proc. INTERSPEECH 2023},
+  pages={2208--2212},
+  doi={10.21437/Interspeech.2023-1194}
+}
+@inproceedings{kim2023branchformer,
+  title={E-branchformer: Branchformer with enhanced merging for speech recognition},
+  author={Kim, Kwangyoun and Wu, Felix and Peng, Yifan and Pan, Jing and Sridhar, Prashant and Han, Kyu J and Watanabe, Shinji},
+  booktitle={2022 IEEE Spoken Language Technology Workshop (SLT)},
+  pages={84--91},
+  year={2023},
+  organization={IEEE}
+}
+@InProceedings{pmlr-v162-peng22a,
+  title = 	 {Branchformer: Parallel {MLP}-Attention Architectures to Capture Local and Global Context for Speech Recognition and Understanding},
+  author =       {Peng, Yifan and Dalmia, Siddharth and Lane, Ian and Watanabe, Shinji},
+  booktitle = 	 {Proceedings of the 39th International Conference on Machine Learning},
+  pages = 	 {17627--17643},
+  year = 	 {2022},
+  editor = 	 {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan},
+  volume = 	 {162},
+  series = 	 {Proceedings of Machine Learning Research},
+  month = 	 {17--23 Jul},
+  publisher =    {PMLR},
+  pdf = 	 {https://proceedings.mlr.press/v162/peng22a/peng22a.pdf},
+  url = 	 {https://proceedings.mlr.press/v162/peng22a.html},
+  abstract = 	 {Conformer has proven to be effective in many speech processing tasks. It combines the benefits of extracting local dependencies using convolutions and global dependencies using self-attention. Inspired by this, we propose a more flexible, interpretable and customizable encoder alternative, Branchformer, with parallel branches for modeling various ranged dependencies in end-to-end speech processing. In each encoder layer, one branch employs self-attention or its variant to capture long-range dependencies, while the other branch utilizes an MLP module with convolutional gating (cgMLP) to extract local relationships. We conduct experiments on several speech recognition and spoken language understanding benchmarks. Results show that our model outperforms both Transformer and cgMLP. It also matches with or outperforms state-of-the-art results achieved by Conformer. Furthermore, we show various strategies to reduce computation thanks to the two-branch architecture, including the ability to have variable inference complexity in a single trained model. The weights learned for merging branches indicate how local and global dependencies are utilized in different layers, which benefits model designing.}
+}
+@inproceedings{watanabe2018espnet,
+  author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
+  title={{ESPnet}: End-to-End Speech Processing Toolkit},
+  year={2018},
+  booktitle={Proceedings of Interspeech},
+  pages={2207--2211},
+  doi={10.21437/Interspeech.2018-1456},
+  url={http://dx.doi.org/10.21437/Interspeech.2018-1456}
+}
+```
diff --git a/data/token_list/bpe_unigram50000/bpe.model b/data/token_list/bpe_unigram50000/bpe.model
new file mode 100644
index 0000000000000000000000000000000000000000..57f6cad77341e1c1f7ea472538544ca6d4c5cbe1
--- /dev/null
+++ b/data/token_list/bpe_unigram50000/bpe.model
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af3103a5e6dbaea47c4ca88e2ae92e536ff335c40876de2b2963d2af52746453
+size 1053520
diff --git a/exp/s2t_stats_raw_bpe50000/train/feats_stats.npz b/exp/s2t_stats_raw_bpe50000/train/feats_stats.npz
new file mode 100644
index 0000000000000000000000000000000000000000..b4e32c14f9cac87d66cc9b9bce606eb6557c4fcd
--- /dev/null
+++ b/exp/s2t_stats_raw_bpe50000/train/feats_stats.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f1171609665f1de3c8b99eb2f23392e69a32c6d7027361b9645e0ffaa5fdeae6
+size 1402
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/RESULTS.md b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/RESULTS.md
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/config.yaml b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..8e9f2f02677752679079b7ff0b9308f7237702b5
--- /dev/null
+++ b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/config.yaml
@@ -0,0 +1,50271 @@
+config: conf/train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn.yaml
+print_config: false
+log_level: INFO
+drop_last_iter: false
+dry_run: false
+iterator_type: sequence
+valid_iterator_type: null
+output_dir: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000
+ngpu: 1
+seed: 42
+num_workers: 6
+num_att_plot: 0
+dist_backend: nccl
+dist_init_method: file:///weka/home-pengyf/espnet-owsm-train/egs2/owsm_v3.1_licensefree_nosa/s2t1/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/.dist_init_d681a8ba-15d8-4a99-88f9-39beba562b1a
+dist_world_size: 16
+dist_rank: 0
+local_rank: 0
+dist_master_addr: null
+dist_master_port: null
+dist_launcher: slurm
+multiprocessing_distributed: true
+unused_parameters: false
+sharded_ddp: false
+cudnn_enabled: true
+cudnn_benchmark: false
+cudnn_deterministic: true
+collect_stats: false
+write_collected_feats: false
+max_epoch: 45
+patience: null
+val_scheduler_criterion:
+- valid
+- loss
+early_stopping_criterion:
+- valid
+- loss
+- min
+best_model_criterion:
+-   - valid
+    - acc
+    - max
+-   - valid
+    - total_count
+    - max
+keep_nbest_models: 5
+nbest_averaging_interval: 0
+grad_clip: 5.0
+grad_clip_type: 2.0
+grad_noise: false
+accum_grad: 1
+no_forward_run: false
+resume: true
+train_dtype: float32
+use_amp: true
+log_interval: 500
+use_matplotlib: true
+use_tensorboard: true
+create_graph_in_tensorboard: false
+use_wandb: true
+wandb_project: null
+wandb_id: null
+wandb_entity: null
+wandb_name: null
+wandb_model_log_interval: -1
+detect_anomaly: false
+use_lora: false
+save_lora_only: true
+lora_conf: {}
+pretrain_path: null
+init_param: []
+ignore_init_mismatch: false
+freeze_param: []
+num_iters_per_epoch: 15000
+batch_size: 256
+valid_batch_size: null
+batch_bins: 1000000
+valid_batch_bins: null
+train_shape_file:
+- exp/s2t_stats_raw_bpe50000/splits3/speech_shape
+- exp/s2t_stats_raw_bpe50000/splits3/text_prev_shape.bpe
+- exp/s2t_stats_raw_bpe50000/splits3/text_ctc_shape.bpe
+- exp/s2t_stats_raw_bpe50000/splits3/text_shape.bpe
+valid_shape_file:
+- exp/s2t_stats_raw_bpe50000/valid/speech_shape
+- exp/s2t_stats_raw_bpe50000/valid/text_prev_shape.bpe
+- exp/s2t_stats_raw_bpe50000/valid/text_ctc_shape.bpe
+- exp/s2t_stats_raw_bpe50000/valid/text_shape.bpe
+batch_type: unsorted
+valid_batch_type: null
+fold_length:
+- 80000
+- 150
+- 150
+- 150
+sort_in_batch: descending
+shuffle_within_batch: false
+sort_batch: descending
+multiple_iterator: true
+chunk_length: 500
+chunk_shift_ratio: 0.5
+num_cache_chunks: 1024
+chunk_excluded_key_prefixes: []
+chunk_default_fs: null
+train_data_path_and_name_and_type:
+-   - exp/s2t_stats_raw_bpe50000/splits3/wav.scp
+    - speech
+    - kaldi_ark
+-   - exp/s2t_stats_raw_bpe50000/splits3/text.prev
+    - text_prev
+    - text
+-   - exp/s2t_stats_raw_bpe50000/splits3/text.ctc
+    - text_ctc
+    - text
+-   - exp/s2t_stats_raw_bpe50000/splits3/text
+    - text
+    - text
+valid_data_path_and_name_and_type:
+-   - dump/raw/dev/wav.scp
+    - speech
+    - kaldi_ark
+-   - dump/raw/dev/text.prev
+    - text_prev
+    - text
+-   - dump/raw/dev/text.ctc
+    - text_ctc
+    - text
+-   - dump/raw/dev/text
+    - text
+    - text
+allow_variable_data_keys: false
+max_cache_size: 0.0
+max_cache_fd: 32
+allow_multi_rates: false
+valid_max_cache_size: null
+exclude_weight_decay: false
+exclude_weight_decay_conf: {}
+optim: adamw
+optim_conf:
+    lr: 0.0005
+    betas:
+    - 0.9
+    - 0.98
+    eps: 1.0e-06
+    weight_decay: 0.0
+scheduler: piecewiselinearwarmuplr
+scheduler_conf:
+    warmup_steps_list:
+    - 0
+    - 30000
+    - 60000
+    warmup_lr_list:
+    - 0.0
+    - 5.0e-05
+    - 0.0005
+token_list:
+- <blank>
+- <unk>
+- <na>
+- <nospeech>
+- <abk>
+- <afr>
+- <amh>
+- <ara>
+- <asm>
+- <ast>
+- <aze>
+- <bak>
+- <bas>
+- <bel>
+- <ben>
+- <bos>
+- <bre>
+- <bul>
+- <cat>
+- <ceb>
+- <ces>
+- <chv>
+- <ckb>
+- <cmn>
+- <cnh>
+- <cym>
+- <dan>
+- <deu>
+- <dgd>
+- <div>
+- <ell>
+- <eng>
+- <epo>
+- <est>
+- <eus>
+- <fas>
+- <fil>
+- <fin>
+- <fra>
+- <frr>
+- <ful>
+- <gle>
+- <glg>
+- <grn>
+- <guj>
+- <hat>
+- <hau>
+- <heb>
+- <hin>
+- <hrv>
+- <hsb>
+- <hun>
+- <hye>
+- <ibo>
+- <ina>
+- <ind>
+- <isl>
+- <ita>
+- <jav>
+- <jpn>
+- <kab>
+- <kam>
+- <kan>
+- <kat>
+- <kaz>
+- <kea>
+- <khm>
+- <kin>
+- <kir>
+- <kmr>
+- <kor>
+- <lao>
+- <lav>
+- <lga>
+- <lin>
+- <lit>
+- <ltz>
+- <lug>
+- <luo>
+- <mal>
+- <mar>
+- <mas>
+- <mdf>
+- <mhr>
+- <mkd>
+- <mlt>
+- <mon>
+- <mri>
+- <mrj>
+- <mya>
+- <myv>
+- <nan>
+- <nep>
+- <nld>
+- <nno>
+- <nob>
+- <npi>
+- <nso>
+- <nya>
+- <oci>
+- <ori>
+- <orm>
+- <ory>
+- <pan>
+- <pol>
+- <por>
+- <pus>
+- <quy>
+- <roh>
+- <ron>
+- <rus>
+- <sah>
+- <sat>
+- <sin>
+- <skr>
+- <slk>
+- <slv>
+- <sna>
+- <snd>
+- <som>
+- <sot>
+- <spa>
+- <srd>
+- <srp>
+- <sun>
+- <swa>
+- <swe>
+- <swh>
+- <tam>
+- <tat>
+- <tel>
+- <tgk>
+- <tgl>
+- <tha>
+- <tig>
+- <tir>
+- <tok>
+- <tpi>
+- <tsn>
+- <tuk>
+- <tur>
+- <twi>
+- <uig>
+- <ukr>
+- <umb>
+- <urd>
+- <uzb>
+- <vie>
+- <vot>
+- <wol>
+- <xho>
+- <yor>
+- <yue>
+- <zho>
+- <zul>
+- <asr>
+- <st_ara>
+- <st_cat>
+- <st_ces>
+- <st_cym>
+- <st_deu>
+- <st_eng>
+- <st_est>
+- <st_fas>
+- <st_fra>
+- <st_ind>
+- <st_ita>
+- <st_jpn>
+- <st_lav>
+- <st_mon>
+- <st_nld>
+- <st_por>
+- <st_ron>
+- <st_rus>
+- <st_slv>
+- <st_spa>
+- <st_swe>
+- <st_tam>
+- <st_tur>
+- <st_vie>
+- <st_zho>
+- <notimestamps>
+- <0.00>
+- <0.02>
+- <0.04>
+- <0.06>
+- <0.08>
+- <0.10>
+- <0.12>
+- <0.14>
+- <0.16>
+- <0.18>
+- <0.20>
+- <0.22>
+- <0.24>
+- <0.26>
+- <0.28>
+- <0.30>
+- <0.32>
+- <0.34>
+- <0.36>
+- <0.38>
+- <0.40>
+- <0.42>
+- <0.44>
+- <0.46>
+- <0.48>
+- <0.50>
+- <0.52>
+- <0.54>
+- <0.56>
+- <0.58>
+- <0.60>
+- <0.62>
+- <0.64>
+- <0.66>
+- <0.68>
+- <0.70>
+- <0.72>
+- <0.74>
+- <0.76>
+- <0.78>
+- <0.80>
+- <0.82>
+- <0.84>
+- <0.86>
+- <0.88>
+- <0.90>
+- <0.92>
+- <0.94>
+- <0.96>
+- <0.98>
+- <1.00>
+- <1.02>
+- <1.04>
+- <1.06>
+- <1.08>
+- <1.10>
+- <1.12>
+- <1.14>
+- <1.16>
+- <1.18>
+- <1.20>
+- <1.22>
+- <1.24>
+- <1.26>
+- <1.28>
+- <1.30>
+- <1.32>
+- <1.34>
+- <1.36>
+- <1.38>
+- <1.40>
+- <1.42>
+- <1.44>
+- <1.46>
+- <1.48>
+- <1.50>
+- <1.52>
+- <1.54>
+- <1.56>
+- <1.58>
+- <1.60>
+- <1.62>
+- <1.64>
+- <1.66>
+- <1.68>
+- <1.70>
+- <1.72>
+- <1.74>
+- <1.76>
+- <1.78>
+- <1.80>
+- <1.82>
+- <1.84>
+- <1.86>
+- <1.88>
+- <1.90>
+- <1.92>
+- <1.94>
+- <1.96>
+- <1.98>
+- <2.00>
+- <2.02>
+- <2.04>
+- <2.06>
+- <2.08>
+- <2.10>
+- <2.12>
+- <2.14>
+- <2.16>
+- <2.18>
+- <2.20>
+- <2.22>
+- <2.24>
+- <2.26>
+- <2.28>
+- <2.30>
+- <2.32>
+- <2.34>
+- <2.36>
+- <2.38>
+- <2.40>
+- <2.42>
+- <2.44>
+- <2.46>
+- <2.48>
+- <2.50>
+- <2.52>
+- <2.54>
+- <2.56>
+- <2.58>
+- <2.60>
+- <2.62>
+- <2.64>
+- <2.66>
+- <2.68>
+- <2.70>
+- <2.72>
+- <2.74>
+- <2.76>
+- <2.78>
+- <2.80>
+- <2.82>
+- <2.84>
+- <2.86>
+- <2.88>
+- <2.90>
+- <2.92>
+- <2.94>
+- <2.96>
+- <2.98>
+- <3.00>
+- <3.02>
+- <3.04>
+- <3.06>
+- <3.08>
+- <3.10>
+- <3.12>
+- <3.14>
+- <3.16>
+- <3.18>
+- <3.20>
+- <3.22>
+- <3.24>
+- <3.26>
+- <3.28>
+- <3.30>
+- <3.32>
+- <3.34>
+- <3.36>
+- <3.38>
+- <3.40>
+- <3.42>
+- <3.44>
+- <3.46>
+- <3.48>
+- <3.50>
+- <3.52>
+- <3.54>
+- <3.56>
+- <3.58>
+- <3.60>
+- <3.62>
+- <3.64>
+- <3.66>
+- <3.68>
+- <3.70>
+- <3.72>
+- <3.74>
+- <3.76>
+- <3.78>
+- <3.80>
+- <3.82>
+- <3.84>
+- <3.86>
+- <3.88>
+- <3.90>
+- <3.92>
+- <3.94>
+- <3.96>
+- <3.98>
+- <4.00>
+- <4.02>
+- <4.04>
+- <4.06>
+- <4.08>
+- <4.10>
+- <4.12>
+- <4.14>
+- <4.16>
+- <4.18>
+- <4.20>
+- <4.22>
+- <4.24>
+- <4.26>
+- <4.28>
+- <4.30>
+- <4.32>
+- <4.34>
+- <4.36>
+- <4.38>
+- <4.40>
+- <4.42>
+- <4.44>
+- <4.46>
+- <4.48>
+- <4.50>
+- <4.52>
+- <4.54>
+- <4.56>
+- <4.58>
+- <4.60>
+- <4.62>
+- <4.64>
+- <4.66>
+- <4.68>
+- <4.70>
+- <4.72>
+- <4.74>
+- <4.76>
+- <4.78>
+- <4.80>
+- <4.82>
+- <4.84>
+- <4.86>
+- <4.88>
+- <4.90>
+- <4.92>
+- <4.94>
+- <4.96>
+- <4.98>
+- <5.00>
+- <5.02>
+- <5.04>
+- <5.06>
+- <5.08>
+- <5.10>
+- <5.12>
+- <5.14>
+- <5.16>
+- <5.18>
+- <5.20>
+- <5.22>
+- <5.24>
+- <5.26>
+- <5.28>
+- <5.30>
+- <5.32>
+- <5.34>
+- <5.36>
+- <5.38>
+- <5.40>
+- <5.42>
+- <5.44>
+- <5.46>
+- <5.48>
+- <5.50>
+- <5.52>
+- <5.54>
+- <5.56>
+- <5.58>
+- <5.60>
+- <5.62>
+- <5.64>
+- <5.66>
+- <5.68>
+- <5.70>
+- <5.72>
+- <5.74>
+- <5.76>
+- <5.78>
+- <5.80>
+- <5.82>
+- <5.84>
+- <5.86>
+- <5.88>
+- <5.90>
+- <5.92>
+- <5.94>
+- <5.96>
+- <5.98>
+- <6.00>
+- <6.02>
+- <6.04>
+- <6.06>
+- <6.08>
+- <6.10>
+- <6.12>
+- <6.14>
+- <6.16>
+- <6.18>
+- <6.20>
+- <6.22>
+- <6.24>
+- <6.26>
+- <6.28>
+- <6.30>
+- <6.32>
+- <6.34>
+- <6.36>
+- <6.38>
+- <6.40>
+- <6.42>
+- <6.44>
+- <6.46>
+- <6.48>
+- <6.50>
+- <6.52>
+- <6.54>
+- <6.56>
+- <6.58>
+- <6.60>
+- <6.62>
+- <6.64>
+- <6.66>
+- <6.68>
+- <6.70>
+- <6.72>
+- <6.74>
+- <6.76>
+- <6.78>
+- <6.80>
+- <6.82>
+- <6.84>
+- <6.86>
+- <6.88>
+- <6.90>
+- <6.92>
+- <6.94>
+- <6.96>
+- <6.98>
+- <7.00>
+- <7.02>
+- <7.04>
+- <7.06>
+- <7.08>
+- <7.10>
+- <7.12>
+- <7.14>
+- <7.16>
+- <7.18>
+- <7.20>
+- <7.22>
+- <7.24>
+- <7.26>
+- <7.28>
+- <7.30>
+- <7.32>
+- <7.34>
+- <7.36>
+- <7.38>
+- <7.40>
+- <7.42>
+- <7.44>
+- <7.46>
+- <7.48>
+- <7.50>
+- <7.52>
+- <7.54>
+- <7.56>
+- <7.58>
+- <7.60>
+- <7.62>
+- <7.64>
+- <7.66>
+- <7.68>
+- <7.70>
+- <7.72>
+- <7.74>
+- <7.76>
+- <7.78>
+- <7.80>
+- <7.82>
+- <7.84>
+- <7.86>
+- <7.88>
+- <7.90>
+- <7.92>
+- <7.94>
+- <7.96>
+- <7.98>
+- <8.00>
+- <8.02>
+- <8.04>
+- <8.06>
+- <8.08>
+- <8.10>
+- <8.12>
+- <8.14>
+- <8.16>
+- <8.18>
+- <8.20>
+- <8.22>
+- <8.24>
+- <8.26>
+- <8.28>
+- <8.30>
+- <8.32>
+- <8.34>
+- <8.36>
+- <8.38>
+- <8.40>
+- <8.42>
+- <8.44>
+- <8.46>
+- <8.48>
+- <8.50>
+- <8.52>
+- <8.54>
+- <8.56>
+- <8.58>
+- <8.60>
+- <8.62>
+- <8.64>
+- <8.66>
+- <8.68>
+- <8.70>
+- <8.72>
+- <8.74>
+- <8.76>
+- <8.78>
+- <8.80>
+- <8.82>
+- <8.84>
+- <8.86>
+- <8.88>
+- <8.90>
+- <8.92>
+- <8.94>
+- <8.96>
+- <8.98>
+- <9.00>
+- <9.02>
+- <9.04>
+- <9.06>
+- <9.08>
+- <9.10>
+- <9.12>
+- <9.14>
+- <9.16>
+- <9.18>
+- <9.20>
+- <9.22>
+- <9.24>
+- <9.26>
+- <9.28>
+- <9.30>
+- <9.32>
+- <9.34>
+- <9.36>
+- <9.38>
+- <9.40>
+- <9.42>
+- <9.44>
+- <9.46>
+- <9.48>
+- <9.50>
+- <9.52>
+- <9.54>
+- <9.56>
+- <9.58>
+- <9.60>
+- <9.62>
+- <9.64>
+- <9.66>
+- <9.68>
+- <9.70>
+- <9.72>
+- <9.74>
+- <9.76>
+- <9.78>
+- <9.80>
+- <9.82>
+- <9.84>
+- <9.86>
+- <9.88>
+- <9.90>
+- <9.92>
+- <9.94>
+- <9.96>
+- <9.98>
+- <10.00>
+- <10.02>
+- <10.04>
+- <10.06>
+- <10.08>
+- <10.10>
+- <10.12>
+- <10.14>
+- <10.16>
+- <10.18>
+- <10.20>
+- <10.22>
+- <10.24>
+- <10.26>
+- <10.28>
+- <10.30>
+- <10.32>
+- <10.34>
+- <10.36>
+- <10.38>
+- <10.40>
+- <10.42>
+- <10.44>
+- <10.46>
+- <10.48>
+- <10.50>
+- <10.52>
+- <10.54>
+- <10.56>
+- <10.58>
+- <10.60>
+- <10.62>
+- <10.64>
+- <10.66>
+- <10.68>
+- <10.70>
+- <10.72>
+- <10.74>
+- <10.76>
+- <10.78>
+- <10.80>
+- <10.82>
+- <10.84>
+- <10.86>
+- <10.88>
+- <10.90>
+- <10.92>
+- <10.94>
+- <10.96>
+- <10.98>
+- <11.00>
+- <11.02>
+- <11.04>
+- <11.06>
+- <11.08>
+- <11.10>
+- <11.12>
+- <11.14>
+- <11.16>
+- <11.18>
+- <11.20>
+- <11.22>
+- <11.24>
+- <11.26>
+- <11.28>
+- <11.30>
+- <11.32>
+- <11.34>
+- <11.36>
+- <11.38>
+- <11.40>
+- <11.42>
+- <11.44>
+- <11.46>
+- <11.48>
+- <11.50>
+- <11.52>
+- <11.54>
+- <11.56>
+- <11.58>
+- <11.60>
+- <11.62>
+- <11.64>
+- <11.66>
+- <11.68>
+- <11.70>
+- <11.72>
+- <11.74>
+- <11.76>
+- <11.78>
+- <11.80>
+- <11.82>
+- <11.84>
+- <11.86>
+- <11.88>
+- <11.90>
+- <11.92>
+- <11.94>
+- <11.96>
+- <11.98>
+- <12.00>
+- <12.02>
+- <12.04>
+- <12.06>
+- <12.08>
+- <12.10>
+- <12.12>
+- <12.14>
+- <12.16>
+- <12.18>
+- <12.20>
+- <12.22>
+- <12.24>
+- <12.26>
+- <12.28>
+- <12.30>
+- <12.32>
+- <12.34>
+- <12.36>
+- <12.38>
+- <12.40>
+- <12.42>
+- <12.44>
+- <12.46>
+- <12.48>
+- <12.50>
+- <12.52>
+- <12.54>
+- <12.56>
+- <12.58>
+- <12.60>
+- <12.62>
+- <12.64>
+- <12.66>
+- <12.68>
+- <12.70>
+- <12.72>
+- <12.74>
+- <12.76>
+- <12.78>
+- <12.80>
+- <12.82>
+- <12.84>
+- <12.86>
+- <12.88>
+- <12.90>
+- <12.92>
+- <12.94>
+- <12.96>
+- <12.98>
+- <13.00>
+- <13.02>
+- <13.04>
+- <13.06>
+- <13.08>
+- <13.10>
+- <13.12>
+- <13.14>
+- <13.16>
+- <13.18>
+- <13.20>
+- <13.22>
+- <13.24>
+- <13.26>
+- <13.28>
+- <13.30>
+- <13.32>
+- <13.34>
+- <13.36>
+- <13.38>
+- <13.40>
+- <13.42>
+- <13.44>
+- <13.46>
+- <13.48>
+- <13.50>
+- <13.52>
+- <13.54>
+- <13.56>
+- <13.58>
+- <13.60>
+- <13.62>
+- <13.64>
+- <13.66>
+- <13.68>
+- <13.70>
+- <13.72>
+- <13.74>
+- <13.76>
+- <13.78>
+- <13.80>
+- <13.82>
+- <13.84>
+- <13.86>
+- <13.88>
+- <13.90>
+- <13.92>
+- <13.94>
+- <13.96>
+- <13.98>
+- <14.00>
+- <14.02>
+- <14.04>
+- <14.06>
+- <14.08>
+- <14.10>
+- <14.12>
+- <14.14>
+- <14.16>
+- <14.18>
+- <14.20>
+- <14.22>
+- <14.24>
+- <14.26>
+- <14.28>
+- <14.30>
+- <14.32>
+- <14.34>
+- <14.36>
+- <14.38>
+- <14.40>
+- <14.42>
+- <14.44>
+- <14.46>
+- <14.48>
+- <14.50>
+- <14.52>
+- <14.54>
+- <14.56>
+- <14.58>
+- <14.60>
+- <14.62>
+- <14.64>
+- <14.66>
+- <14.68>
+- <14.70>
+- <14.72>
+- <14.74>
+- <14.76>
+- <14.78>
+- <14.80>
+- <14.82>
+- <14.84>
+- <14.86>
+- <14.88>
+- <14.90>
+- <14.92>
+- <14.94>
+- <14.96>
+- <14.98>
+- <15.00>
+- <15.02>
+- <15.04>
+- <15.06>
+- <15.08>
+- <15.10>
+- <15.12>
+- <15.14>
+- <15.16>
+- <15.18>
+- <15.20>
+- <15.22>
+- <15.24>
+- <15.26>
+- <15.28>
+- <15.30>
+- <15.32>
+- <15.34>
+- <15.36>
+- <15.38>
+- <15.40>
+- <15.42>
+- <15.44>
+- <15.46>
+- <15.48>
+- <15.50>
+- <15.52>
+- <15.54>
+- <15.56>
+- <15.58>
+- <15.60>
+- <15.62>
+- <15.64>
+- <15.66>
+- <15.68>
+- <15.70>
+- <15.72>
+- <15.74>
+- <15.76>
+- <15.78>
+- <15.80>
+- <15.82>
+- <15.84>
+- <15.86>
+- <15.88>
+- <15.90>
+- <15.92>
+- <15.94>
+- <15.96>
+- <15.98>
+- <16.00>
+- <16.02>
+- <16.04>
+- <16.06>
+- <16.08>
+- <16.10>
+- <16.12>
+- <16.14>
+- <16.16>
+- <16.18>
+- <16.20>
+- <16.22>
+- <16.24>
+- <16.26>
+- <16.28>
+- <16.30>
+- <16.32>
+- <16.34>
+- <16.36>
+- <16.38>
+- <16.40>
+- <16.42>
+- <16.44>
+- <16.46>
+- <16.48>
+- <16.50>
+- <16.52>
+- <16.54>
+- <16.56>
+- <16.58>
+- <16.60>
+- <16.62>
+- <16.64>
+- <16.66>
+- <16.68>
+- <16.70>
+- <16.72>
+- <16.74>
+- <16.76>
+- <16.78>
+- <16.80>
+- <16.82>
+- <16.84>
+- <16.86>
+- <16.88>
+- <16.90>
+- <16.92>
+- <16.94>
+- <16.96>
+- <16.98>
+- <17.00>
+- <17.02>
+- <17.04>
+- <17.06>
+- <17.08>
+- <17.10>
+- <17.12>
+- <17.14>
+- <17.16>
+- <17.18>
+- <17.20>
+- <17.22>
+- <17.24>
+- <17.26>
+- <17.28>
+- <17.30>
+- <17.32>
+- <17.34>
+- <17.36>
+- <17.38>
+- <17.40>
+- <17.42>
+- <17.44>
+- <17.46>
+- <17.48>
+- <17.50>
+- <17.52>
+- <17.54>
+- <17.56>
+- <17.58>
+- <17.60>
+- <17.62>
+- <17.64>
+- <17.66>
+- <17.68>
+- <17.70>
+- <17.72>
+- <17.74>
+- <17.76>
+- <17.78>
+- <17.80>
+- <17.82>
+- <17.84>
+- <17.86>
+- <17.88>
+- <17.90>
+- <17.92>
+- <17.94>
+- <17.96>
+- <17.98>
+- <18.00>
+- <18.02>
+- <18.04>
+- <18.06>
+- <18.08>
+- <18.10>
+- <18.12>
+- <18.14>
+- <18.16>
+- <18.18>
+- <18.20>
+- <18.22>
+- <18.24>
+- <18.26>
+- <18.28>
+- <18.30>
+- <18.32>
+- <18.34>
+- <18.36>
+- <18.38>
+- <18.40>
+- <18.42>
+- <18.44>
+- <18.46>
+- <18.48>
+- <18.50>
+- <18.52>
+- <18.54>
+- <18.56>
+- <18.58>
+- <18.60>
+- <18.62>
+- <18.64>
+- <18.66>
+- <18.68>
+- <18.70>
+- <18.72>
+- <18.74>
+- <18.76>
+- <18.78>
+- <18.80>
+- <18.82>
+- <18.84>
+- <18.86>
+- <18.88>
+- <18.90>
+- <18.92>
+- <18.94>
+- <18.96>
+- <18.98>
+- <19.00>
+- <19.02>
+- <19.04>
+- <19.06>
+- <19.08>
+- <19.10>
+- <19.12>
+- <19.14>
+- <19.16>
+- <19.18>
+- <19.20>
+- <19.22>
+- <19.24>
+- <19.26>
+- <19.28>
+- <19.30>
+- <19.32>
+- <19.34>
+- <19.36>
+- <19.38>
+- <19.40>
+- <19.42>
+- <19.44>
+- <19.46>
+- <19.48>
+- <19.50>
+- <19.52>
+- <19.54>
+- <19.56>
+- <19.58>
+- <19.60>
+- <19.62>
+- <19.64>
+- <19.66>
+- <19.68>
+- <19.70>
+- <19.72>
+- <19.74>
+- <19.76>
+- <19.78>
+- <19.80>
+- <19.82>
+- <19.84>
+- <19.86>
+- <19.88>
+- <19.90>
+- <19.92>
+- <19.94>
+- <19.96>
+- <19.98>
+- <20.00>
+- <20.02>
+- <20.04>
+- <20.06>
+- <20.08>
+- <20.10>
+- <20.12>
+- <20.14>
+- <20.16>
+- <20.18>
+- <20.20>
+- <20.22>
+- <20.24>
+- <20.26>
+- <20.28>
+- <20.30>
+- <20.32>
+- <20.34>
+- <20.36>
+- <20.38>
+- <20.40>
+- <20.42>
+- <20.44>
+- <20.46>
+- <20.48>
+- <20.50>
+- <20.52>
+- <20.54>
+- <20.56>
+- <20.58>
+- <20.60>
+- <20.62>
+- <20.64>
+- <20.66>
+- <20.68>
+- <20.70>
+- <20.72>
+- <20.74>
+- <20.76>
+- <20.78>
+- <20.80>
+- <20.82>
+- <20.84>
+- <20.86>
+- <20.88>
+- <20.90>
+- <20.92>
+- <20.94>
+- <20.96>
+- <20.98>
+- <21.00>
+- <21.02>
+- <21.04>
+- <21.06>
+- <21.08>
+- <21.10>
+- <21.12>
+- <21.14>
+- <21.16>
+- <21.18>
+- <21.20>
+- <21.22>
+- <21.24>
+- <21.26>
+- <21.28>
+- <21.30>
+- <21.32>
+- <21.34>
+- <21.36>
+- <21.38>
+- <21.40>
+- <21.42>
+- <21.44>
+- <21.46>
+- <21.48>
+- <21.50>
+- <21.52>
+- <21.54>
+- <21.56>
+- <21.58>
+- <21.60>
+- <21.62>
+- <21.64>
+- <21.66>
+- <21.68>
+- <21.70>
+- <21.72>
+- <21.74>
+- <21.76>
+- <21.78>
+- <21.80>
+- <21.82>
+- <21.84>
+- <21.86>
+- <21.88>
+- <21.90>
+- <21.92>
+- <21.94>
+- <21.96>
+- <21.98>
+- <22.00>
+- <22.02>
+- <22.04>
+- <22.06>
+- <22.08>
+- <22.10>
+- <22.12>
+- <22.14>
+- <22.16>
+- <22.18>
+- <22.20>
+- <22.22>
+- <22.24>
+- <22.26>
+- <22.28>
+- <22.30>
+- <22.32>
+- <22.34>
+- <22.36>
+- <22.38>
+- <22.40>
+- <22.42>
+- <22.44>
+- <22.46>
+- <22.48>
+- <22.50>
+- <22.52>
+- <22.54>
+- <22.56>
+- <22.58>
+- <22.60>
+- <22.62>
+- <22.64>
+- <22.66>
+- <22.68>
+- <22.70>
+- <22.72>
+- <22.74>
+- <22.76>
+- <22.78>
+- <22.80>
+- <22.82>
+- <22.84>
+- <22.86>
+- <22.88>
+- <22.90>
+- <22.92>
+- <22.94>
+- <22.96>
+- <22.98>
+- <23.00>
+- <23.02>
+- <23.04>
+- <23.06>
+- <23.08>
+- <23.10>
+- <23.12>
+- <23.14>
+- <23.16>
+- <23.18>
+- <23.20>
+- <23.22>
+- <23.24>
+- <23.26>
+- <23.28>
+- <23.30>
+- <23.32>
+- <23.34>
+- <23.36>
+- <23.38>
+- <23.40>
+- <23.42>
+- <23.44>
+- <23.46>
+- <23.48>
+- <23.50>
+- <23.52>
+- <23.54>
+- <23.56>
+- <23.58>
+- <23.60>
+- <23.62>
+- <23.64>
+- <23.66>
+- <23.68>
+- <23.70>
+- <23.72>
+- <23.74>
+- <23.76>
+- <23.78>
+- <23.80>
+- <23.82>
+- <23.84>
+- <23.86>
+- <23.88>
+- <23.90>
+- <23.92>
+- <23.94>
+- <23.96>
+- <23.98>
+- <24.00>
+- <24.02>
+- <24.04>
+- <24.06>
+- <24.08>
+- <24.10>
+- <24.12>
+- <24.14>
+- <24.16>
+- <24.18>
+- <24.20>
+- <24.22>
+- <24.24>
+- <24.26>
+- <24.28>
+- <24.30>
+- <24.32>
+- <24.34>
+- <24.36>
+- <24.38>
+- <24.40>
+- <24.42>
+- <24.44>
+- <24.46>
+- <24.48>
+- <24.50>
+- <24.52>
+- <24.54>
+- <24.56>
+- <24.58>
+- <24.60>
+- <24.62>
+- <24.64>
+- <24.66>
+- <24.68>
+- <24.70>
+- <24.72>
+- <24.74>
+- <24.76>
+- <24.78>
+- <24.80>
+- <24.82>
+- <24.84>
+- <24.86>
+- <24.88>
+- <24.90>
+- <24.92>
+- <24.94>
+- <24.96>
+- <24.98>
+- <25.00>
+- <25.02>
+- <25.04>
+- <25.06>
+- <25.08>
+- <25.10>
+- <25.12>
+- <25.14>
+- <25.16>
+- <25.18>
+- <25.20>
+- <25.22>
+- <25.24>
+- <25.26>
+- <25.28>
+- <25.30>
+- <25.32>
+- <25.34>
+- <25.36>
+- <25.38>
+- <25.40>
+- <25.42>
+- <25.44>
+- <25.46>
+- <25.48>
+- <25.50>
+- <25.52>
+- <25.54>
+- <25.56>
+- <25.58>
+- <25.60>
+- <25.62>
+- <25.64>
+- <25.66>
+- <25.68>
+- <25.70>
+- <25.72>
+- <25.74>
+- <25.76>
+- <25.78>
+- <25.80>
+- <25.82>
+- <25.84>
+- <25.86>
+- <25.88>
+- <25.90>
+- <25.92>
+- <25.94>
+- <25.96>
+- <25.98>
+- <26.00>
+- <26.02>
+- <26.04>
+- <26.06>
+- <26.08>
+- <26.10>
+- <26.12>
+- <26.14>
+- <26.16>
+- <26.18>
+- <26.20>
+- <26.22>
+- <26.24>
+- <26.26>
+- <26.28>
+- <26.30>
+- <26.32>
+- <26.34>
+- <26.36>
+- <26.38>
+- <26.40>
+- <26.42>
+- <26.44>
+- <26.46>
+- <26.48>
+- <26.50>
+- <26.52>
+- <26.54>
+- <26.56>
+- <26.58>
+- <26.60>
+- <26.62>
+- <26.64>
+- <26.66>
+- <26.68>
+- <26.70>
+- <26.72>
+- <26.74>
+- <26.76>
+- <26.78>
+- <26.80>
+- <26.82>
+- <26.84>
+- <26.86>
+- <26.88>
+- <26.90>
+- <26.92>
+- <26.94>
+- <26.96>
+- <26.98>
+- <27.00>
+- <27.02>
+- <27.04>
+- <27.06>
+- <27.08>
+- <27.10>
+- <27.12>
+- <27.14>
+- <27.16>
+- <27.18>
+- <27.20>
+- <27.22>
+- <27.24>
+- <27.26>
+- <27.28>
+- <27.30>
+- <27.32>
+- <27.34>
+- <27.36>
+- <27.38>
+- <27.40>
+- <27.42>
+- <27.44>
+- <27.46>
+- <27.48>
+- <27.50>
+- <27.52>
+- <27.54>
+- <27.56>
+- <27.58>
+- <27.60>
+- <27.62>
+- <27.64>
+- <27.66>
+- <27.68>
+- <27.70>
+- <27.72>
+- <27.74>
+- <27.76>
+- <27.78>
+- <27.80>
+- <27.82>
+- <27.84>
+- <27.86>
+- <27.88>
+- <27.90>
+- <27.92>
+- <27.94>
+- <27.96>
+- <27.98>
+- <28.00>
+- <28.02>
+- <28.04>
+- <28.06>
+- <28.08>
+- <28.10>
+- <28.12>
+- <28.14>
+- <28.16>
+- <28.18>
+- <28.20>
+- <28.22>
+- <28.24>
+- <28.26>
+- <28.28>
+- <28.30>
+- <28.32>
+- <28.34>
+- <28.36>
+- <28.38>
+- <28.40>
+- <28.42>
+- <28.44>
+- <28.46>
+- <28.48>
+- <28.50>
+- <28.52>
+- <28.54>
+- <28.56>
+- <28.58>
+- <28.60>
+- <28.62>
+- <28.64>
+- <28.66>
+- <28.68>
+- <28.70>
+- <28.72>
+- <28.74>
+- <28.76>
+- <28.78>
+- <28.80>
+- <28.82>
+- <28.84>
+- <28.86>
+- <28.88>
+- <28.90>
+- <28.92>
+- <28.94>
+- <28.96>
+- <28.98>
+- <29.00>
+- <29.02>
+- <29.04>
+- <29.06>
+- <29.08>
+- <29.10>
+- <29.12>
+- <29.14>
+- <29.16>
+- <29.18>
+- <29.20>
+- <29.22>
+- <29.24>
+- <29.26>
+- <29.28>
+- <29.30>
+- <29.32>
+- <29.34>
+- <29.36>
+- <29.38>
+- <29.40>
+- <29.42>
+- <29.44>
+- <29.46>
+- <29.48>
+- <29.50>
+- <29.52>
+- <29.54>
+- <29.56>
+- <29.58>
+- <29.60>
+- <29.62>
+- <29.64>
+- <29.66>
+- <29.68>
+- <29.70>
+- <29.72>
+- <29.74>
+- <29.76>
+- <29.78>
+- <29.80>
+- <29.82>
+- <29.84>
+- <29.86>
+- <29.88>
+- <29.90>
+- <29.92>
+- <29.94>
+- <29.96>
+- <29.98>
+- <30.00>
+- ▁
+- ▁the
+- ▁and
+- ▁of
+- s
+- ▁to
+- ▁a
+- .
+- ▁in
+- ▁i
+- ''''
+- ▁he
+- ▁that
+- ▁was
+- ▁it
+- ▁his
+- ','
+- ▁for
+- ▁with
+- '-'
+- ▁is
+- ▁you
+- ▁as
+- ▁had
+- t
+- ▁de
+- ▁be
+- ▁her
+- ▁not
+- e
+- ▁but
+- ▁on
+- ▁at
+- ▁she
+- ▁by
+- ▁have
+- ▁they
+- ▁him
+- ▁this
+- ▁which
+- ▁all
+- ▁from
+- ▁la
+- ▁so
+- ▁me
+- ▁my
+- ▁were
+- ▁we
+- ▁one
+- ▁no
+- n
+- a
+- en
+- d
+- ▁an
+- ▁or
+- ▁are
+- ▁said
+- ▁there
+- ▁their
+- o
+- ▁when
+- ▁would
+- ▁if
+- ▁them
+- ▁who
+- ▁en
+- ▁been
+- ▁what
+- ing
+- er
+- ▁will
+- ▁do
+- i
+- ▁out
+- ed
+- m
+- ▁man
+- ▁up
+- ▁then
+- es
+- ▁more
+- ▁into
+- ▁could
+- y
+- ▁now
+- r
+- ▁some
+- ▁your
+- ▁very
+- ▁has
+- ▁time
+- ▁die
+- ▁can
+- re
+- ly
+- ▁about
+- ▁que
+- ▁little
+- ▁than
+- ▁like
+- ▁only
+- ▁un
+- ▁our
+- ▁over
+- ▁upon
+- ▁any
+- ▁did
+- ▁other
+- ▁two
+- '?'
+- ▁und
+- ▁l
+- ▁know
+- ▁see
+- ▁well
+- ▁y
+- ▁before
+- ▁after
+- ▁made
+- ▁d
+- ▁der
+- ▁good
+- ▁el
+- ▁down
+- ▁its
+- ▁should
+- te
+- ▁great
+- ▁such
+- ▁us
+- ▁these
+- ▁come
+- ▁le
+- ▁men
+- ▁go
+- ▁how
+- ▁must
+- ▁may
+- ▁first
+- ▁na
+- ▁much
+- ▁never
+- ▁old
+- ▁mr
+- ▁where
+- ▁came
+- ▁way
+- in
+- ▁am
+- u
+- ’
+- ll
+- ▁long
+- ▁back
+- ▁day
+- ▁here
+- ▁own
+- ▁mu
+- ▁himself
+- ▁even
+- ▁again
+- ▁life
+- ▁say
+- ▁those
+- ▁through
+- ▁most
+- ▁just
+- ▁think
+- ▁make
+- ▁se
+- ▁might
+- ▁went
+- ve
+- ▁al
+- ▁many
+- ▁don
+- ▁too
+- ▁every
+- ▁et
+- ▁shall
+- ▁thought
+- ▁being
+- ▁hand
+- ▁away
+- ▁still
+- ▁La
+- ▁den
+- ▁without
+- j
+- na
+- ▁while
+- ▁people
+- ▁les
+- ▁take
+- de
+- ar
+- ▁last
+- ▁eyes
+- ▁god
+- ▁o
+- 'on'
+- ▁van
+- ▁yet
+- ▁ku
+- ▁des
+- al
+- is
+- ▁three
+- ▁found
+- ▁also
+- ▁get
+- ▁nothing
+- ▁du
+- ▁place
+- ▁same
+- ▁house
+- ▁though
+- ▁n
+- to
+- ▁under
+- ▁new
+- ▁let
+- ▁once
+- ▁The
+- ▁another
+- ▁ya
+- ▁face
+- ▁à
+- ▁off
+- ▁saw
+- ▁things
+- an
+- ▁per
+- '!'
+- ▁young
+- ▁war
+- ▁father
+- ▁head
+- ▁es
+- ▁tell
+- ▁ever
+- ▁er
+- ▁te
+- ▁right
+- ▁zu
+- l
+- ▁del
+- ▁far
+- ▁night
+- k
+- ▁world
+- ▁put
+- ▁left
+- ▁ne
+- ▁I
+- ta
+- ▁work
+- ▁because
+- ▁con
+- ▁part
+- ▁took
+- ▁give
+- ▁look
+- ▁love
+- est
+- le
+- ▁why
+- ▁di
+- ▁always
+- ▁het
+- ▁A
+- ▁looked
+- ▁door
+- z
+- ▁thing
+- ▁against
+- as
+- ▁years
+- ▁mind
+- la
+- ne
+- ▁een
+- ▁home
+- ▁seemed
+- ▁sie
+- ▁moment
+- ▁going
+- ▁asked
+- ten
+- ▁heart
+- ▁got
+- il
+- ▁mrs
+- ▁son
+- ▁something
+- ▁each
+- se
+- é
+- ▁mother
+- ▁told
+- ▁knew
+- ▁few
+- ra
+- man
+- ▁sir
+- ▁heard
+- ga
+- ▁find
+- ▁das
+- da
+- ▁і
+- ▁una
+- ▁woman
+- ▁between
+- ge
+- ▁oh
+- st
+- ▁met
+- ▁done
+- ▁half
+- ▁wa
+- ▁king
+- ▁room
+- ▁both
+- ▁dat
+- ▁water
+- ▁side
+- os
+- ▁yes
+- ▁als
+- ▁better
+- ▁ein
+- ▁seen
+- us
+- ▁called
+- ▁soon
+- е
+- ▁quite
+- ▁name
+- ▁va
+- ▁course
+- ▁want
+- ▁ich
+- ▁enough
+- ▁having
+- ro
+- ▁il
+- ▁не
+- ▁ik
+- ch
+- ▁ko
+- ▁whole
+- ▁light
+- ▁mit
+- ▁nicht
+- ▁miss
+- do
+- c
+- 'no'
+- ▁ni
+- ▁ver
+- ▁lord
+- mo
+- it
+- ▁sich
+- ▁whom
+- ▁nor
+- un
+- ▁lady
+- а
+- ho
+- ment
+- ▁gave
+- ness
+- ▁El
+- ▁von
+- ▁set
+- ▁almost
+- ▁however
+- ▁hands
+- ▁days
+- ▁end
+- ▁began
+- م
+- ▁white
+- ▁w
+- я
+- м
+- ▁je
+- f
+- ▁turned
+- ▁на
+- ▁auf
+- у
+- ▁ha
+- h
+- ▁da
+- ▁poor
+- ▁among
+- ▁ist
+- ▁morning
+- ▁until
+- ▁ba
+- ▁por
+- ▁ma
+- ▁su
+- ka
+- ▁felt
+- ma
+- ▁estas
+- ▁words
+- ▁"
+- ▁op
+- ▁kind
+- ▁four
+- ▁since
+- ▁best
+- ▁country
+- ▁death
+- ▁voice
+- ▁si
+- we
+- ▁hundred
+- ▁brought
+- ▁small
+- me
+- й
+- ▁next
+- ers
+- ▁word
+- be
+- 。
+- ▁perhaps
+- ▁does
+- ▁boy
+- w
+- ▁myself
+- ▁stood
+- ▁anything
+- ▁girl
+- ce
+- ▁round
+- ▁з
+- ni
+- ▁hij
+- ▁ge
+- va
+- ▁est
+- ▁herself
+- ▁los
+- 이
+- ▁u
+- ▁dear
+- ▁matter
+- ▁thus
+- b
+- p
+- ▁rather
+- ▁taken
+- ▁till
+- ▁full
+- wa
+- den
+- ▁dem
+- ы
+- ▁par
+- ▁five
+- je
+- ▁present
+- za
+- ▁friend
+- ▁sent
+- ▁high
+- ba
+- ▁together
+- ▁twenty
+- li
+- g
+- ▁kaj
+- ▁He
+- ▁mi
+- ▁air
+- ▁general
+- ▁thou
+- el
+- ▁true
+- at
+- ▁indeed
+- ▁near
+- ▁open
+- he
+- ▁rest
+- ▁power
+- ▁large
+- ▁children
+- ▁sure
+- ke
+- ▁keep
+- ▁themselves
+- ▁whose
+- ko
+- ▁fire
+- ▁feet
+- ▁wife
+- ▁others
+- men
+- ▁given
+- ▁cannot
+- ▁nature
+- sa
+- ▁sea
+- go
+- ▁qui
+- ▁money
+- ti
+- ▁certain
+- ▁believe
+- ▁help
+- ▁gone
+- ir
+- ▁cried
+- ▁along
+- th
+- ▁black
+- ▁па
+- ca
+- ▁ў
+- ▁used
+- ▁c
+- ▁second
+- ▁case
+- ▁point
+- zi
+- ▁often
+- ▁child
+- ▁passed
+- ter
+- ▁niet
+- ha
+- ▁sat
+- ▁fact
+- ▁hear
+- or
+- ▁year
+- ▁alone
+- ▁S
+- ia
+- ▁order
+- ▁그
+- ▁known
+- ri
+- ▁therefore
+- ▁during
+- ▁use
+- ▁call
+- ▁li
+- et
+- lo
+- т
+- ▁im
+- ó
+- ▁dans
+- ▁ten
+- ▁zijn
+- ▁behind
+- em
+- ▁lay
+- ▁body
+- ▁state
+- ▁hope
+- ▁really
+- ی
+- 가
+- ya
+- ▁above
+- ▁looking
+- ▁pas
+- ▁leave
+- ▁ye
+- х
+- ton
+- ▁answered
+- ▁land
+- ▁sa
+- ▁replied
+- ▁um
+- and
+- ▁b
+- ▁city
+- ▁red
+- ▁thousand
+- der
+- ▁dead
+- ▁lo
+- ▁mean
+- ▁six
+- by
+- ▁held
+- ▁hard
+- ▁within
+- на
+- ▁fell
+- ▁around
+- ▁less
+- ▁earth
+- ▁least
+- ry
+- ▁already
+- ▁whether
+- ur
+- с
+- ▁hour
+- ن
+- ▁speak
+- ▁read
+- ▁thy
+- ▁itself
+- co
+- ▁és
+- ▁became
+- ▁Il
+- '"'
+- ▁women
+- ▁reason
+- ▁eine
+- ▁거
+- ▁une
+- ▁у
+- ▁L
+- ▁care
+- ▁hat
+- ci
+- ':'
+- ▁soul
+- ▁says
+- ▁friends
+- ▁several
+- ▁person
+- ▁live
+- ant
+- ▁coming
+- ▁short
+- x
+- ▁means
+- ▁question
+- ▁In
+- ▁sort
+- ▁times
+- ▁pour
+- ت
+- ▁за
+- ▁dark
+- ка
+- ▁wie
+- oj
+- ▁au
+- ▁ground
+- ▁possible
+- di
+- ie
+- ▁form
+- ▁human
+- ▁amb
+- ла
+- ▁either
+- sch
+- ▁feel
+- ki
+- ▁business
+- ▁town
+- ▁lost
+- ▁com
+- ▁bed
+- ▁turn
+- ▁qu
+- ▁En
+- ш
+- ▁zij
+- ▁big
+- nt
+- ▁sur
+- ▁Li
+- ▁family
+- ja
+- ver
+- ▁won
+- ▁suddenly
+- ▁become
+- к
+- ▁It
+- ▁everything
+- ▁manner
+- ▁later
+- aba
+- ▁kept
+- 는
+- ▁truth
+- ▁master
+- ▁aus
+- ▁ser
+- ▁sun
+- à
+- ▁bi
+- um
+- ▁letter
+- ▁didn
+- ▁wish
+- ▁rose
+- ▁spirit
+- ▁evening
+- ▁making
+- ▁fear
+- ▁need
+- ation
+- ▁close
+- ▁returned
+- ▁idea
+- ze
+- ▁brother
+- ▁om
+- ▁show
+- ▁tu
+- so
+- 도
+- ▁strong
+- ▁sense
+- ▁church
+- the
+- ▁ask
+- ▁sight
+- ▁para
+- ▁understand
+- able
+- ▁aan
+- ▁sometimes
+- ▁across
+- ▁number
+- ▁different
+- 的
+- son
+- ▁table
+- ▁river
+- ▁talk
+- ل
+- ▁bring
+- р
+- ent
+- ▁vor
+- am
+- ▁towards
+- ▁return
+- vo
+- ▁answer
+- ▁beautiful
+- ▁real
+- ▁아
+- ▁else
+- mi
+- п
+- ▁Le
+- ity
+- ▁fine
+- ▁arm
+- ▁public
+- ▁ready
+- ▁voor
+- day
+- ▁story
+- ▁doubt
+- ▁followed
+- mu
+- д
+- ▁strange
+- ▁p
+- ▁auch
+- ▁gu
+- ▁aber
+- ▁z
+- ▁—
+- v
+- ▁stand
+- ▁captain
+- gen
+- less
+- ▁law
+- ▁hold
+- ▁sin
+- ،
+- ▁suppose
+- ye
+- ▁able
+- ▁bad
+- ▁spoke
+- ▁што
+- ▁run
+- ▁sound
+- tu
+- лі
+- ▁free
+- ig
+- ting
+- ▁john
+- ▁да
+- land
+- ste
+- ▁Es
+- ▁arms
+- ▁cold
+- н
+- ▁ja
+- ▁car
+- 지
+- ▁book
+- л
+- ▁english
+- ak
+- ado
+- ▁eye
+- ▁front
+- ▁za
+- ▁art
+- si
+- ▁wind
+- ▁pass
+- ▁ad
+- ling
+- ▁deep
+- ▁husband
+- ▁received
+- і
+- ▁remember
+- ▁cause
+- id
+- ▁А
+- ▁past
+- ▁horse
+- ▁reached
+- ▁ce
+- ▁street
+- та
+- ▁cut
+- ▁road
+- ada
+- du
+- ▁haar
+- ▁근데
+- ▁england
+- ▁era
+- ▁daß
+- yo
+- ▁У
+- ▁appeared
+- ▁hair
+- ▁happy
+- ▁saying
+- га
+- ▁wanted
+- ▁pro
+- ▁longer
+- ▁gold
+- ▁seems
+- ه
+- ▁early
+- ▁vi
+- ▁blood
+- ▁che
+- ▁안
+- ▁ought
+- ▁subject
+- aj
+- ▁carried
+- 을
+- hi
+- ▁po
+- ▁em
+- 고
+- ▁low
+- ▁Die
+- ▁wild
+- ▁taking
+- ў
+- ▁dan
+- ▁clear
+- ▁line
+- ven
+- zo
+- ▁none
+- ▁feeling
+- bo
+- ▁forward
+- po
+- ▁company
+- ▁nach
+- ▁further
+- ▁eighteen
+- ▁ran
+- ▁continued
+- ▁toward
+- ▁certainly
+- ر
+- ▁change
+- ▁hours
+- ▁maar
+- ра
+- ▁window
+- ▁dit
+- ad
+- ▁seven
+- ▁common
+- ▁seem
+- ▁daughter
+- pa
+- ine
+- ic
+- ▁De
+- wi
+- ▁natural
+- ▁fellow
+- ▁beyond
+- ▁á
+- ku
+- 에
+- ▁las
+- ▁neither
+- ▁doing
+- ▁court
+- ▁green
+- ▁vous
+- ▁force
+- ▁meet
+- ▁tried
+- ны
+- '".'
+- ▁living
+- ted
+- ▁although
+- ▁doctor
+- ▁school
+- ▁bu
+- im
+- ▁lui
+- ▁interest
+- ▁mal
+- ▁opened
+- ▁noch
+- ▁play
+- ку
+- ▁position
+- ▁O
+- ▁account
+- room
+- ▁pretty
+- ▁mir
+- ▁fair
+- ▁glad
+- ▁fall
+- ▁character
+- ▁party
+- bi
+- ▁necessary
+- ▁els
+- ty
+- ▁estis
+- ▁blue
+- ▁added
+- ▁für
+- ▁act
+- ▁co
+- ▁except
+- q
+- ▁plus
+- ▁silence
+- ▁miles
+- ▁eight
+- gi
+- ▁government
+- ▁self
+- ▁sleep
+- ▁ki
+- ▁thirty
+- ▁k
+- кі
+- ▁nearly
+- ▁yo
+- ▁nur
+- ▁nu
+- ▁bear
+- ش
+- ▁cap
+- ؟
+- les
+- ▁mar
+- ▁foot
+- ▁prince
+- ▁mine
+- que
+- ken
+- fa
+- ▁bo
+- ▁christ
+- ▁entered
+- ▁D
+- zu
+- ny
+- ▁wall
+- ▁following
+- ▁ah
+- ng
+- ▁ab
+- ren
+- ▁knowledge
+- ▁minutes
+- ▁heaven
+- ▁M
+- ▁happened
+- ▁view
+- ▁late
+- ▁non
+- ▁Mu
+- ▁lived
+- ▁army
+- ▁try
+- ▁ara
+- ▁а
+- ▁chance
+- ▁queen
+- ▁ship
+- ▁bit
+- ▁hem
+- ▁forth
+- ▁Re
+- ist
+- ▁sister
+- ▁attention
+- ▁N
+- ▁peace
+- led
+- ▁yourself
+- ▁и
+- ▁뭐
+- ▁pay
+- ▁dr
+- ▁french
+- ▁über
+- ▁어
+- ung
+- ait
+- ▁hall
+- ▁ago
+- ▁como
+- ▁fifty
+- да
+- ley
+- ▁Se
+- tion
+- ner
+- ▁trouble
+- ▁afraid
+- ▁purpose
+- ▁gentleman
+- ▁service
+- ▁می
+- ▁age
+- з
+- ▁important
+- ▁north
+- ion
+- ي
+- и
+- ▁distance
+- ▁led
+- vi
+- ▁secret
+- ▁third
+- ▁save
+- ▁Na
+- jo
+- ▁zoo
+- ▁thinking
+- ▁hi
+- ut
+- pe
+- ck
+- ▁с
+- ▁walked
+- au
+- ▁wir
+- ▁و
+- ▁laid
+- ▁south
+- ek
+- ▁length
+- ish
+- ▁died
+- ▁comes
+- ▁pleasure
+- ▁standing
+- ▁trees
+- ▁пра
+- ▁sweet
+- ▁effect
+- ж
+- ю
+- ▁slowly
+- ▁probably
+- ▁muri
+- ▁wait
+- ger
+- 어
+- ▁wrong
+- ▁smile
+- ▁paper
+- wo
+- ▁whatever
+- ▁heavy
+- ana
+- ▁london
+- ▁гэта
+- ita
+- ▁desire
+- ▁week
+- ▁mary
+- ▁dis
+- ▁rich
+- ▁deal
+- ▁quiet
+- ▁drew
+- ▁pre
+- tes
+- ▁pri
+- ▁faith
+- не
+- ▁instead
+- ▁ill
+- ina
+- ▁wonder
+- ▁st
+- ▁simple
+- ▁mon
+- ▁lips
+- ▁camp
+- ▁tree
+- nie
+- ▁object
+- tt
+- ▁hardly
+- ru
+- ▁Per
+- ▁stay
+- ▁started
+- ▁remained
+- elle
+- د
+- clock
+- ▁mouth
+- í
+- ▁T
+- ▁post
+- ▁stopped
+- ▁top
+- ▁caught
+- era
+- ▁immediately
+- ▁history
+- ez
+- ▁send
+- ▁C
+- ▁exclaimed
+- ▁U
+- ▁uncle
+- و
+- ▁fast
+- ▁No
+- ▁strength
+- ▁mij
+- io
+- ▁einen
+- ai
+- ▁sind
+- ▁chief
+- ▁joy
+- ▁dinner
+- ▁boys
+- ▁g
+- ▁mais
+- ▁uit
+- ▁food
+- fe
+- gu
+- ▁kon
+- ▁office
+- ты
+- ▁placed
+- ▁getting
+- ▁walk
+- ▁Al
+- ▁floor
+- ▁moved
+- ▁fa
+- ▁bright
+- ▁written
+- ▁outside
+- ▁zich
+- ▁Ma
+- ▁single
+- ▁wood
+- ▁grew
+- ▁watch
+- ▁E
+- ▁в
+- ▁по
+- ▁nine
+- ▁beauty
+- ▁loved
+- ▁anar
+- ▁visit
+- ز
+- age
+- ▁seeing
+- ▁follow
+- ▁ka
+- ster
+- ▁months
+- vu
+- ▁impossible
+- ▁ад
+- ▁Mi
+- ▁broken
+- ▁dog
+- ard
+- ▁Les
+- ▁wat
+- ▁forty
+- ring
+- end
+- ▁easy
+- ▁stone
+- une
+- ▁stop
+- ▁werden
+- ▁boat
+- ▁greater
+- ▁sit
+- ▁v
+- ні
+- ва
+- ▁please
+- ▁figure
+- ▁filled
+- ▁Ha
+- ▁hier
+- ▁І
+- ▁sitting
+- ▁вы
+- ▁eat
+- eth
+- ▁cast
+- ▁laughed
+- ▁Er
+- 은
+- ▁evil
+- ▁bar
+- ▁battle
+- ▁quickly
+- ▁enemy
+- ▁presence
+- ful
+- ▁waiting
+- ▁according
+- ▁ihm
+- 서
+- ▁struck
+- ле
+- ▁chair
+- ▁wo
+- ▁wurde
+- ▁avec
+- ▁music
+- ▁beginning
+- ▁married
+- ма
+- ▁mich
+- ▁makes
+- ▁hatte
+- ▁duty
+- ▁sein
+- ▁kwa
+- ler
+- ▁oku
+- ▁west
+- ▁plain
+- ▁ihr
+- cu
+- ▁Ba
+- ▁plan
+- ▁막
+- ▁fight
+- ▁més
+- ▁regard
+- ▁especially
+- ▁due
+- ▁perfect
+- ik
+- ▁danger
+- ▁village
+- ▁pa
+- ▁below
+- ▁bien
+- ▁durch
+- ate
+- ▁particular
+- ▁opinion
+- ▁beside
+- ч
+- ▁ho
+- г
+- ▁showed
+- ▁tears
+- ble
+- ▁step
+- ▁france
+- ▁vol
+- ▁latter
+- des
+- ▁grand
+- ham
+- '...'
+- ism
+- ▁f
+- ben
+- ▁garden
+- ▁girls
+- ▁holy
+- ▁brown
+- che
+- ▁states
+- ▁afternoon
+- ▁fer
+- ▁besides
+- lar
+- ▁fresh
+- ▁future
+- аў
+- ▁snow
+- ▁appearance
+- ▁silent
+- ▁winter
+- ▁wide
+- ▁result
+- mer
+- ▁cross
+- bu
+- ▁Be
+- ▁aba
+- ды
+- ب
+- ura
+- ▁conversation
+- ▁entre
+- ▁carry
+- ▁sudden
+- ▁horses
+- ▁grace
+- sh
+- ▁tom
+- ▁comme
+- ▁hot
+- ▁learned
+- ▁kandi
+- ▁merely
+- ▁raised
+- ▁arrived
+- ▁wrote
+- س
+- ran
+- ▁straight
+- ▁thoughts
+- ля
+- ▁youth
+- ▁knows
+- ▁various
+- A
+- ць
+- ▁Я
+- ▁society
+- ▁unto
+- ▁leaving
+- ▁action
+- ▁meant
+- ча
+- ль
+- ▁broke
+- chen
+- ▁sky
+- ▁summer
+- ▁terrible
+- ▁born
+- ▁condition
+- bra
+- ▁direction
+- ▁tre
+- ▁write
+- ▁Ur
+- ul
+- zen
+- ▁expression
+- ▁dé
+- ▁ke
+- ▁unter
+- ▁journey
+- ▁An
+- ▁wird
+- ▁spring
+- ▁trying
+- ée
+- ban
+- lu
+- gan
+- ▁pale
+- ber
+- ft
+- ▁madame
+- ▁experience
+- мі
+- ▁books
+- ▁easily
+- ▁나
+- ▁seu
+- 를
+- rs
+- gel
+- ▁main
+- ▁Sa
+- ▁giving
+- ef
+- ▁wenn
+- és
+- ▁jack
+- ▁cy
+- ▁entirely
+- ▁minute
+- ▁american
+- ▁usual
+- ▁umu
+- ▁drink
+- ▁george
+- ▁kan
+- ▁hill
+- ▁corner
+- her
+- ли
+- ▁pain
+- pi
+- fer
+- ▁من
+- up
+- ▁gi
+- ▁letters
+- ▁Ni
+- ▁aunt
+- ▁middle
+- ▁lives
+- ▁Ne
+- ।
+- ▁worth
+- ▁expected
+- ▁speaking
+- ▁tone
+- ▁twelve
+- ▁h
+- ▁nun
+- nya
+- car
+- ▁chapter
+- ▁waren
+- ▁talking
+- sten
+- ▁covered
+- ▁unless
+- ▁inter
+- lan
+- ▁nie
+- ▁train
+- ▁fit
+- ▁occasion
+- ▁respect
+- ▁afterwards
+- ira
+- ▁couldn
+- ▁finally
+- ▁ses
+- ▁elle
+- ò
+- ▁warm
+- ende
+- ▁sub
+- ▁trust
+- ▁escape
+- ▁spot
+- ▁cry
+- ;
+- ▁grave
+- ▁lot
+- ▁bei
+- ▁naar
+- ▁lie
+- ▁game
+- ▁success
+- в
+- ▁그냥
+- ure
+- ▁dream
+- ▁turning
+- ▁bank
+- ▁reply
+- ▁instant
+- ▁jesus
+- ji
+- ▁به
+- ▁command
+- ▁K
+- ▁influence
+- á
+- ▁sign
+- ▁island
+- ▁Але
+- ▁appear
+- ▁cette
+- но
+- ▁einem
+- ▁ihn
+- bar
+- nu
+- ▁내가
+- ▁forget
+- ▁places
+- ning
+- ▁persons
+- ▁mere
+- ▁learn
+- ous
+- ▁soft
+- ▁running
+- ▁Ĝi
+- ▁remain
+- fu
+- kin
+- ▁field
+- mar
+- ▁charge
+- ▁silver
+- о
+- ▁rock
+- 만
+- ▁wouldn
+- ▁한
+- ▁killed
+- house
+- ria
+- ▁forest
+- ari
+- ▁surprise
+- ▁Ad
+- ▁simply
+- ▁begin
+- ▁marriage
+- ▁march
+- han
+- ▁board
+- cy
+- len
+- ▁news
+- ▁considered
+- ▁touch
+- con
+- ▁pleasant
+- б
+- ary
+- ▁consider
+- ▁scarcely
+- per
+- ▁laugh
+- ▁class
+- ▁period
+- ▁einer
+- ▁shot
+- ▁move
+- ▁wished
+- è
+- ▁ngo
+- ▁reach
+- ▁allowed
+- ndo
+- 다
+- ▁difficult
+- morrow
+- ▁bij
+- ▁peter
+- ció
+- ▁generally
+- lich
+- ▁space
+- ▁mark
+- ▁closed
+- ▁Vi
+- ▁alle
+- ▁safe
+- ian
+- ▁Ka
+- ▁lead
+- ▁private
+- ▁race
+- ville
+- ▁dress
+- ▁former
+- my
+- ▁Wa
+- ▁breath
+- ▁break
+- ▁그런
+- ▁ari
+- ▁beneath
+- ▁picture
+- ▁circumstances
+- ▁uns
+- ▁tea
+- ▁We
+- nd
+- ▁sus
+- cha
+- ▁iron
+- ▁glass
+- ▁noble
+- mal
+- ▁observed
+- ▁Do
+- ▁ta
+- ▁Der
+- like
+- ▁piece
+- ▁somewhat
+- té
+- ▁B
+- ▁scene
+- ▁paid
+- ▁bound
+- ▁system
+- ▁changed
+- ▁discovered
+- ▁darkness
+- ا
+- if
+- ▁quick
+- dan
+- ق
+- night
+- ▁pot
+- ▁judge
+- ▁enter
+- ▁Com
+- ▁doch
+- ▁się
+- ▁divine
+- ▁tra
+- ам
+- ▁personal
+- don
+- ▁special
+- ▁neck
+- ▁tall
+- ▁sont
+- lin
+- ▁mountain
+- ▁parts
+- ▁united
+- ▁ze
+- ▁Un
+- ▁vint
+- ▁religion
+- ▁guard
+- 기
+- ▁пры
+- ▁paris
+- ▁greatest
+- ▁isn
+- ▁goes
+- ▁ladies
+- ▁fifteen
+- ▁determined
+- ▁believed
+- ▁thank
+- ▁walls
+- ▁fortune
+- ▁mis
+- ▁sharp
+- ры
+- ▁vast
+- ida
+- ojn
+- ▁wonderful
+- ▁built
+- ▁jo
+- ▁seine
+- ▁nous
+- ▁ex
+- ся
+- ▁más
+- ▁ب
+- ▁broad
+- ▁justice
+- ▁rain
+- ▁nog
+- ies
+- ▁hath
+- ford
+- ▁start
+- ▁kun
+- ▁Sie
+- ▁perfectly
+- wood
+- ▁decided
+- ker
+- ive
+- ▁wel
+- ▁gij
+- ▁presently
+- ▁ancient
+- ك
+- ▁modern
+- ▁sixty
+- ▁surely
+- ▁colonel
+- ▁yn
+- ▁declared
+- ▁steps
+- ▁flowers
+- sha
+- це
+- ▁ob
+- ɛ
+- ▁east
+- ▁working
+- ▁habit
+- ▁likely
+- ▁watched
+- ‘
+- sen
+- ron
+- ▁wieder
+- tic
+- ▁sorry
+- ▁exactly
+- ▁진짜
+- ▁crowd
+- ▁bw
+- ▁understood
+- ▁seat
+- ▁zou
+- ▁This
+- ▁ru
+- fi
+- port
+- ▁thick
+- ▁lying
+- ▁threw
+- ▁station
+- ▁social
+- gar
+- ▁deux
+- ier
+- ▁smiled
+- ▁happiness
+- ▁value
+- mente
+- ▁soldiers
+- ▁На
+- bre
+- ▁dels
+- una
+- ▁worse
+- ▁language
+- ▁yellow
+- op
+- ▁cyane
+- ▁як
+- чы
+- tor
+- ▁sake
+- ▁situation
+- ع
+- ▁study
+- ▁offered
+- ▁drawing
+- ▁moon
+- ama
+- ▁monsieur
+- ▁attempt
+- ве
+- ▁fond
+- bel
+- ▁shook
+- ▁gran
+- ▁Si
+- ▁cat
+- ▁spent
+- ▁clothes
+- ▁shore
+- ▁pleased
+- ▁passing
+- ando
+- tra
+- ▁henry
+- ix
+- ▁ears
+- sta
+- ▁honor
+- ▁narrow
+- red
+- ▁zum
+- ▁proper
+- ▁sobre
+- ía
+- ▁яго
+- ▁Va
+- ▁fixed
+- ▁haben
+- ▁expect
+- ▁formed
+- ▁meeting
+- ▁animal
+- ▁existence
+- ite
+- ▁promise
+- ▁wine
+- kan
+- ol
+- ▁leaves
+- шы
+- ем
+- ▁tout
+- ▁honour
+- ▁nice
+- ▁était
+- ante
+- де
+- eg
+- ▁P
+- 야
+- ▁notice
+- ▁spite
+- ▁fallen
+- but
+- ▁mad
+- ▁wise
+- ▁receive
+- ▁royal
+- го
+- ▁G
+- ک
+- ▁evidence
+- ло
+- ▁memory
+- ▁dropped
+- ▁è
+- ▁fi
+- field
+- ▁cent
+- ▁ourselves
+- min
+- dy
+- ju
+- ▁passion
+- ach
+- ▁mijn
+- ▁bis
+- ▁bird
+- ▁york
+- ▁Ver
+- ▁supposed
+- ▁christian
+- ▁Das
+- ▁sick
+- ▁dos
+- ▁prepared
+- ▁golden
+- ▁rise
+- ▁duke
+- ▁repeated
+- ▁named
+- ▁box
+- tte
+- ▁그래서
+- ▁She
+- ali
+- pen
+- ▁vain
+- ▁german
+- tres
+- ▁다
+- ▁drawn
+- ▁shut
+- ▁works
+- ▁speech
+- ▁finished
+- ▁dry
+- ▁creature
+- ▁building
+- of
+- ▁square
+- ▁castle
+- ▁pe
+- ▁dollars
+- ▁birds
+- ▁count
+- ▁complete
+- line
+- ▁pocket
+- ons
+- ▁listen
+- cho
+- way
+- ▁song
+- ą
+- ▁knowing
+- igen
+- ▁passage
+- ▁kill
+- ▁movement
+- ▁bill
+- 是
+- ▁difficulty
+- ▁sad
+- ▁fate
+- ▁daar
+- ex
+- ▁Els
+- ▁major
+- ان
+- ▁ubu
+- ▁color
+- ▁lower
+- ▁lines
+- ▁courage
+- ▁spoken
+- ▁officers
+- ▁houses
+- 대
+- ▁played
+- ▁opportunity
+- ɣ
+- ah
+- nes
+- ▁deg
+- dor
+- yi
+- ▁writing
+- ▁Bu
+- ▁gray
+- ▁principal
+- ▁president
+- ▁allow
+- ▁month
+- ம்
+- ▁hu
+- ▁paul
+- ag
+- ▁officer
+- one
+- ▁mountains
+- ere
+- ▁weeks
+- ▁servant
+- ба
+- ▁nga
+- ▁bread
+- cia
+- ▁ear
+- ▁princess
+- ▁effort
+- ▁ur
+- ▁note
+- ▁looks
+- mon
+- ▁cost
+- ▁이제
+- ige
+- ▁band
+- ▁lake
+- پ
+- ·
+- ня
+- ▁attack
+- ▁property
+- ▁mass
+- ▁questions
+- ▁comfort
+- ▁guess
+- ▁stranger
+- ▁moral
+- ▁F
+- ▁glance
+- ▁equal
+- ▁inside
+- ▁reading
+- ▁marry
+- ▁tan
+- ▁health
+- ▁ay
+- ▁grass
+- ç
+- ▁higher
+- can
+- ê
+- ▁fancy
+- ▁pray
+- ▁С
+- ▁capital
+- ▁win
+- ḥ
+- ▁proved
+- ন
+- ▁oder
+- eur
+- ▁را
+- heid
+- му
+- ▁anxious
+- mos
+- ▁roman
+- ▁surface
+- ement
+- ▁serious
+- ▁holding
+- ▁glory
+- ▁casa
+- ▁ti
+- ▁animals
+- ▁pen
+- ▁tired
+- ano
+- ▁sword
+- ▁original
+- ▁altogether
+- ▁path
+- ▁pride
+- sche
+- ▁avait
+- ро
+- ities
+- nce
+- ▁thin
+- ▁Mo
+- tin
+- ▁carefully
+- ▁usually
+- ▁hurt
+- ▁ordered
+- ▁hung
+- ▁mehr
+- shi
+- ▁blow
+- ▁nobody
+- ▁ri
+- ▁weather
+- ▁Ta
+- ▁அ
+- ou
+- ▁pure
+- ▁meaning
+- ▁pan
+- ▁waited
+- ▁beat
+- ▁Ki
+- ▁Elle
+- ▁ibi
+- ▁coast
+- ▁national
+- ▁example
+- ▁indian
+- ▁gentlemen
+- ▁hotel
+- ▁opposite
+- ▁quietly
+- ▁Sant
+- ▁spread
+- ▁caused
+- ▁worked
+- ▁prove
+- ▁final
+- ▁search
+- ▁Aba
+- ▁remembered
+- ▁dare
+- ныя
+- ▁cases
+- ▁produced
+- tter
+- ▁pounds
+- ▁conduct
+- art
+- ▁gate
+- 로
+- ▁promised
+- ▁fish
+- ud
+- ▁kann
+- ▁storm
+- ▁rule
+- ▁edge
+- ▁breast
+- king
+- ▁weak
+- ava
+- ▁valley
+- ▁ankaŭ
+- ▁centre
+- ▁talked
+- ▁lose
+- ▁possession
+- ▁terror
+- ▁proud
+- ▁kuri
+- tro
+- ▁laws
+- ف
+- ▁در
+- ire
+- ▁Hi
+- ▁grow
+- ▁fingers
+- ▁Co
+- lie
+- ▁ben
+- ▁shoulder
+- ▁dick
+- ▁Te
+- ated
+- ▁curious
+- ę
+- ▁forgotten
+- tan
+- ▁ju
+- ▁difference
+- ▁calm
+- ▁charles
+- ▁support
+- র
+- ▁popular
+- ú
+- ах
+- tel
+- ме
+- ▁seventy
+- ▁breakfast
+- ▁quarter
+- ▁rode
+- ▁drive
+- ▁brave
+- ▁matters
+- ▁companion
+- for
+- ▁orders
+- ▁So
+- ▁alles
+- ▁degree
+- ▁seventeen
+- ▁bottom
+- ▁engaged
+- om
+- let
+- ▁offer
+- ▁serve
+- ▁forced
+- ▁clean
+- 나
+- ▁shadow
+- ▁control
+- ▁america
+- ▁ordinary
+- ▁group
+- ▁evidently
+- cer
+- ▁To
+- ▁gar
+- ▁william
+- uka
+- ▁col
+- ▁police
+- ▁draw
+- ▁diese
+- ▁Ge
+- ▁dann
+- ▁sand
+- ▁everybody
+- ▁palace
+- ries
+- ▁local
+- ▁progress
+- nen
+- dos
+- S
+- ▁Su
+- ▁bent
+- well
+- ▁served
+- ▁“
+- ти
+- ҡ
+- ▁angry
+- ise
+- ▁opening
+- ▁ook
+- ł
+- ▁stream
+- ▁trade
+- ä
+- ▁cousin
+- ▁political
+- ▁surprised
+- ▁Ku
+- ▁watching
+- ▁kingdom
+- се
+- ▁я
+- ▁europe
+- ▁fool
+- ▁material
+- ▁pu
+- nou
+- ха
+- ile
+- ▁greatly
+- iz
+- ▁waar
+- ▁star
+- nde
+- ▁authority
+- chi
+- ▁mas
+- ▁coat
+- ido
+- ▁kwi
+- ▁bay
+- ▁seva
+- nia
+- ḍ
+- ▁sing
+- ни
+- ▁ball
+- ▁gentle
+- ند
+- ging
+- ту
+- ▁individual
+- ▁thrown
+- ence
+- lle
+- ▁confidence
+- ▁seek
+- ▁sagte
+- ▁information
+- ot
+- ▁Mar
+- ▁walking
+- ▁They
+- ▁desert
+- vis
+- ▁Me
+- ▁ways
+- ▁settled
+- ▁rising
+- ▁prayer
+- ▁거야
+- ▁finding
+- ▁weight
+- 는데
+- ▁shoulders
+- res
+- ▁servants
+- ize
+- ▁gun
+- win
+- ▁با
+- ▁season
+- ▁direct
+- ▁nti
+- val
+- dem
+- ▁shown
+- nda
+- ▁fashion
+- side
+- ▁ride
+- ▁Ra
+- ▁base
+- ▁prison
+- ▁bell
+- ▁noticed
+- ▁similar
+- ▁empty
+- ▁familiar
+- за
+- ▁touched
+- ▁fort
+- bla
+- ▁doesn
+- ern
+- ▁loss
+- hu
+- ▁frank
+- ▁ideas
+- ▁legs
+- ▁sought
+- ▁instance
+- ▁از
+- ▁heads
+- ▁kuba
+- ▁dir
+- ▁port
+- ▁sufficient
+- ▁happen
+- nni
+- ació
+- ▁fait
+- ▁Гэта
+- sse
+- kel
+- hen
+- ▁explained
+- ischen
+- ▁delight
+- น
+- ▁ice
+- ▁smoke
+- ▁woods
+- ▁dieser
+- ▁schon
+- ▁ён
+- ▁interesting
+- ▁taste
+- ▁content
+- los
+- ▁couple
+- ▁nation
+- ▁affairs
+- ▁judgment
+- tar
+- ▁british
+- 해
+- ▁mag
+- ▁catch
+- ▁min
+- ▁pli
+- ▁failed
+- ▁burst
+- ▁wij
+- ▁amount
+- ▁false
+- ▁james
+- ▁rate
+- ▁naturally
+- ▁nta
+- ▁immer
+- ь
+- ▁actually
+- ▁members
+- ▁blind
+- ▁alive
+- ▁telling
+- 면
+- ▁explain
+- ▁minister
+- ▁whispered
+- ส
+- ▁fat
+- ▁professor
+- tre
+- ges
+- ▁feelings
+- ▁slow
+- ▁aside
+- fo
+- ▁lu
+- vin
+- ▁denn
+- ▁раз
+- то
+- ▁pity
+- ▁ou
+- ▁És
+- ▁growing
+- ▁было
+- ▁prevent
+- ▁buy
+- ▁loud
+- ▁streets
+- ▁refused
+- ▁share
+- ▁nose
+- ▁task
+- ▁century
+- ▁otherwise
+- ▁agreed
+- ally
+- ▁baby
+- ns
+- ▁chi
+- ▁pero
+- ▁religious
+- ▁obliged
+- ย
+- ale
+- ▁considerable
+- ▁tender
+- ▁add
+- ▁Di
+- ▁intended
+- ▁names
+- ▁grey
+- ▁ces
+- ▁indians
+- ero
+- af
+- ship
+- ▁advantage
+- ff
+- ey
+- ▁imagine
+- ▁murder
+- cher
+- ▁pointed
+- ▁bow
+- ▁mighty
+- ▁excellent
+- ▁troops
+- ▁twice
+- 리
+- ▁keeping
+- cht
+- ▁aniran
+- ▁park
+- ▁Ro
+- ▁gives
+- ▁priest
+- ▁carrer
+- ▁Tu
+- ▁bon
+- ▁apparently
+- ▁distant
+- ▁són
+- lla
+- ▁그거
+- ▁Je
+- ▁required
+- ▁stock
+- ▁windows
+- itat
+- ▁af
+- ▁bag
+- خ
+- ▁게
+- rwa
+- ▁brain
+- ▁تو
+- ▁brief
+- ▁gathered
+- use
+- ▁Pro
+- ▁eene
+- ▁rome
+- ▁때
+- ▁noise
+- ▁club
+- ▁report
+- ance
+- ▁좀
+- ▁bat
+- jn
+- ▁suggested
+- 요
+- ▁dressed
+- ▁importance
+- ▁devil
+- ▁salt
+- E
+- ▁grant
+- ▁superior
+- ▁saint
+- ▁است
+- ▁grown
+- ▁crime
+- са
+- ▁hills
+- ▁для
+- ▁forms
+- து
+- ко
+- ▁san
+- ß
+- ▁moving
+- ré
+- ▁sam
+- fen
+- те
+- )
+- rie
+- ▁labor
+- ▁leading
+- ▁needed
+- ward
+- ▁lies
+- ▁cover
+- ▁rough
+- ▁terms
+- ▁della
+- ▁science
+- ▁flesh
+- ▁flat
+- ▁Gu
+- ▁pieces
+- ▁accept
+- iki
+- ▁haven
+- ▁military
+- amu
+- ▁sorrow
+- ▁stage
+- ▁Bo
+- ▁sixteen
+- ң
+- ▁tongue
+- ▁sides
+- via
+- ▁dust
+- su
+- vel
+- ▁events
+- ▁remarked
+- ▁hearts
+- ▁mentioned
+- ▁habe
+- ▁firm
+- ▁conditions
+- ▁double
+- ▁liked
+- ▁eighty
+- ▁saved
+- I
+- ▁honest
+- ▁daily
+- ▁gut
+- ▁아니
+- ding
+- ну
+- ac
+- ▁Ri
+- ▁aid
+- ві
+- ris
+- ▁ships
+- ▁hebben
+- ▁union
+- ▁stars
+- nk
+- ▁temple
+- ▁seized
+- ▁handsome
+- цца
+- ▁ci
+- ż
+- ▁succeeded
+- van
+- ▁мы
+- ▁molt
+- ▁faint
+- ▁carriage
+- ▁liberty
+- ▁taught
+- ду
+- ▁bridge
+- ▁ring
+- ▁fourth
+- ▁listened
+- ▁wore
+- ре
+- ▁contrary
+- ர
+- ats
+- сы
+- ika
+- mbi
+- ▁soldier
+- ably
+- ▁powerful
+- ▁aware
+- ü
+- ال
+- ▁pair
+- ▁corn
+- ▁dangerous
+- ▁zur
+- ▁apart
+- ▁cha
+- ▁hearing
+- ▁faces
+- ▁papers
+- ▁philip
+- ▁throw
+- ata
+- ▁directly
+- ▁smiling
+- ité
+- ▁unknown
+- bert
+- ә
+- ▁style
+- ▁수
+- ▁sans
+- ▁level
+- ▁Da
+- ok
+- ▁david
+- time
+- ▁size
+- ▁fly
+- ▁mistress
+- ▁possibly
+- ша
+- لا
+- tas
+- ating
+- ай
+- ▁bitter
+- ▁나는
+- ▁sunday
+- ▁busy
+- ▁cruel
+- ▁lifted
+- ▁که
+- ien
+- ▁cal
+- ▁supper
+- ▁heat
+- ▁putting
+- ▁Pa
+- ▁sum
+- ▁worthy
+- aient
+- ▁slight
+- ▁Po
+- ▁cya
+- ▁huge
+- ▁affair
+- ▁Bi
+- ▁fill
+- ▁cook
+- 잖아
+- ungen
+- iza
+- ▁aux
+- ▁particularly
+- cle
+- ▁nevertheless
+- ▁nk
+- ▁governor
+- ▁enemies
+- ▁fully
+- ió
+- ▁teeth
+- ных
+- оў
+- ▁satisfied
+- wer
+- ▁(
+- ▁seiner
+- ▁demanded
+- ▁yi
+- ▁emperor
+- ć
+- zer
+- ым
+- ▁joined
+- ▁lovely
+- ▁این
+- mes
+- ▁fruit
+- ▁geen
+- ▁completely
+- ▁occupied
+- ▁entire
+- ▁virtue
+- ned
+- ista
+- are
+- cio
+- pre
+- ен
+- bye
+- ▁W
+- ▁favor
+- ▁ihre
+- nye
+- 한
+- ▁practice
+- ▁farther
+- ▁zi
+- sel
+- ▁flight
+- ▁takes
+- ción
+- ▁bore
+- ▁nom
+- ▁Ya
+- ▁impression
+- лы
+- nis
+- ▁pardon
+- stand
+- ▁shape
+- ▁points
+- ▁measure
+- же
+- 에서
+- ab
+- ▁har
+- ▁education
+- ▁wealth
+- ▁occurred
+- ▁interested
+- ▁carrying
+- ▁playing
+- ▁regarded
+- ros
+- ala
+- ▁price
+- ▁laughing
+- ▁foreign
+- ła
+- mba
+- най
+- ▁finger
+- nga
+- ▁wasn
+- ▁fan
+- ub
+- mann
+- ▁suffered
+- ▁bare
+- ▁esta
+- ▁ariko
+- ▁wy
+- ▁будзе
+- ich
+- ves
+- ▁ба
+- las
+- ▁kitchen
+- ▁drove
+- ▁famous
+- ▁presented
+- ▁crossed
+- ras
+- ▁uh
+- law
+- வ
+- ps
+- gira
+- ins
+- ▁Con
+- ▁també
+- ▁Ru
+- ▁powers
+- par
+- ▁ain
+- 게
+- mmer
+- ▁There
+- ▁spirits
+- ▁frequently
+- illa
+- ▁regular
+- ▁excitement
+- ▁mercy
+- دا
+- ▁yesterday
+- ▁colour
+- ▁suffer
+- gera
+- ▁risk
+- ▁civil
+- र
+- eb
+- ▁peculiar
+- ▁willing
+- の
+- iĝis
+- ndi
+- rin
+- ▁hast
+- ▁robert
+- vor
+- ▁З
+- ▁dull
+- ▁잘
+- ▁nearer
+- ▁shop
+- tz
+- ▁native
+- ▁hay
+- ther
+- ▁species
+- ▁bir
+- ini
+- ▁calling
+- ▁так
+- ▁relief
+- umu
+- ▁deck
+- ▁marked
+- ▁wounded
+- ▁été
+- ola
+- ▁pulled
+- xi
+- ▁beg
+- ▁roof
+- ▁spiritual
+- ▁asleep
+- ▁gegen
+- ত
+- ▁smith
+- ▁absolutely
+- ▁behold
+- ▁parents
+- ▁gain
+- ▁speed
+- ▁necessity
+- ▁milk
+- ▁rapidly
+- ▁eu
+- ▁zal
+- ito
+- ▁iran
+- ▁falling
+- ▁tri
+- ▁job
+- жа
+- ▁ahead
+- oli
+- 人
+- ▁drop
+- ▁possessed
+- ▁ré
+- ula
+- ▁mil
+- ▁cre
+- ▁sheep
+- ato
+- ▁wear
+- ▁너무
+- ney
+- ▁deze
+- ▁desired
+- ▁freedom
+- ament
+- ▁bodies
+- tura
+- ▁highest
+- ม
+- ▁meer
+- ▁truly
+- ▁ceased
+- ▁dogs
+- ▁clearly
+- ▁dance
+- ▁slave
+- ▁facts
+- ▁described
+- ▁imagination
+- ▁hidden
+- ▁crown
+- ▁gods
+- ▁edward
+- ow
+- ▁claim
+- сці
+- ▁appears
+- ani
+- lon
+- out
+- ▁suffering
+- ▁throughout
+- ▁join
+- table
+- ▁fighting
+- tta
+- ▁remains
+- ▁Па
+- ▁anger
+- ona
+- ▁sold
+- ▁worden
+- ▁wisdom
+- ▁upper
+- ▁dozen
+- 스
+- ய
+- ▁selbst
+- sti
+- ▁obtained
+- ▁favour
+- ní
+- ib
+- ▁height
+- ▁trail
+- ▁ж
+- ▁driven
+- ▁flower
+- ▁hoe
+- อ
+- ▁author
+- ▁demand
+- ▁harm
+- ▁worst
+- ▁abantu
+- ▁struggle
+- work
+- ▁mention
+- ▁knight
+- ▁prisoner
+- ак
+- ▁papa
+- ▁anybody
+- ▁Sch
+- ▁advanced
+- ▁problem
+- ▁В
+- ▁entrance
+- ▁moments
+- ▁seated
+- ▁admit
+- ▁sehr
+- ▁burning
+- ▁process
+- ▁пад
+- ▁satisfaction
+- she
+- ▁midst
+- ▁Ab
+- ▁fault
+- ▁shame
+- ▁Du
+- า
+- ug
+- ▁loose
+- 、
+- ▁suit
+- ▁R
+- ▁circle
+- ▁cloud
+- ▁لە
+- nte
+- 有
+- ▁angel
+- ▁triumph
+- ▁não
+- ▁enjoy
+- mbe
+- ▁instantly
+- ▁kiel
+- lia
+- ▁press
+- ▁ئا
+- ▁harry
+- ▁cool
+- iko
+- sco
+- 一
+- heit
+- ▁provided
+- ▁store
+- ▁frightened
+- qui
+- das
+- ▁buried
+- ▁nineteen
+- ▁sei
+- î
+- zwa
+- ▁fun
+- mas
+- ▁Ar
+- ▁expressed
+- ▁welcome
+- ora
+- ▁rocks
+- ▁accepted
+- mbo
+- ▁excuse
+- ▁hun
+- ▁advice
+- ▁같애
+- ▁pi
+- ▁andere
+- ▁аб
+- ▁rank
+- ▁hate
+- ру
+- ▁іх
+- ▁proceeded
+- ▁birth
+- ▁principle
+- ized
+- schen
+- č
+- ▁stones
+- ▁dia
+- ү
+- ц
+- blo
+- five
+- ue
+- ▁motion
+- ▁zwei
+- ▁majesty
+- ▁eleven
+- ▁portion
+- ▁stories
+- ▁equally
+- ▁removed
+- ▁onder
+- ▁throat
+- š
+- ▁hoped
+- ▁knees
+- ă
+- ella
+- ▁ninety
+- ▁wants
+- ▁sons
+- 시
+- ▁leur
+- ▁key
+- ▁sprang
+- ▁jane
+- uko
+- ▁belief
+- ▁ocean
+- mwe
+- ▁council
+- ▁You
+- ▁awful
+- ▁heb
+- ▁rare
+- ▁paused
+- ▁lad
+- iri
+- ▁grief
+- ▁많이
+- ▁pel
+- ▁mile
+- ▁cup
+- ▁cabin
+- ▁hence
+- ▁skin
+- ▁type
+- ▁host
+- дзе
+- ▁careful
+- ▁message
+- ▁What
+- ▁Ho
+- ▁rec
+- ▁admitted
+- ▁са
+- ▁해
+- ர்
+- ang
+- ▁established
+- ▁mai
+- était
+- ▁fields
+- ф
+- ▁iki
+- ▁unable
+- ри
+- ح
+- ▁somewhere
+- ▁fin
+- ▁arrival
+- ▁gradually
+- ▁worship
+- ky
+- ino
+- ▁exist
+- ▁kam
+- ▁yari
+- int
+- ▁market
+- isi
+- ▁cro
+- ▁extraordinary
+- all
+- ▁begun
+- ati
+- ius
+- ▁mor
+- ▁lamp
+- во
+- 在
+- ▁hurried
+- ▁disappeared
+- 인
+- ged
+- ▁leg
+- ▁dying
+- ▁advance
+- tie
+- ▁larger
+- ▁forces
+- ▁mental
+- ▁choose
+- ▁continue
+- lichen
+- dia
+- ▁Fa
+- ▁friendship
+- ▁waters
+- ego
+- ▁blessed
+- ▁countenance
+- ▁بە
+- oni
+- ▁faithful
+- ▁remarkable
+- idad
+- ▁ama
+- ▁mistake
+- ▁hare
+- ▁tres
+- ล
+- ге
+- ▁intention
+- ient
+- ş
+- ▁б
+- ▁central
+- ▁solid
+- ▁plenty
+- pro
+- ▁vessel
+- ▁origin
+- ▁stick
+- ▁resolved
+- ▁oil
+- ▁signal
+- ▁quatre
+- ▁chamber
+- ▁gaze
+- ▁features
+- ▁sympathy
+- ▁fut
+- place
+- ▁evident
+- ▁active
+- க்
+- sis
+- eren
+- ▁dreadful
+- ▁member
+- мы
+- ▁concerning
+- ▁нас
+- ▁farm
+- ▁meine
+- ap
+- ▁actual
+- غ
+- ▁method
+- ▁splendid
+- ▁jim
+- qu
+- ▁innocent
+- ▁absence
+- hin
+- ils
+- ▁dawn
+- ▁heeft
+- ці
+- ło
+- ▁design
+- 라
+- ▁mein
+- ব
+- ▁cyangwa
+- ▁gently
+- ▁express
+- mont
+- ▁understanding
+- 하고
+- ▁sell
+- sto
+- ہ
+- ▁sol
+- цы
+- ▁everywhere
+- ▁Ye
+- bed
+- ▁hit
+- ด
+- med
+- ▁horror
+- osa
+- ▁conscious
+- ▁kindly
+- па
+- ▁louis
+- ▁hole
+- ▁Im
+- вы
+- ▁increased
+- ▁kra
+- ▁produce
+- ▁record
+- tive
+- ▁Y
+- ▁zeide
+- ным
+- ▁mystery
+- ста
+- tal
+- ▁sang
+- ▁despair
+- ▁rooms
+- ▁conscience
+- kwa
+- ▁affection
+- ет
+- ments
+- гу
+- bri
+- ▁sacred
+- dra
+- ▁richard
+- ▁temps
+- ische
+- mis
+- ▁ended
+- ▁ging
+- ▁bearing
+- dd
+- ▁mat
+- nic
+- ▁dim
+- ▁bought
+- ▁stairs
+- ▁neza
+- ques
+- ▁supply
+- ди
+- pin
+- ph
+- ▁cor
+- dre
+- ▁stepped
+- ก
+- ▁ruin
+- ▁strike
+- ▁match
+- ▁ibyo
+- ▁His
+- ▁closely
+- ▁ĝi
+- ▁약간
+- ▁pressed
+- ▁mort
+- ▁tiu
+- N
+- cent
+- ▁picked
+- ▁yeah
+- ▁concerned
+- ▁torn
+- ▁maybe
+- ▁joe
+- ▁super
+- izi
+- ▁list
+- ▁moreover
+- ▁choice
+- ibi
+- ▁film
+- ▁больш
+- ▁또
+- ▁gli
+- ▁இ
+- ப்
+- ▁employed
+- ▁rushed
+- tis
+- hn
+- ▁constant
+- kla
+- ▁jetzt
+- ▁до
+- ▁tent
+- ▁wondered
+- ▁guns
+- ▁asking
+- ▁college
+- ▁К
+- ▁ganz
+- ▁wings
+- ▁kuko
+- ▁visible
+- ▁extent
+- ▁gas
+- imi
+- ▁excited
+- ▁vision
+- B
+- rem
+- ▁energy
+- ette
+- ▁eager
+- ▁tis
+- ▁fra
+- ▁imp
+- anar
+- ▁dar
+- ▁fu
+- ▁term
+- ▁clouds
+- ▁efforts
+- ▁este
+- ▁patient
+- ▁escaped
+- ▁chose
+- ▁yards
+- че
+- ▁bal
+- set
+- ▁sooner
+- ▁weary
+- stone
+- ها
+- eux
+- ▁trial
+- ▁où
+- ▁meat
+- ▁softly
+- aga
+- stra
+- ▁우리
+- ын
+- ▁spend
+- ▁fue
+- ▁wohl
+- ▁él
+- ▁elizabeth
+- ที่
+- eau
+- ▁пера
+- ▁whenever
+- cie
+- mond
+- ε
+- ▁doors
+- ▁singing
+- ▁deeply
+- ▁physical
+- ки
+- ▁fail
+- ▁то
+- ▁sed
+- ▁nay
+- ▁capable
+- ▁temper
+- ▁friendly
+- ▁wake
+- ▁artist
+- ▁On
+- ▁fought
+- ە
+- ▁shouted
+- ▁helped
+- ▁safety
+- ▁zijne
+- ▁aussi
+- T
+- ▁fairly
+- ▁poet
+- san
+- лан
+- ▁treated
+- ▁bishop
+- ▁Мы
+- ▁ease
+- жы
+- ▁brothers
+- ▁relations
+- O
+- نا
+- ▁plej
+- ▁uko
+- ная
+- ▁obtain
+- ▁attitude
+- ▁gained
+- gue
+- met
+- rio
+- ▁wicked
+- ▁zei
+- บ
+- ▁bin
+- cal
+- ▁chosen
+- ше
+- ▁tal
+- ▁rw
+- ▁lover
+- ▁inquired
+- ▁jest
+- ▁han
+- рэ
+- ▁está
+- ▁Ga
+- rat
+- nos
+- 네
+- ▁discover
+- ▁maid
+- pt
+- ▁вельмі
+- ▁genius
+- 'off'
+- ▁sta
+- ▁southern
+- rick
+- ▁grande
+- ▁nations
+- ▁As
+- dis
+- ▁precious
+- ric
+- ▁useful
+- ง
+- ▁exact
+- ▁flying
+- ▁theory
+- ▁machine
+- ▁worn
+- ক
+- ▁arthur
+- ▁throne
+- ▁voyage
+- ன்
+- ▁becomes
+- uch
+- ▁consequence
+- 데
+- biri
+- ▁V
+- né
+- ▁hell
+- ▁fierce
+- 大
+- ▁becoming
+- sion
+- ▁trip
+- ▁había
+- ▁connection
+- ▁voices
+- qua
+- ▁seinen
+- iyi
+- ▁approached
+- ▁tower
+- ▁Lo
+- ▁extreme
+- stro
+- ssa
+- ▁avoid
+- ▁development
+- gwa
+- ▁tie
+- ға
+- 为
+- ин
+- lang
+- ▁violent
+- ▁intelligence
+- ▁zo
+- ▁figures
+- ▁fleet
+- த்த
+- ▁proof
+- tum
+- ▁approach
+- ▁못
+- ▁range
+- ▁attend
+- ▁Ein
+- ▁pause
+- ces
+- ▁viel
+- мо
+- ▁surrounded
+- ec
+- ▁И
+- ▁fo
+- ▁relation
+- ▁earl
+- ties
+- ▁delicate
+- ▁numerous
+- ▁même
+- ▁series
+- ▁moi
+- ▁witness
+- del
+- ▁victory
+- ▁consideration
+- ▁older
+- ▁thoroughly
+- toj
+- ▁minds
+- ▁regret
+- lé
+- ан
+- ṛ
+- ▁Не
+- ▁katika
+- ▁keen
+- ▁cottage
+- ▁raise
+- oc
+- نی
+- ▁miserable
+- ম
+- ▁anne
+- liche
+- ▁somebody
+- ▁secure
+- ▁reg
+- looking
+- nzi
+- ev
+- mel
+- ▁refuse
+- ▁acquaintance
+- ▁Tom
+- ▁region
+- mie
+- ▁accompanied
+- ▁내
+- ▁image
+- rá
+- ▁crew
+- ▁keine
+- نى
+- ▁gay
+- ат
+- tur
+- ▁interrupted
+- ▁толькі
+- ร
+- ▁nurse
+- ists
+- ▁hopes
+- ▁tail
+- ž
+- ▁accident
+- ▁ili
+- ▁ont
+- ▁teach
+- ▁previous
+- wed
+- od
+- ▁somehow
+- ▁plans
+- ▁bold
+- ▁hy
+- ▁Bar
+- ▁slept
+- ▁arose
+- ▁credit
+- ▁Her
+- ▁description
+- ву
+- ▁contra
+- гә
+- ski
+- ▁hide
+- ▁vers
+- ering
+- kon
+- ▁nodded
+- ▁guide
+- ▁kindness
+- ▁آ
+- ▁flame
+- ▁returning
+- ▁humble
+- ▁wound
+- ▁request
+- rà
+- sz
+- ▁iyo
+- ▁flew
+- ▁whilst
+- ▁hurry
+- ▁afford
+- ▁bringing
+- zin
+- ▁shining
+- ▁cattle
+- ▁fe
+- ▁wholly
+- ▁anyone
+- ▁meanwhile
+- ▁widow
+- ▁curiosity
+- ▁appointed
+- ▁anti
+- ▁peu
+- ▁title
+- ν
+- tant
+- hand
+- ▁showing
+- ▁proposed
+- vers
+- ntu
+- ▁Ce
+- ▁mysterious
+- ami
+- ▁assured
+- ▁guilty
+- را
+- த
+- ▁amongst
+- ▁быў
+- ▁row
+- ▁stared
+- ецца
+- iro
+- ▁partly
+- ep
+- ▁trans
+- ▁dignity
+- ▁subjects
+- چ
+- ▁career
+- ▁donc
+- ▁eternal
+- tern
+- ▁pick
+- ▁bid
+- ara
+- ▁branches
+- ubu
+- cur
+- ▁dreams
+- ار
+- ▁companions
+- ▁número
+- бы
+- yer
+- ▁Ko
+- ▁silk
+- ▁fur
+- ▁informed
+- ▁address
+- гі
+- ▁rapid
+- ▁werd
+- ▁parte
+- ▁build
+- ▁pushed
+- ▁successful
+- sia
+- ▁lang
+- ▁bull
+- mut
+- ▁plant
+- ▁lest
+- 가지고
+- dro
+- ▁stands
+- ▁sacrifice
+- ▁fled
+- wy
+- ▁souls
+- ▁Z
+- ▁utterly
+- ▁comfortable
+- ▁custom
+- ▁consent
+- ▁confess
+- ▁Ich
+- bwa
+- ▁wit
+- ▁luck
+- ▁J
+- ▁smooth
+- ▁remark
+- 구
+- ری
+- ▁wolf
+- ▁rear
+- ▁tin
+- ▁discovery
+- als
+- ▁washington
+- ▁younger
+- ▁brilliant
+- ped
+- ▁laughter
+- ▁empire
+- α
+- ▁listening
+- ▁western
+- ▁steel
+- wel
+- ▁yours
+- fall
+- gin
+- eu
+- न
+- ▁needs
+- ▁rang
+- ▁vida
+- ▁todo
+- ін
+- ▁punishment
+- sie
+- ▁су
+- low
+- ▁objects
+- ▁ўсё
+- ▁practical
+- ▁pra
+- ▁lack
+- ẓ
+- ط
+- ▁All
+- ▁말
+- uf
+- ▁thanks
+- ▁observe
+- ▁sounds
+- ▁soil
+- ▁fatal
+- ▁meal
+- ▁fellows
+- ▁margaret
+- 일
+- ҫ
+- ert
+- ▁managed
+- ▁weg
+- ▁reward
+- ▁pla
+- ▁arranged
+- ▁uma
+- ▁Demà
+- ▁hollow
+- berg
+- ▁forgot
+- ▁accustomed
+- ▁nest
+- cra
+- 사
+- ▁rue
+- ▁gift
+- ▁highly
+- fel
+- ▁pin
+- ▁solemn
+- vuga
+- vil
+- ▁Jo
+- ide
+- ▁praise
+- mm
+- vé
+- ations
+- ▁directed
+- mp
+- ▁kiss
+- ▁clever
+- that
+- gli
+- G
+- ▁foolish
+- ▁parties
+- ▁agree
+- ole
+- ngo
+- ▁Ja
+- ial
+- ▁forgive
+- 랑
+- ▁expedition
+- uri
+- ▁anirà
+- ▁hospital
+- ▁sua
+- ▁muy
+- ▁goed
+- ▁creatures
+- ▁ма
+- ந்த
+- ▁swift
+- amente
+- ▁мо
+- nne
+- ▁explanation
+- ▁thomas
+- амі
+- ▁unhappy
+- ▁duties
+- iert
+- ▁event
+- ▁Ca
+- ▁trembling
+- ▁prisoners
+- ▁bob
+- ▁porque
+- ▁christmas
+- ▁immense
+- ▁labour
+- cul
+- ▁devoted
+- ▁glanced
+- ▁tale
+- ▁Man
+- ▁faire
+- zwe
+- igi
+- ▁attended
+- ▁sank
+- ▁lieutenant
+- ▁distinguished
+- encia
+- kiri
+- ▁route
+- cions
+- ▁ما
+- ades
+- ▁northern
+- ▁waves
+- ▁aho
+- ▁constantly
+- ▁응
+- ▁사
+- ▁reality
+- ▁1
+- स
+- ▁grim
+- ▁bas
+- ture
+- ▁severe
+- ▁connected
+- ▁hanging
+- ▁되게
+- ▁immediate
+- ans
+- ▁results
+- ▁sail
+- ▁destroyed
+- ▁reasons
+- ▁anxiety
+- ▁stern
+- тар
+- ыр
+- ▁thousands
+- ▁zurück
+- ▁principles
+- ants
+- ▁article
+- tus
+- ▁divided
+- kt
+- ▁numbers
+- ▁etwas
+- ▁tour
+- ▁Ti
+- ▁hin
+- ▁stuff
+- ical
+- ▁protection
+- ▁nichts
+- eix
+- ▁charm
+- ▁italian
+- ▁suspicion
+- ▁separate
+- ▁track
+- ”
+- ▁addressed
+- до
+- ▁stretched
+- ▁pipe
+- ▁consciousness
+- ▁spanish
+- ks
+- pping
+- ▁이런
+- ▁But
+- ▁gihe
+- ça
+- ▁counsel
+- lijk
+- pu
+- ும்
+- ▁queer
+- ▁delighted
+- ▁recognized
+- ▁murmured
+- ▁fourteen
+- க
+- ute
+- ▁objection
+- ▁bwa
+- yn
+- ▁kissed
+- ▁exercise
+- ▁pictures
+- ▁current
+- ▁signs
+- ov
+- ▁primer
+- dic
+- ▁cheeks
+- wn
+- ▁shows
+- är
+- ei
+- ▁dont
+- ▁test
+- mit
+- ▁vent
+- に
+- ▁rolled
+- ▁ideal
+- ўся
+- ▁cities
+- ون
+- lli
+- lina
+- ▁travel
+- ▁pull
+- aux
+- uma
+- ▁convinced
+- ▁во
+- ▁eh
+- ▁cave
+- lit
+- ▁bru
+- ▁너
+- iw
+- ▁cow
+- ▁«
+- ▁яны
+- ми
+- ▁New
+- ▁inner
+- ▁yard
+- ть
+- ▁о
+- ▁sal
+- тә
+- ▁mia
+- shobora
+- ▁wurden
+- ன
+- ▁assistance
+- ▁follows
+- ▁accordingly
+- ▁També
+- ▁swept
+- ▁fox
+- ام
+- ▁horrible
+- ele
+- ▁confusion
+- ▁odd
+- ▁지금
+- ▁feared
+- ación
+- ▁tiny
+- wu
+- ▁ol
+- ▁reader
+- ▁havas
+- oku
+- ▁martin
+- ▁page
+- ▁learning
+- ▁earnest
+- ▁committed
+- ▁alarm
+- ▁argument
+- ▁announced
+- لی
+- ▁source
+- ▁blame
+- ▁ĉi
+- ▁dich
+- ão
+- ای
+- ▁interests
+- ▁sentence
+- ▁spare
+- ▁od
+- ice
+- ▁lonely
+- ▁patience
+- ▁unu
+- ob
+- ▁ihnen
+- ea
+- ▁joseph
+- ón
+- ▁하
+- 아
+- ▁permit
+- ▁wishes
+- ux
+- ▁increase
+- ▁For
+- ua
+- ▁Am
+- ▁And
+- ▁glorious
+- ▁rush
+- rus
+- ▁universal
+- dig
+- kuru
+- ▁dass
+- ▁setting
+- ▁к
+- ▁secretary
+- ▁ei
+- ▁hang
+- ▁plainly
+- ▁coffee
+- ▁bless
+- ▁dijo
+- rit
+- ке
+- ▁prefer
+- ▁composed
+- ▁quality
+- ▁charming
+- oso
+- ▁treasure
+- ▁delay
+- ▁extremely
+- ▁quarrel
+- ▁그렇게
+- ▁knife
+- ugu
+- iti
+- kora
+- ▁population
+- ▁slightly
+- ▁deed
+- ▁conclusion
+- 들
+- ▁bald
+- ▁drei
+- ▁яе
+- ▁seldom
+- ▁household
+- ▁lit
+- ▁oak
+- ▁skill
+- ▁baron
+- ▁feast
+- ▁hart
+- 주
+- ▁그게
+- ados
+- ▁hungry
+- ées
+- ▁reign
+- ம
+- ▁ohne
+- ▁kings
+- ▁але
+- ▁commanded
+- ▁mamma
+- ▁card
+- ▁leader
+- ▁aim
+- ▁konnte
+- нь
+- ▁element
+- ▁ella
+- ▁igihe
+- stein
+- は
+- hr
+- ▁pur
+- P
+- ais
+- gy
+- bes
+- ▁august
+- ▁doubtless
+- ▁час
+- ▁rid
+- 으로
+- ▁stir
+- ▁그때
+- ▁wil
+- ▁icyo
+- ▁tomb
+- cli
+- сто
+- ▁fix
+- ▁compelled
+- ল
+- ▁burden
+- ô
+- ▁ali
+- ▁perceived
+- 까
+- ▁тут
+- ▁thrust
+- har
+- ▁warning
+- ▁hunting
+- ▁dien
+- 장
+- ▁lights
+- iva
+- ▁slipped
+- ls
+- ар
+- े
+- ▁wi
+- nna
+- ▁county
+- ھ
+- ell
+- ▁vom
+- ▁flag
+- ▁frame
+- ▁details
+- lis
+- ▁interview
+- ▁hut
+- ▁rage
+- ▁encore
+- pan
+- gri
+- ▁cart
+- ▁absolute
+- uk
+- ே
+- ▁ati
+- ayo
+- ▁countries
+- ▁2
+- ▁pod
+- ێ
+- ▁inform
+- dar
+- ram
+- ▁està
+- ▁eagerly
+- ▁sous
+- ▁ashamed
+- ▁können
+- net
+- ▁reform
+- ▁district
+- ▁hari
+- ▁Rwanda
+- ▁nord
+- ▁created
+- ència
+- ▁오
+- ▁striking
+- ▁forehead
+- gra
+- ple
+- ▁peut
+- ▁check
+- ▁fairy
+- ▁correct
+- ions
+- ▁admiration
+- ▁crying
+- ▁alas
+- â
+- ▁ghost
+- ▁riding
+- ▁dire
+- ғы
+- yan
+- uti
+- ▁male
+- ит
+- bb
+- ید
+- pf
+- ▁wet
+- wan
+- ▁savage
+- ique
+- ▁trois
+- ▁Hu
+- ▁cuando
+- ā
+- gon
+- ு
+- ▁inhabitants
+- ▁Ён
+- mia
+- recht
+- ▁lighted
+- ▁rent
+- ▁profit
+- ▁million
+- war
+- og
+- án
+- ▁Aus
+- ▁però
+- ▁chap
+- ▁있어
+- pp
+- مان
+- ejo
+- ▁Cap
+- ▁roll
+- light
+- ssen
+- त
+- ▁revolution
+- ▁pope
+- ▁fest
+- ▁manera
+- ▁contained
+- ▁ere
+- eta
+- ip
+- water
+- ▁shake
+- ▁mbere
+- ▁sigh
+- some
+- ▁cu
+- ▁이렇게
+- ▁occasionally
+- ▁flung
+- wing
+- ili
+- ▁sleeping
+- ère
+- ން
+- endo
+- ▁thunder
+- mb
+- ▁bya
+- ท
+- ▁dared
+- tent
+- gna
+- ▁про
+- ▁visited
+- ▁ignorant
+- دە
+- ▁kugira
+- 不
+- ▁rope
+- geze
+- ▁giant
+- ▁slaves
+- ▁mate
+- ▁smell
+- ▁mud
+- ▁Son
+- stre
+- ▁At
+- ▁useless
+- ▁destroy
+- های
+- ▁ĉe
+- ▁failure
+- ς
+- ▁breaking
+- ▁mortal
+- ▁verse
+- ▁guests
+- ▁яшчэ
+- ▁wave
+- lik
+- ▁uttered
+- ▁pink
+- ▁жа
+- ▁departure
+- hel
+- ▁cloth
+- ▁whence
+- hl
+- ▁pie
+- uda
+- ▁있는
+- aire
+- iu
+- ró
+- ▁addition
+- ▁더
+- ▁utter
+- ï
+- ج
+- ▁extra
+- ▁enormous
+- ▁trick
+- ▁permitted
+- ▁abroad
+- ▁parliament
+- ▁introduced
+- ▁services
+- ▁aka
+- ে
+- tig
+- ▁alors
+- ▁eggs
+- ▁todos
+- ست
+- ▁beast
+- ▁published
+- зе
+- ▁inclined
+- ▁او
+- ▁unfortunate
+- mina
+- ▁gy
+- лю
+- ima
+- ▁shelter
+- ▁gazed
+- nden
+- ▁armed
+- ador
+- ▁distinct
+- ▁helen
+- ▁mission
+- ▁female
+- ▁mister
+- ▁saddle
+- ▁billy
+- bona
+- лу
+- ś
+- ▁butter
+- ▁greek
+- ▁sisters
+- ▁tem
+- ▁weakness
+- over
+- ▁Umu
+- ა
+- ▁diesem
+- ▁stated
+- ▁wondering
+- ু
+- ▁begged
+- ▁degrees
+- ▁ebi
+- ▁pack
+- ▁department
+- ieren
+- far
+- ▁fever
+- ▁comment
+- rate
+- ▁mill
+- ▁그래
+- лася
+- ▁proportion
+- lock
+- ▁literature
+- ▁qué
+- ▁rod
+- more
+- gun
+- ▁tio
+- ▁cur
+- ▁disposition
+- zy
+- pot
+- ì
+- ▁mac
+- ay
+- ▁seinem
+- ▁articles
+- ▁eines
+- ▁volen
+- ▁constitution
+- taj
+- ین
+- ▁Car
+- head
+- lem
+- ▁jam
+- ▁atmosphere
+- ille
+- ving
+- ▁troubled
+- 니까
+- ▁bari
+- ▁vier
+- ▁dwell
+- ▁shade
+- ▁painted
+- ▁со
+- ▁merry
+- ▁menschen
+- ▁cab
+- ▁shock
+- ви
+- ▁க
+- ▁decision
+- ▁toch
+- ▁omu
+- ▁troba
+- die
+- ं
+- ▁remote
+- ▁vie
+- ily
+- ▁tied
+- การ
+- ▁approaching
+- ▁library
+- ▁gal
+- ▁bowed
+- ▁val
+- ▁israel
+- rait
+- ▁하고
+- ▁ji
+- ▁acts
+- ▁sono
+- ▁enjoyed
+- ▁classes
+- ▁destruction
+- ▁sah
+- ▁spain
+- ▁pare
+- bone
+- cre
+- ability
+- ▁lands
+- ▁limit
+- zar
+- ▁res
+- ▁disease
+- ▁valuable
+- ▁mode
+- ▁misery
+- وا
+- ▁uniform
+- nas
+- ▁lion
+- ▁dum
+- ▁hastily
+- ▁Go
+- ▁earlier
+- ston
+- ▁emotion
+- ▁난
+- ▁date
+- wie
+- kira
+- э
+- eh
+- ▁waste
+- ▁teaching
+- ▁weiß
+- ▁mayor
+- ▁할
+- eza
+- ▁burned
+- க்க
+- ▁ry
+- ▁О
+- ▁bosom
+- ▁Ili
+- ▁tegen
+- ▁mounted
+- zel
+- ▁ages
+- ▁shoot
+- ▁italy
+- ▁division
+- ▁rendered
+- ▁causes
+- ▁ال
+- ▁commander
+- ▁banks
+- leg
+- ew
+- ▁pré
+- ▁że
+- ▁instrument
+- ▁using
+- ▁dread
+- ▁farmer
+- প
+- bil
+- ako
+- ▁possess
+- ▁flash
+- ліся
+- ▁deu
+- ▁pet
+- ▁nervous
+- ца
+- ▁aspect
+- ▁retired
+- ▁motor
+- ▁bones
+- сты
+- ▁violence
+- eḍ
+- ▁mount
+- ▁scheme
+- ▁aloud
+- ▁germany
+- ▁affected
+- ▁branch
+- ▁„
+- ▁difficulties
+- var
+- bat
+- ▁vague
+- ▁towns
+- ▁shadows
+- ▁domestic
+- ▁fifth
+- ▁barn
+- 자
+- ▁wooden
+- ule
+- être
+- ici
+- ▁steady
+- ▁statement
+- ▁model
+- ▁plants
+- ▁official
+- nge
+- olo
+- bli
+- ▁sufficiently
+- ▁irà
+- ▁ко
+- ▁jacob
+- ▁bottle
+- ▁interior
+- ▁philosophy
+- isa
+- кә
+- ▁discussion
+- ▁scattered
+- ▁rival
+- ▁mijnheer
+- ▁eating
+- ▁trace
+- ▁hid
+- ா
+- ▁applied
+- ▁belong
+- ▁uw
+- gé
+- ▁appeal
+- ▁trop
+- tsi
+- ▁magic
+- ▁boats
+- ▁notes
+- fl
+- ▁benefit
+- ▁june
+- ▁Yuba
+- ▁nou
+- ▁kleine
+- rig
+- ▁sugar
+- ▁depend
+- ▁doen
+- tha
+- ▁Pi
+- ▁belonged
+- ▁stayed
+- zione
+- ▁ŝi
+- ▁delightful
+- ▁forever
+- ▁hunt
+- tir
+- ▁profession
+- ▁staring
+- rt
+- ▁bur
+- ور
+- cker
+- ▁beloved
+- ▁invited
+- nyi
+- ▁realized
+- گ
+- ▁seg
+- ▁community
+- ات
+- ▁absurd
+- ▁loving
+- юць
+- gre
+- ▁acquainted
+- sk
+- ▁negro
+- lus
+- ▁accomplished
+- ص
+- 니
+- ▁supreme
+- ▁musical
+- ▁writer
+- вер
+- ▁delivered
+- ▁metal
+- ▁resist
+- M
+- یم
+- ▁estate
+- ingen
+- ▁nascut
+- ▁smaller
+- ▁cheek
+- ই
+- vol
+- ▁lift
+- ajn
+- gang
+- ▁pol
+- ▁policy
+- ▁desk
+- 中
+- แ
+- iya
+- ▁islands
+- ▁resolution
+- ▁aquest
+- ▁acu
+- fort
+- 거든
+- ▁allen
+- ▁emp
+- ▁ம
+- ▁coup
+- сь
+- any
+- ▁baba
+- ▁commission
+- ▁bara
+- зі
+- igo
+- ده
+- ِ
+- ▁prize
+- ▁volume
+- ▁quand
+- ▁represented
+- ▁quantity
+- ▁sex
+- ▁تا
+- ▁generous
+- ▁amid
+- тра
+- ians
+- fra
+- ▁developed
+- tier
+- ▁midnight
+- 치
+- ▁intellectual
+- ▁ihrer
+- ii
+- ▁fame
+- board
+- ▁attached
+- ▁mist
+- ▁egypt
+- ▁jean
+- dde
+- ▁Gi
+- ▁là
+- ▁kinds
+- ல்
+- ара
+- ▁independent
+- ▁insisted
+- ès
+- ▁chest
+- ▁shoes
+- ▁forma
+- ▁render
+- ▁awake
+- ▁veel
+- ۆ
+- ser
+- ▁total
+- ман
+- ▁desperate
+- ْ
+- ▁gukora
+- ▁arrive
+- zeit
+- ▁log
+- vert
+- ▁require
+- pel
+- ▁qualities
+- ▁Lu
+- ▁haste
+- ▁normal
+- iḍ
+- ▁fla
+- kana
+- ▁infinite
+- ▁trap
+- 의
+- ▁railway
+- gere
+- chten
+- ▁하는
+- ▁settle
+- ▁concluded
+- ▁zag
+- put
+- ▁guest
+- ▁combat
+- ▁uno
+- ▁buck
+- èrent
+- dale
+- ▁effects
+- ыя
+- ▁wheel
+- ▁handed
+- ▁heel
+- ала
+- cé
+- ▁maig
+- enden
+- ▁changes
+- ▁magnificent
+- ee
+- ▁proceed
+- ▁Да
+- ▁bible
+- ▁text
+- ▁Um
+- ▁cada
+- ▁abandon
+- ▁citizens
+- ▁burn
+- 하는
+- ▁pat
+- म
+- ▁Nach
+- ▁recovered
+- ▁tak
+- ▁properly
+- ▁ambition
+- hy
+- ▁За
+- ▁novel
+- ets
+- ▁folk
+- ▁vin
+- نىڭ
+- ▁defence
+- ▁plate
+- ▁tide
+- ▁tempo
+- ▁strongly
+- ▁pres
+- ath
+- ude
+- quer
+- ▁flood
+- ▁net
+- ے
+- rum
+- hearted
+- ▁دە
+- ▁committee
+- ▁altar
+- din
+- tem
+- ▁shortly
+- ▁Par
+- ▁ɣer
+- ค
+- ▁Ama
+- ▁anywhere
+- ▁movements
+- тэ
+- ের
+- ▁distinction
+- nel
+- ▁recht
+- ▁growth
+- und
+- зы
+- ▁multitude
+- unk
+- ▁أ
+- eva
+- ▁wilt
+- ▁beard
+- ▁mur
+- ▁vessels
+- ▁fri
+- nar
+- tiva
+- ▁jews
+- atu
+- нага
+- ▁sovereign
+- ▁sé
+- cla
+- ▁melancholy
+- uli
+- ▁urged
+- tó
+- ▁wretched
+- ▁ignorance
+- ▁rights
+- ▁ends
+- schaft
+- ▁гэтага
+- aka
+- ▁locked
+- ▁audience
+- ▁pit
+- ▁families
+- ▁dame
+- ▁leaned
+- ▁var
+- ij
+- ой
+- ▁thine
+- ер
+- ▁Tra
+- ிய
+- ▁gates
+- ▁kunnen
+- bul
+- ▁бер
+- ▁incident
+- 단
+- ▁আ
+- ▁treat
+- ▁crowded
+- ▁humanity
+- bin
+- ▁likewise
+- ▁fil
+- ▁tous
+- ী
+- anza
+- ▁utmost
+- ▁stronger
+- nta
+- ▁shone
+- lay
+- ▁thinks
+- ator
+- ▁puis
+- ▁moet
+- wen
+- ▁reputation
+- تی
+- ▁thirteen
+- ▁priests
+- ▁при
+- ▁wedding
+- ▁eens
+- pri
+- ▁mixed
+- cz
+- ▁dancing
+- ▁observation
+- ▁più
+- ▁chin
+- ▁fence
+- ▁seus
+- ▁mankind
+- ▁ihren
+- ▁frequent
+- ▁nearest
+- mus
+- »
+- ▁lock
+- ▁reference
+- ▁рас
+- ▁maiden
+- ▁kuwa
+- ▁herr
+- ▁extended
+- ▁tard
+- ▁señor
+- ▁lawyer
+- ▁substance
+- ▁liberal
+- ría
+- ▁iets
+- C
+- ▁manners
+- ▁tar
+- ▁cas
+- ▁보
+- ▁Post
+- ▁borne
+- ▁была
+- ▁calls
+- ▁beach
+- ска
+- ▁performed
+- ▁brow
+- ▁cries
+- ever
+- lic
+- ▁gusa
+- ▁noon
+- ▁ruth
+- ▁distress
+- K
+- ▁umuntu
+- ▁povas
+- ▁goodness
+- ▁afterward
+- ▁ɣef
+- '",'
+- ▁seriously
+- су
+- ▁lately
+- ▁porte
+- rse
+- ▁grandfather
+- ▁jumped
+- ом
+- ▁être
+- ö
+- ▁komen
+- ▁mood
+- ▁strain
+- 了
+- ▁related
+- ▁habits
+- ▁map
+- roj
+- ▁congress
+- ما
+- ▁victim
+- ▁Saint
+- heim
+- ▁manifest
+- が
+- ▁slip
+- ▁scientific
+- ▁rob
+- ▁San
+- ▁poetry
+- ход
+- ▁recent
+- ified
+- ▁ward
+- ery
+- ▁remaining
+- ▁gather
+- isme
+- ▁gro
+- ▁area
+- ▁fired
+- ▁poured
+- ▁mille
+- ▁나도
+- ▁거기
+- mme
+- ▁zeit
+- ი
+- nza
+- ska
+- ▁prospect
+- ல
+- ▁elements
+- ▁conviction
+- ▁bade
+- ُ
+- ▁fro
+- рт
+- ▁Des
+- ▁center
+- ▁decide
+- pon
+- һа
+- ▁granted
+- ▁crack
+- jar
+- ulo
+- kar
+- ო
+- ▁legal
+- ิ
+- dda
+- ▁studied
+- 드
+- az
+- ▁St
+- ▁variety
+- ▁té
+- D
+- ▁teacher
+- ▁obey
+- tse
+- ▁painful
+- ▁zeer
+- ▁byose
+- ▁yield
+- ▁medicine
+- qa
+- ட
+- ▁H
+- ▁muttered
+- ▁boots
+- ▁alice
+- ▁heavily
+- ▁eastern
+- rez
+- ches
+- aron
+- ▁Zu
+- ▁feed
+- rent
+- ▁vanished
+- lam
+- air
+- ி
+- av
+- ▁ntabwo
+- ▁쫌
+- F
+- ▁opposition
+- ▁hitherto
+- ▁views
+- ett
+- ▁leaning
+- 할
+- burg
+- ▁bride
+- ral
+- iv
+- ▁rya
+- ▁unusual
+- abi
+- ▁رو
+- ▁seeking
+- ▁உ
+- ▁Sta
+- ray
+- four
+- ▁absent
+- ▁assure
+- com
+- ▁contact
+- ▁darling
+- лар
+- ▁hadn
+- ▁apparent
+- еш
+- will
+- ▁firmly
+- ▁stupid
+- ▁chiefly
+- ▁suggestion
+- ▁왜
+- ▁sole
+- born
+- ▁ёсць
+- ▁india
+- ▁enthusiasm
+- ▁assumed
+- lot
+- ▁hor
+- tera
+- ▁stem
+- tri
+- ▁startled
+- اد
+- ▁alike
+- town
+- ▁production
+- 려
+- ▁chain
+- asi
+- vy
+- ▁shaking
+- isto
+- ▁standard
+- 다고
+- you
+- ▁Fi
+- ▁actions
+- ▁mistaken
+- rea
+- ▁section
+- ade
+- ▁Ke
+- чи
+- ства
+- ▁treatment
+- ▁гэтым
+- dio
+- ▁sounded
+- স
+- aɣ
+- ▁separated
+- ▁veil
+- ▁huis
+- ▁wherever
+- ela
+- ▁gleich
+- erde
+- ▁sighed
+- anga
+- uw
+- clo
+- ▁Jordi
+- nch
+- ▁ім
+- û
+- ▁padre
+- octubre
+- tik
+- ▁там
+- non
+- iger
+- ▁singular
+- ▁latin
+- ▁daughters
+- ▁Mal
+- ▁team
+- ▁tur
+- ane
+- ▁nombre
+- dom
+- ▁jar
+- ica
+- 라고
+- ▁probable
+- ete
+- ▁shirt
+- ▁charity
+- ▁purple
+- ▁sentiment
+- ▁everyone
+- ▁lincoln
+- ▁owner
+- ▁regiment
+- ▁palm
+- ene
+- َ
+- gro
+- ▁retreat
+- ▁fou
+- ▁э
+- ors
+- ▁contest
+- ▁smart
+- ▁merchant
+- ▁tend
+- ▁Diese
+- ▁protest
+- ▁copy
+- бе
+- ▁cared
+- fold
+- ▁madam
+- ▁leven
+- ▁brand
+- ▁gw
+- ded
+- ▁av
+- ▁driving
+- ▁einmal
+- ▁campaign
+- tí
+- ▁sí
+- elles
+- ▁fet
+- ▁steep
+- ▁puc
+- ▁от
+- ▁bath
+- ▁adventure
+- ة
+- ▁hated
+- ▁wandering
+- keit
+- ▁ஆ
+- ▁conflict
+- தி
+- ▁reported
+- ▁contrast
+- ▁instinct
+- ▁astonishment
+- ▁betty
+- ▁Д
+- ▁калі
+- ▁rested
+- ▁continually
+- ▁sensible
+- ▁entering
+- cas
+- ▁lange
+- ▁doctrine
+- ago
+- ▁whisper
+- ▁cri
+- ▁intense
+- 수
+- ▁irish
+- vous
+- ły
+- ▁Pri
+- ▁opera
+- ill
+- ▁stiff
+- gh
+- ▁revealed
+- ▁ப
+- gui
+- fal
+- uz
+- ▁tells
+- ▁kein
+- ▁ram
+- ▁acted
+- ▁detective
+- dal
+- ▁cord
+- ил
+- ыш
+- ▁pues
+- ▁structure
+- kken
+- ▁pierre
+- ▁porta
+- تا
+- ▁transport
+- бо
+- ▁cheerful
+- ▁breeze
+- ▁poste
+- ▁hero
+- kul
+- ▁bud
+- ▁sport
+- ต
+- ст
+- sent
+- ▁attempted
+- ▁departed
+- ▁sud
+- ▁gibt
+- ▁someone
+- ▁imagined
+- ▁après
+- ▁slightest
+- hanga
+- 하
+- ▁agent
+- ▁astonished
+- ▁vice
+- ว
+- 진
+- bia
+- ttle
+- ې
+- gie
+- rimo
+- gle
+- ▁deliver
+- ech
+- gla
+- ▁venture
+- ▁examination
+- ▁devotion
+- nor
+- ▁diesen
+- ▁pursued
+- abu
+- ▁agreeable
+- là
+- jen
+- mat
+- pit
+- ▁protect
+- ▁urbo
+- ▁cell
+- ▁succeed
+- elijk
+- twa
+- ▁millions
+- ▁rude
+- ▁clerk
+- ▁experiment
+- ▁schi
+- ▁Vor
+- ▁terwijl
+- ▁hab
+- ▁Ŝi
+- ו
+- rí
+- ▁cease
+- ▁consequently
+- ى
+- ▁spell
+- gal
+- ▁centuries
+- ▁Pe
+- ose
+- ▁francis
+- ▁quarters
+- ▁rat
+- сі
+- ction
+- tje
+- ▁alter
+- gua
+- ked
+- ente
+- ret
+- ▁bwo
+- ▁hesitated
+- ▁dining
+- ▁ح
+- tions
+- ▁adopted
+- ▁mild
+- ered
+- ▁hätte
+- ▁dal
+- '6'
+- ▁descended
+- ▁quit
+- を
+- ს
+- yon
+- ▁infant
+- mor
+- ▁theatre
+- ▁shield
+- ▁examined
+- ▁occasions
+- uff
+- ▁prayers
+- ▁platform
+- ▁descend
+- мен
+- ▁cloak
+- ▁issue
+- ▁falls
+- ▁resumed
+- ▁tam
+- ▁également
+- ith
+- ▁coach
+- gate
+- ▁manage
+- ▁coal
+- ▁inferior
+- ▁gets
+- ▁rules
+- not
+- ▁Em
+- ▁Is
+- ▁deny
+- kte
+- ▁frau
+- ▁drinking
+- ved
+- ▁kate
+- pat
+- ison
+- ▁estat
+- ▁abandoned
+- ▁rot
+- ▁buildings
+- ▁perfection
+- ax
+- ▁murmur
+- bon
+- ▁endure
+- ▁esto
+- ▁senses
+- ▁notion
+- oon
+- å
+- õ
+- ▁including
+- ▁collection
+- ▁kwam
+- ▁newspaper
+- ▁idle
+- ▁sharply
+- ▁같은
+- ▁sollte
+- eyed
+- ▁apartment
+- ▁Una
+- ched
+- ▁drank
+- ▁majority
+- ▁throwing
+- ▁gratitude
+- ▁marble
+- ▁repeat
+- ▁winds
+- ▁piano
+- ▁songs
+- door
+- ага
+- ов
+- ▁readily
+- ▁essential
+- ▁wollte
+- cken
+- 긴
+- ▁april
+- llen
+- ▁fins
+- ▁sins
+- ▁pace
+- ▁anderen
+- ▁passat
+- ▁tren
+- comp
+- ▁rifle
+- ▁impulse
+- ▁мяне
+- U
+- hora
+- ▁Barcelona
+- tia
+- ▁buri
+- iga
+- ▁badly
+- ▁bark
+- zza
+- ▁candle
+- ڭ
+- rar
+- ema
+- quin
+- ض
+- ▁wearing
+- ▁femme
+- ▁lane
+- ▁twee
+- ▁furnished
+- ▁можна
+- ▁inches
+- ced
+- ▁таксама
+- ▁conception
+- ▁catholic
+- ▁wilde
+- iter
+- ▁zeggen
+- ▁represent
+- ё
+- द
+- ▁Auf
+- ▁welche
+- glo
+- ▁deserted
+- agost
+- ▁très
+- aban
+- 和
+- ▁safely
+- ▁poverty
+- ▁professional
+- ▁Can
+- ае
+- ▁climbed
+- ▁ng
+- ▁concern
+- род
+- ▁elder
+- क
+- ▁каб
+- ▁reflection
+- ▁turns
+- nny
+- ▁literary
+- esa
+- pha
+- cus
+- ▁tag
+- ▁heavens
+- ▁votre
+- ▁steadily
+- ▁maintain
+- ih
+- stru
+- ▁kiu
+- ▁marie
+- ▁activity
+- دى
+- ▁dragged
+- ▁char
+- ori
+- ▁ந
+- ▁یک
+- ▁province
+- ▁pine
+- mul
+- ▁–
+- cc
+- ▁мне
+- lf
+- ▁located
+- pper
+- two
+- '4'
+- ▁sunshine
+- ая
+- ration
+- ▁intent
+- ▁reduced
+- lä
+- ▁defend
+- ines
+- ▁owing
+- ▁russian
+- ल
+- fan
+- вя
+- book
+- ▁woe
+- ena
+- ▁knocked
+- ▁virginia
+- ▁compared
+- uta
+- isch
+- ono
+- cri
+- ▁freely
+- kom
+- oro
+- дзі
+- ▁possibility
+- ▁трэба
+- ▁quart
+- ▁tap
+- lí
+- ▁lucy
+- ▁рэ
+- ▁stout
+- ▁soll
+- ▁criminal
+- تى
+- тер
+- ▁railroad
+- ight
+- пу
+- lung
+- ium
+- ▁hunger
+- لى
+- ▁overcome
+- ▁belonging
+- ▁nacht
+- oh
+- til
+- ▁dun
+- ▁Wir
+- ▁Mae
+- von
+- ▁sixth
+- rai
+- lent
+- ▁medical
+- ▁belle
+- ▁bwe
+- ▁X
+- ▁kwe
+- ▁refer
+- 던
+- ▁amusement
+- ▁refuge
+- ▁pressure
+- elo
+- tle
+- ▁tro
+- ▁suspected
+- им
+- ▁plu
+- tten
+- ▁obu
+- ał
+- ▁realize
+- ▁furniture
+- ▁aŭ
+- shed
+- ime
+- ▁student
+- ग
+- rk
+- ▁ndetse
+- els
+- ▁opinions
+- tered
+- ▁건
+- ▁poison
+- ▁ils
+- ▁خ
+- ▁sensation
+- ▁folks
+- ▁contract
+- ▁iyi
+- த்
+- ▁tant
+- gg
+- 년
+- ▁finish
+- ▁Kon
+- ▁vez
+- ▁invitation
+- ▁threatened
+- ▁morgen
+- ▁director
+- ▁arch
+- ▁acting
+- ৰ
+- ▁july
+- ▁swiftly
+- ▁sia
+- tara
+- ную
+- ▁tear
+- व
+- ▁ச
+- ddi
+- ▁mc
+- cen
+- ▁describe
+- worth
+- ▁concealed
+- ▁angels
+- ▁lamb
+- ▁anche
+- ▁shell
+- ▁china
+- zan
+- ▁돼
+- ▁gathering
+- ▁experienced
+- ▁plot
+- ▁tijd
+- ▁couch
+- ▁disturbed
+- ▁Sed
+- ▁corps
+- cor
+- usa
+- ▁positive
+- teil
+- unt
+- ▁marks
+- ▁secured
+- ▁grateful
+- ▁ugly
+- cons
+- ▁disposed
+- ▁election
+- دی
+- ▁circumstance
+- gas
+- ▁nonsense
+- ▁load
+- ▁gab
+- rü
+- ▁steam
+- ▁landed
+- ▁knock
+- ▁intellect
+- ▁zusammen
+- ▁Т
+- ▁brook
+- ▁farewell
+- ▁mare
+- ▁tight
+- ▁autumn
+- ҡы
+- bal
+- ▁sorts
+- ▁sale
+- bus
+- ▁cela
+- จ
+- ngu
+- qi
+- ▁reflected
+- burn
+- ▁reaching
+- ▁gifts
+- พ
+- ọ
+- ▁пу
+- app
+- ▁Cor
+- اب
+- ▁scotland
+- mini
+- ▁بود
+- ▁John
+- xa
+- dzie
+- ▁blank
+- ▁glow
+- ▁rev
+- 上
+- dri
+- phi
+- ▁poco
+- ▁purposes
+- бу
+- ▁profound
+- ▁kar
+- ▁muss
+- ▁hugh
+- ▁machen
+- ▁starting
+- ▁febrer
+- い
+- เป็น
+- ▁있
+- ▁expense
+- ▁lee
+- kunda
+- ▁flor
+- ▁weet
+- ▁handkerchief
+- ▁음
+- ▁flu
+- ▁kama
+- ι
+- ▁score
+- ▁cosa
+- ▁swear
+- ▁methods
+- ▁cheap
+- ի
+- kle
+- ▁imbere
+- yard
+- ▁былі
+- ▁sending
+- ▁millor
+- ▁intelligent
+- ▁exchange
+- ▁attacked
+- ▁Los
+- ▁beasts
+- ▁lust
+- log
+- ▁gra
+- ▁horizon
+- uba
+- ▁sis
+- ▁pilot
+- ▁naked
+- ▁Ju
+- tzen
+- mir
+- ▁etc
+- ▁deeper
+- ▁ўжо
+- mpa
+- ▁hasta
+- ▁glimpse
+- пе
+- ▁industry
+- ▁leurs
+- ▁alla
+- gard
+- ▁arrest
+- ▁premier
+- ▁prop
+- fle
+- ▁pointing
+- ▁exhausted
+- ▁kw
+- ▁fed
+- ów
+- ▁rien
+- urs
+- ▁prepare
+- ▁aquel
+- ▁owe
+- ious
+- ▁detail
+- bit
+- ▁crossing
+- ▁데
+- ▁disappointment
+- bera
+- ▁johnson
+- isse
+- ▁strangers
+- по
+- ▁lightly
+- ▁begins
+- ila
+- ciones
+- ую
+- ▁та
+- ▁fearful
+- ▁crept
+- ▁error
+- ▁remind
+- ▁번
+- ▁aga
+- ▁drunk
+- ▁nights
+- ▁disappointed
+- ▁bone
+- bur
+- ▁assist
+- lum
+- dad
+- ▁lesson
+- bura
+- ▁contempt
+- gana
+- ▁training
+- nze
+- rian
+- hum
+- ▁sailed
+- ▁staff
+- аць
+- ▁perform
+- ▁included
+- ▁blessing
+- dir
+- ▁그니까
+- ▁cure
+- cs
+- '9'
+- ▁That
+- ▁jump
+- ▁landing
+- ▁dios
+- ▁yer
+- nnen
+- gat
+- ▁happens
+- cks
+- tti
+- ▁Э
+- ▁sap
+- ▁vend
+- ▁merit
+- प
+- ▁spa
+- ▁mid
+- ▁climb
+- gged
+- ▁pound
+- ▁бар
+- bare
+- ▁niets
+- ▁leaf
+- гра
+- ▁visitor
+- ▁moja
+- ▁homme
+- ▁tremendous
+- 까지
+- ▁przy
+- ▁confession
+- nal
+- whi
+- '2'
+- bury
+- ей
+- ▁orange
+- '8'
+- ▁pistol
+- ▁lunch
+- ▁Zeit
+- ▁rub
+- “
+- ▁favourite
+- ▁яна
+- ▁suspect
+- ▁charged
+- ▁perceive
+- ▁completed
+- lor
+- י
+- ▁purchase
+- ▁rep
+- ▁penny
+- мер
+- kim
+- eo
+- ▁characters
+- ▁motive
+- gram
+- ▁jak
+- ▁heute
+- ên
+- ▁document
+- form
+- spir
+- ▁zien
+- ▁increasing
+- ▁بر
+- ▁chapel
+- ▁gazing
+- یا
+- ▁roar
+- ▁provide
+- ▁acquired
+- ▁pursuit
+- ▁formerly
+- rä
+- ▁involved
+- ▁beiden
+- ▁mirror
+- ident
+- ▁wäre
+- ▁dressing
+- ▁meantime
+- ▁straw
+- llo
+- ▁govern
+- ▁european
+- nder
+- ▁pal
+- ht
+- ▁illustration
+- ▁rein
+- ▁eve
+- ө
+- ▁talent
+- ▁mismo
+- lier
+- ▁parted
+- ▁bedroom
+- ▁notwithstanding
+- би
+- ▁minor
+- ▁loves
+- ▁directions
+- ▁marched
+- ume
+- ▁genuine
+- tat
+- ▁phrase
+- ▁کی
+- ▁gloom
+- ▁organ
+- ▁cru
+- ▁handle
+- ▁jealous
+- zon
+- ▁mont
+- stan
+- ▁feeble
+- ▁remove
+- mic
+- abo
+- ubi
+- ▁fury
+- ▁cutting
+- ▁application
+- ▁provisions
+- ரி
+- jya
+- gt
+- ▁exposed
+- ▁crow
+- ▁continent
+- shire
+- ▁augen
+- ▁drag
+- ப
+- ▁vroeg
+- ▁jones
+- ▁closer
+- ▁erst
+- gia
+- ismo
+- kazi
+- ▁auto
+- ▁Mont
+- ▁lag
+- ▁bor
+- ▁cock
+- ▁fre
+- igkeit
+- ▁Ка
+- ▁definite
+- flu
+- ej
+- ▁beau
+- ▁Ма
+- br
+- тан
+- ▁Veni
+- ▁punt
+- ▁entra
+- ▁crystal
+- gut
+- ▁qual
+- ▁enterprise
+- ▁driver
+- igis
+- ario
+- ▁sink
+- ▁moses
+- ป
+- ▁exceedingly
+- ▁Ben
+- ▁stra
+- adi
+- ▁scale
+- ის
+- ▁operation
+- vas
+- ▁pretend
+- lant
+- ▁anyway
+- ▁tones
+- ▁pig
+- ▁relative
+- ▁rivers
+- ▁disse
+- ▁swung
+- ▁versch
+- ▁pos
+- ▁natives
+- ▁brush
+- ▁dort
+- azi
+- ▁fille
+- get
+- ▁gre
+- ▁knee
+- ə
+- sin
+- ▁П
+- ▁hiding
+- enda
+- ▁Mit
+- uza
+- fite
+- ṣ
+- ▁ale
+- ▁scarce
+- ப்ப
+- nja
+- ▁walter
+- nat
+- pped
+- ▁doorway
+- nim
+- ▁schools
+- ▁seed
+- ▁Pu
+- ▁jerusalem
+- ▁program
+- ▁ellen
+- ▁inquiry
+- ▁execution
+- ▁sober
+- ryo
+- sky
+- ▁hastened
+- ▁gesch
+- ▁nnyo
+- ▁poem
+- ా
+- ▁dag
+- ▁practically
+- fre
+- ▁rolling
+- ▁roads
+- 生
+- ▁seventh
+- nto
+- post
+- ▁бы
+- ▁preserved
+- ▁intimate
+- ваць
+- ▁supported
+- ▁receiving
+- ▁pages
+- ор
+- ▁fears
+- ▁helpless
+- its
+- ▁joke
+- ас
+- ▁Ntabwo
+- shing
+- ▁violet
+- '7'
+- ▁restored
+- ▁today
+- ▁revenge
+- ▁그리고
+- gara
+- ▁mwa
+- ▁reasonable
+- ▁hundreds
+- ▁rug
+- ▁hoofd
+- ▁squire
+- ▁project
+- ▁wash
+- ▁dorothy
+- ъ
+- ▁rabbit
+- rac
+- ▁tanto
+- rra
+- ▁causa
+- ▁virgin
+- fern
+- ▁generation
+- ▁د
+- kins
+- ▁craft
+- ▁bush
+- ▁lightning
+- ▁fois
+- ▁dirty
+- ▁estaba
+- rel
+- ▁declare
+- ▁eaten
+- long
+- گە
+- ▁preserve
+- ▁basket
+- ore
+- uze
+- ör
+- ▁dieses
+- ▁oude
+- ▁ق
+- uni
+- ▁wives
+- aĵo
+- ▁zonder
+- ▁universe
+- ▁commercial
+- лек
+- ▁ме
+- ▁würde
+- rais
+- ▁folly
+- ▁hatred
+- ▁exception
+- ▁beheld
+- ▁terug
+- ▁creation
+- ▁bou
+- V
+- uc
+- ▁tramp
+- ▁wrath
+- ▁whereas
+- ▁dearest
+- ▁politics
+- ▁characteristic
+- inte
+- mma
+- ▁permission
+- ▁jour
+- ▁shouldn
+- ▁silly
+- ▁weer
+- ▁string
+- ▁satisfy
+- zig
+- Any
+- H
+- sca
+- ▁antwoordde
+- ▁gospel
+- ▁horn
+- werk
+- ▁september
+- rant
+- ised
+- aren
+- ▁Ex
+- tree
+- ▁weder
+- ▁tun
+- ▁leben
+- cing
+- ▁communication
+- ▁funny
+- ▁beating
+- iendo
+- ▁referred
+- dat
+- ▁Tiu
+- isha
+- old
+- ▁unconscious
+- ▁کو
+- ▁neu
+- ▁cents
+- 생
+- ▁verb
+- uku
+- ▁raising
+- lando
+- ▁missed
+- ▁terre
+- ▁upstairs
+- ҙ
+- udi
+- ▁shalt
+- sam
+- ▁flashed
+- ▁dig
+- tab
+- ▁toujours
+- ller
+- ▁cape
+- mt
+- ▁assume
+- وی
+- cel
+- ▁knights
+- ▁belt
+- 트
+- ▁outer
+- ▁tax
+- ▁examine
+- rek
+- wr
+- ▁considering
+- ▁maria
+- ▁pole
+- ▁tr
+- ▁salvation
+- lib
+- ible
+- cat
+- ▁Wie
+- 到
+- ▁Вы
+- ▁silently
+- ▁apt
+- ார்
+- 더라고
+- tä
+- ▁ئە
+- ▁anyhow
+- giye
+- genda
+- ▁porter
+- pie
+- تر
+- ▁아니야
+- zio
+- ▁Omu
+- ▁cleared
+- ▁previously
+- back
+- ega
+- ▁3
+- ▁masters
+- ▁counter
+- ▁imi
+- ▁pains
+- ▁vrouw
+- ▁zeker
+- ▁clay
+- ▁ralph
+- ▁mail
+- cies
+- ▁beaten
+- ▁cotton
+- य
+- ะ
+- ▁authorities
+- ु
+- ▁cinema
+- headed
+- ▁stirred
+- ▁eigen
+- nin
+- eff
+- agi
+- мә
+- ▁font
+- uto
+- cr
+- رو
+- ▁kit
+- ช
+- uga
+- ▁کار
+- ▁Sol
+- ▁distinctly
+- ▁belongs
+- ▁recall
+- ذ
+- ова
+- lig
+- ▁quelques
+- tru
+- rd
+- reich
+- ▁solution
+- ▁gesture
+- fully
+- ▁suis
+- ▁Ibi
+- ▁leather
+- ▁runs
+- minded
+- ▁cop
+- ▁arrangement
+- ▁phil
+- mine
+- ▁چ
+- rze
+- ▁müssen
+- ▁бо
+- sso
+- ▁برای
+- elli
+- ▁monte
+- ▁indi
+- бі
+- ▁Za
+- ▁heir
+- ▁daylight
+- ▁petit
+- ▁weil
+- ▁ع
+- nz
+- ▁limited
+- γ
+- ▁sage
+- ▁unexpected
+- ▁buryo
+- bro
+- ▁Or
+- ▁christians
+- gur
+- ▁cardinal
+- ▁sp
+- ▁paint
+- ffe
+- 出
+- ▁earnestly
+- sek
+- ▁dix
+- ▁oath
+- ئ
+- ▁princes
+- ▁stuck
+- кам
+- ▁ireland
+- ▁roz
+- este
+- ▁twa
+- গ
+- ▁terra
+- ▁кан
+- eɣ
+- ▁nooit
+- tanga
+- ▁passa
+- ▁можа
+- ▁resting
+- dam
+- ▁هم
+- kka
+- saba
+- ο
+- ▁intend
+- R
+- ▁cheer
+- ա
+- cut
+- ▁culture
+- ▁Men
+- ▁Пра
+- ▁measures
+- ▁album
+- ▁limbs
+- ▁وَ
+- ▁주
+- ▁dies
+- ▁depths
+- ▁offering
+- ▁touching
+- ал
+- ▁trunk
+- จะ
+- ӹ
+- ▁munsi
+- ▁moest
+- ▁ließ
+- ▁mademoiselle
+- ▁eki
+- ▁gown
+- 地
+- ▁bre
+- ▁생각
+- ▁relieved
+- ▁alleen
+- ▁establishment
+- ▁fitted
+- jja
+- ▁leaped
+- ▁zero
+- ▁tribe
+- ▁border
+- ▁맞아
+- ▁laŭ
+- ▁saturday
+- ▁slope
+- ▁hint
+- fata
+- ▁cyo
+- ▁captured
+- 제
+- ▁tiempo
+- wala
+- ▁block
+- iɣ
+- bau
+- ▁murderer
+- ▁일
+- ▁display
+- ld
+- ço
+- ▁bro
+- ▁slavery
+- bie
+- ▁ainsi
+- eka
+- ▁review
+- ▁heap
+- ▁enable
+- ▁tribes
+- ▁wing
+- ண
+- ▁engine
+- ▁instructions
+- ctor
+- ology
+- стаў
+- ▁ці
+- ▁conducted
+- ▁père
+- fon
+- ▁bang
+- ▁mun
+- ▁Kan
+- ▁resistance
+- col
+- mmen
+- rah
+- adas
+- ▁creek
+- শ
+- ▁mot
+- ér
+- ▁duchess
+- მ
+- ▁patron
+- jes
+- โ
+- ▁yu
+- eye
+- tto
+- ▁planet
+- ▁confused
+- uit
+- ▁elsewhere
+- ▁column
+- ▁tales
+- ello
+- ▁planta
+- ▁엄청
+- غا
+- ▁noted
+- tori
+- ▁honey
+- ▁nel
+- raj
+- ▁durant
+- acht
+- ▁fog
+- ming
+- ▁ف
+- vre
+- ▁romance
+- ▁bowl
+- tore
+- ▁dues
+- ▁procession
+- юцца
+- кай
+- ▁Sin
+- termin
+- right
+- ▁slain
+- weg
+- ▁obvious
+- ▁agony
+- ▁longing
+- ▁insist
+- lea
+- ▁bend
+- ▁ты
+- zwi
+- ▁containing
+- ▁euch
+- ▁brass
+- ▁performance
+- ур
+- ▁War
+- ▁romans
+- ▁october
+- ▁geweest
+- 年
+- rri
+- ▁smoking
+- ▁sailor
+- ▁gang
+- ▁collar
+- ▁wagon
+- ő
+- рам
+- ▁vest
+- ▁chase
+- ▁passions
+- ▁mé
+- ▁engagement
+- ▁brings
+- ▁collected
+- ▁Esperanto
+- ▁screen
+- чу
+- ▁hal
+- ▁steal
+- ▁weapon
+- ▁Ш
+- ▁zwischen
+- nan
+- ▁grasp
+- ▁flow
+- گر
+- ▁nada
+- pend
+- ▁puzzled
+- ▁drama
+- có
+- ▁gaan
+- 면은
+- ▁opposed
+- 多
+- ▁cares
+- pul
+- ▁pillow
+- ▁jeune
+- arse
+- ▁imperial
+- ▁remarks
+- と
+- iste
+- ▁average
+- istic
+- ▁wandered
+- vá
+- ▁gardens
+- ▁prevented
+- ্য
+- ▁weiter
+- ▁Б
+- ▁lap
+- ▁fastened
+- ّ
+- auf
+- uru
+- izing
+- ▁released
+- ▁sto
+- ▁그럼
+- ▁issued
+- pes
+- ▁celebrated
+- ▁scenes
+- itu
+- ▁Mon
+- ▁anna
+- W
+- rer
+- lim
+- ▁din
+- ▁reserve
+- ▁cards
+- kola
+- mak
+- ▁unpleasant
+- ▁capacity
+- ▁passes
+- sser
+- ▁foe
+- gno
+- ▁avenue
+- ши
+- ▁Ĉu
+- ▁carpet
+- ▁wire
+- ▁recently
+- ▁School
+- spiel
+- stu
+- ран
+- ▁deeds
+- ▁maintained
+- sus
+- gomba
+- лә
+- ▁cet
+- ▁favorite
+- он
+- cí
+- box
+- cro
+- ▁precisely
+- ▁fetch
+- قا
+- ▁aquella
+- lio
+- anya
+- vid
+- zzi
+- ult
+- ological
+- ont
+- ▁کا
+- ▁Nu
+- ▁largely
+- ▁reflect
+- ▁sheet
+- hol
+- ▁könig
+- ▁contains
+- aria
+- script
+- ▁ubwo
+- ▁Pre
+- ▁Ste
+- ▁geld
+- ▁gloomy
+- rich
+- ▁denied
+- ffer
+- ▁defeat
+- дар
+- ▁servi
+- ▁lean
+- ▁això
+- ét
+- ▁product
+- 화
+- ▁solo
+- ▁neighbourhood
+- apo
+- mana
+- ace
+- ▁finds
+- ▁دو
+- ▁worry
+- umwa
+- ▁bose
+- ▁sul
+- ǧ
+- ▁dove
+- ▁russia
+- ün
+- ▁prze
+- bir
+- ไม่
+- ▁gale
+- egu
+- wick
+- ppen
+- elt
+- ▁prey
+- бра
+- ▁exc
+- 우
+- zz
+- ▁بۆ
+- ▁clothing
+- ▁thread
+- ▁indignation
+- top
+- ▁powder
+- ▁regarding
+- umba
+- ▁kumu
+- ห
+- bara
+- ▁groups
+- ▁electric
+- ▁Har
+- ▁responsible
+- ▁include
+- tatu
+- ▁tradition
+- ӱ
+- onde
+- ▁steamer
+- lü
+- 물
+- া
+- yu
+- ▁kom
+- kuba
+- bor
+- ▁sergeant
+- ification
+- ▁university
+- ▁Quan
+- ▁civilization
+- ▁By
+- ▁git
+- ▁hum
+- cis
+- ▁francs
+- ▁wherefore
+- һы
+- ▁convenient
+- ota
+- ▁mingled
+- ▁abruptly
+- ▁wilderness
+- ▁onze
+- ▁propose
+- ▁Mor
+- ▁desires
+- ▁primera
+- ▁messenger
+- ▁physician
+- вае
+- ▁contain
+- 都
+- ange
+- ▁hook
+- ▁security
+- thing
+- ▁welt
+- ▁damp
+- ▁ciutat
+- 科
+- maze
+- ▁respectable
+- ▁ceremony
+- rak
+- ▁jamais
+- ▁Wo
+- kwi
+- ▁mí
+- ▁visitors
+- ▁suo
+- ▁wanting
+- ▁Ik
+- ச்
+- ▁vader
+- ▁funeral
+- hall
+- bank
+- ▁residence
+- ӧ
+- ▁recognize
+- ▁flock
+- ▁muß
+- ▁heavenly
+- ▁polly
+- bol
+- ▁damit
+- ▁slender
+- ▁assurance
+- ▁countess
+- ▁cavalry
+- ▁Aquest
+- ▁eben
+- ▁necessarily
+- ых
+- ▁scent
+- iness
+- ket
+- ▁invisible
+- ▁consists
+- ਾ
+- ▁flo
+- ▁laura
+- ▁persuaded
+- ▁afite
+- ▁permanent
+- で
+- ▁Mas
+- ▁wider
+- emt
+- chte
+- ▁hano
+- σ
+- ▁via
+- ▁commenced
+- mera
+- ข
+- emi
+- ▁rosa
+- ▁vote
+- antes
+- ▁balance
+- ín
+- ▁hatten
+- ▁adam
+- ▁construction
+- ▁territory
+- ▁eighth
+- ▁lassen
+- ▁stole
+- ▁bushes
+- ▁pitch
+- ▁strangely
+- ▁employ
+- ▁based
+- ▁largest
+- ▁grounds
+- cke
+- ▁humour
+- яў
+- ▁arrested
+- ▁virtues
+- ją
+- ▁divers
+- ▁organization
+- ▁dutch
+- fin
+- apa
+- zia
+- ▁existed
+- ▁semi
+- iß
+- bell
+- ▁rays
+- mark
+- ▁operations
+- ▁pays
+- று
+- ▁hamwe
+- oor
+- ▁pile
+- ▁dispute
+- ▁sehen
+- ▁arise
+- вай
+- ▁solitary
+- ▁کرد
+- ▁washed
+- oko
+- faced
+- ▁attracted
+- ир
+- ct
+- sper
+- ▁signed
+- ো
+- ▁reminded
+- ▁torture
+- ▁laying
+- ▁international
+- კ
+- his
+- ▁modest
+- ▁coarse
+- си
+- ث
+- ▁formal
+- ▁fortunate
+- vant
+- ▁meu
+- ▁uyu
+- ▁cunning
+- 했
+- ğ
+- tung
+- ▁saints
+- ters
+- plo
+- ▁foundation
+- ▁jusqu
+- cion
+- ਰ
+- ▁chimney
+- ▁canal
+- ▁discipline
+- jon
+- ▁spoil
+- ▁depth
+- ▁succession
+- ▁homes
+- ▁apple
+- ▁weapons
+- ▁portrait
+- aw
+- ▁ham
+- ▁november
+- ಿ
+- ▁africa
+- ▁bears
+- ▁mostly
+- ▁monarch
+- ▁suggest
+- ▁uwo
+- halten
+- ▁troubles
+- ▁breathing
+- ▁conceal
+- ற்ற
+- ▁sondern
+- ▁Nor
+- ità
+- ▁occur
+- ory
+- ▁pr
+- ▁blew
+- ▁alte
+- ▁enjoyment
+- ruh
+- ▁oni
+- ▁oogen
+- nom
+- ▁бу
+- ▁childhood
+- imu
+- ▁preferred
+- ▁Як
+- ▁assembled
+- ▁Some
+- ▁Au
+- каў
+- цыя
+- ▁moonlight
+- ▁parish
+- czy
+- ▁се
+- ▁prophet
+- ▁holds
+- ▁dish
+- ▁editor
+- ▁tip
+- ▁vulgar
+- ე
+- ▁ха
+- مۇ
+- ▁stephen
+- ▁vital
+- ▁Fu
+- ▁shout
+- ▁serpent
+- ▁telephone
+- ▁confined
+- ister
+- ▁calmly
+- ño
+- ▁losing
+- ▁masses
+- ▁rarely
+- ▁foul
+- ▁egg
+- ▁claims
+- ▁roses
+- ▁churches
+- cum
+- ▁Ze
+- ज
+- eri
+- ▁daring
+- ▁saving
+- ▁photograph
+- boy
+- ▁wrapped
+- 분
+- еп
+- ós
+- ▁lofty
+- ट
+- ▁flank
+- ▁respond
+- tory
+- ball
+- avait
+- ▁aff
+- ▁telegraph
+- kal
+- ▁hunter
+- ▁founded
+- ▁gracious
+- ▁offence
+- ▁foi
+- stall
+- ▁armies
+- ▁த
+- ा
+- ▁mischief
+- ▁neat
+- ெ
+- ▁monde
+- ▁consequences
+- ▁ridiculous
+- ▁aller
+- artagnan
+- ▁nerves
+- ▁donde
+- ▁fanny
+- bet
+- spe
+- zaba
+- ▁obedience
+- swa
+- ि
+- ▁determination
+- iyo
+- ▁contents
+- ']'
+- ▁sala
+- нае
+- ▁studies
+- ▁site
+- iques
+- ▁florence
+- '3'
+- ▁republic
+- ▁sen
+- их
+- here
+- ▁maze
+- ▁timber
+- ▁없어
+- ▁impressed
+- ná
+- ▁그러
+- ▁selected
+- ▁composition
+- kes
+- ibwa
+- ҙа
+- ike
+- ▁guards
+- rn
+- ▁gr
+- ▁kre
+- না
+- ▁uncertain
+- ▁sprach
+- 원
+- ▁channel
+- ▁wrought
+- abantu
+- камі
+- ▁appetite
+- ▁sy
+- sey
+- ▁По
+- ▁tele
+- ▁sunset
+- ▁manager
+- ▁prime
+- ▁adopt
+- wind
+- ▁Gar
+- ▁sweep
+- ▁마
+- ▁ruined
+- row
+- ▁inch
+- ▁ĝin
+- ▁zelf
+- زی
+- ी
+- ▁stolen
+- tür
+- кы
+- dek
+- ▁grain
+- ▁humor
+- hat
+- ▁instruments
+- ▁whip
+- ▁thence
+- ▁inspired
+- ാ
+- lt
+- ▁grandmother
+- ▁dwelling
+- ▁mari
+- tet
+- ▁rend
+- ▁deadly
+- ▁[
+- ▁vos
+- ▁employment
+- ▁desde
+- ▁사람
+- mé
+- ▁pendant
+- ▁conserva
+- ▁Ә
+- num
+- hem
+- gus
+- ▁boston
+- ▁coin
+- ไป
+- ▁sau
+- ington
+- ▁wali
+- ▁thither
+- ▁Wi
+- ▁Ubu
+- ▁villa
+- uloj
+- ▁cau
+- ▁uru
+- ▁gravely
+- ▁vater
+- ▁바
+- ▁fountain
+- ▁outward
+- ▁misfortune
+- ort
+- dé
+- ▁toute
+- ह
+- ▁These
+- hari
+- ▁brethren
+- asse
+- ▁vengeance
+- ▁cos
+- ▁Fe
+- ▁condemned
+- ones
+- ▁bloom
+- ▁bench
+- ▁intervals
+- uld
+- ▁Abantu
+- 정
+- ▁шмат
+- ্
+- ▁Dum
+- ▁risen
+- ▁sil
+- ▁speaks
+- ▁bitterly
+- gam
+- ▁doth
+- ▁ventured
+- ▁insult
+- ▁Hauptstadt
+- ▁trembled
+- ▁root
+- esca
+- fashioned
+- vuit
+- ui
+- ▁burnt
+- ▁няма
+- ienne
+- ▁sagen
+- ▁tiene
+- ▁sofa
+- aŭ
+- ા
+- ▁indifferent
+- nam
+- nell
+- சி
+- iente
+- ▁hopeless
+- ▁weeping
+- tract
+- cos
+- ▁regions
+- ▁что
+- ▁easier
+- kwe
+- ▁из
+- ▁toutes
+- ▁mess
+- ▁assembly
+- ▁pop
+- ▁strict
+- ▁Ci
+- ▁niemand
+- omu
+- май
+- ars
+- می
+- ▁settlement
+- ▁sempre
+- rung
+- ping
+- ▁leap
+- ▁얘기
+- ▁wept
+- ▁Так
+- дан
+- cap
+- ▁ned
+- ▁external
+- ▁stable
+- ▁guardian
+- stri
+- yen
+- ▁aquí
+- கள்
+- ▁mwaka
+- ▁yid
+- ▁sera
+- ▁release
+- ▁illness
+- ▁bite
+- lá
+- rica
+- ▁arme
+- ▁situas
+- ▁tenderness
+- ▁invention
+- tch
+- ▁feather
+- ▁dei
+- ▁soup
+- ▁dreamed
+- ▁promptly
+- 냐
+- ▁mann
+- каз
+- ▁prior
+- ▁travelling
+- aro
+- ▁lace
+- ▁However
+- ▁suitable
+- ▁quelque
+- வி
+- '5'
+- ▁ray
+- bs
+- ▁gute
+- poli
+- ▁guy
+- known
+- ▁reception
+- ▁amused
+- ▁pronounced
+- ieron
+- chu
+- ▁providence
+- ▁reeds
+- ▁testimony
+- ▁yara
+- ▁ability
+- kh
+- ɣa
+- ▁sadly
+- ▁supplied
+- ▁proposal
+- ▁менән
+- ▁anchor
+- ▁murdered
+- ▁stretch
+- ▁lin
+- ▁enabled
+- ▁reckon
+- ▁Por
+- ▁Kaj
+- ▁prayed
+- ▁divide
+- ▁شما
+- ▁así
+- ▁ndi
+- ý
+- sp
+- ▁preparing
+- vent
+- ús
+- ▁owned
+- თ
+- ▁адзін
+- ign
+- ो
+- ▁solitude
+- hal
+- ▁mara
+- ▁shooting
+- ▁heels
+- ▁bundle
+- έ
+- ▁purse
+- দ
+- ▁scho
+- ▁nephew
+- އި
+- ▁quel
+- ▁gott
+- ▁nam
+- nken
+- ▁Gra
+- iĝas
+- horn
+- ▁speaker
+- ▁seize
+- ▁leaders
+- ▁absorbed
+- ▁anders
+- ▁colours
+- η
+- є
+- ▁arts
+- ▁repose
+- mund
+- ▁chez
+- tage
+- rte
+- tine
+- ▁dice
+- itaj
+- ▁neglect
+- سە
+- ▁garments
+- أ
+- down
+- bas
+- ▁submit
+- ▁mask
+- 보
+- 路
+- ▁hebt
+- ▁students
+- ▁push
+- фі
+- ▁jedoch
+- ▁fruits
+- même
+- ▁Bei
+- tos
+- oz
+- ▁surrender
+- ▁ihrem
+- ▁feature
+- ▁accused
+- ▁umwe
+- ▁sailors
+- ದ
+- ▁rwa
+- gil
+- 你
+- ▁ئۇ
+- rij
+- eck
+- ▁slo
+- thi
+- ▁быць
+- ▁kid
+- ▁decent
+- nig
+- ▁scorn
+- ▁spur
+- kura
+- ▁гэты
+- мат
+- ەوە
+- ▁chill
+- ▁hadden
+- ▁mem
+- ▁maison
+- pla
+- ▁자
+- ▁cabinet
+- ▁stomach
+- ▁같이
+- ▁responsibility
+- ví
+- ▁welke
+- ▁bran
+- ▁hammer
+- 이랑
+- ▁comparison
+- ▁undoubtedly
+- exp
+- ▁macht
+- ▁rice
+- ▁Nord
+- ▁wenig
+- ▁정도
+- umva
+- aha
+- hau
+- sses
+- ▁lively
+- ▁stre
+- ของ
+- ▁rushing
+- ▁alarmed
+- ▁sack
+- ▁esti
+- 이야
+- ▁narrative
+- essen
+- taba
+- ▁combination
+- ▁قا
+- weise
+- ▁obeyed
+- ー
+- ▁altijd
+- ▁Ali
+- ▁wherein
+- эр
+- tek
+- ▁paar
+- ▁waist
+- ▁alma
+- ▁occupation
+- டு
+- ▁Maria
+- ▁große
+- سى
+- say
+- ტ
+- ▁fare
+- ▁grows
+- ക
+- ▁hacer
+- ▁indicated
+- 실
+- ▁critical
+- пер
+- ▁goods
+- ▁spear
+- chel
+- ▁shepherd
+- ▁jury
+- ▁poder
+- 子
+- र्
+- ▁pursue
+- ▁Qui
+- ▁curse
+- る
+- نگ
+- ▁eby
+- ▁appointment
+- ál
+- ▁changing
+- lau
+- ▁yonder
+- ▁Aquesta
+- ▁trusted
+- icia
+- ▁colored
+- ▁After
+- ı
+- аз
+- ▁Таму
+- 신
+- ▁Art
+- ▁전
+- ▁eenige
+- fund
+- boat
+- hm
+- ▁nella
+- ify
+- ▁kur
+- ▁Era
+- ▁einige
+- ▁finest
+- ▁administration
+- tag
+- ▁stroke
+- six
+- ▁upward
+- ▁Ve
+- со
+- ese
+- का
+- ▁copper
+- vano
+- ▁bus
+- ▁trat
+- ▁gleam
+- 학
+- کن
+- ▁california
+- abwe
+- ▁wana
+- ▁valor
+- аны
+- ▁folded
+- ▁vanity
+- ▁trained
+- ▁debt
+- mani
+- ▁Fer
+- ▁seas
+- ong
+- ▁elegant
+- ▁napoleon
+- ▁summoned
+- ▁hombre
+- ▁thereby
+- ▁tim
+- lian
+- سا
+- ▁joan
+- ▁writers
+- ▁chairs
+- ▁fancied
+- ▁stopping
+- feld
+- ▁unlike
+- frag
+- ▁summit
+- ▁jewels
+- ി
+- ▁wax
+- ▁searching
+- ▁binnen
+- sur
+- 行
+- kk
+- ്
+- ▁pier
+- ▁Cett
+- ▁depends
+- ▁jove
+- ▁encounter
+- ▁paying
+- ▁chant
+- ▁grup
+- ыл
+- 월
+- ▁werde
+- ▁complex
+- ▁oncle
+- ▁individuals
+- ▁Cal
+- ژ
+- igu
+- வா
+- tego
+- ได้
+- jan
+- ▁brute
+- lat
+- ה
+- ▁trot
+- ▁verd
+- ▁하나
+- رە
+- вяр
+- ып
+- ▁romantic
+- ▁flames
+- ▁bun
+- ▁marquis
+- dies
+- ▁jan
+- ▁leads
+- ▁ek
+- ว่า
+- ▁troviĝas
+- ▁갈
+- enen
+- ▁depart
+- ▁climate
+- ци
+- nio
+- izo
+- kino
+- ▁tobacco
+- tege
+- hör
+- ▁niba
+- ▁봐
+- лен
+- ▁wool
+- ▁commonly
+- ▁seal
+- ▁그러면
+- ▁clara
+- qué
+- ▁poi
+- ▁whit
+- ▁commerce
+- foot
+- now
+- нан
+- ▁contre
+- ṭ
+- ▁lodge
+- ▁dashed
+- ▁rescue
+- ▁aquesta
+- baza
+- nant
+- ння
+- ▁ranks
+- ▁perish
+- ▁awoke
+- ▁dragon
+- ▁frei
+- ▁cream
+- ▁retain
+- த்து
+- ▁ganze
+- ▁syn
+- ▁som
+- ▁پ
+- ం
+- cca
+- ▁viele
+- ▁citizen
+- ▁Bru
+- ▁tune
+- master
+- ▁data
+- ▁lessons
+- ▁kinder
+- ▁lantern
+- ▁function
+- raga
+- ця
+- ▁warned
+- ▁cigar
+- ▁roused
+- ു
+- ▁moss
+- ▁diamond
+- ▁plunged
+- ить
+- ▁otro
+- ▁teatre
+- ▁pool
+- ald
+- lei
+- ▁available
+- 国
+- ▁association
+- ord
+- umi
+- ▁mano
+- ▁impatient
+- ▁response
+- ▁guilt
+- ▁Fo
+- ▁Pol
+- ▁faded
+- ▁lordship
+- ▁gewesen
+- divi
+- ▁swallow
+- mol
+- ▁blast
+- ▁parent
+- ▁recover
+- ▁sensitive
+- ▁blows
+- inda
+- cl
+- ▁liegt
+- ▁affect
+- ใน
+- ▁들어
+- ŝ
+- ▁independence
+- ▁roots
+- ▁tett
+- ▁brick
+- ▁income
+- lets
+- чын
+- ▁cigarette
+- ▁faut
+- ▁uza
+- қ
+- ▁licht
+- ▁visitar
+- ▁pic
+- ्
+- ▁tendency
+- ▁achter
+- ▁dumb
+- ▁caesar
+- ▁prima
+- nak
+- сць
+- ▁flushed
+- ▁revolver
+- cial
+- ▁살
+- ▁distinguish
+- ▁biblioteca
+- ▁britain
+- kem
+- ▁chinese
+- ▁igi
+- ▁schw
+- ▁fishing
+- ▁eagle
+- ▁cannon
+- ▁apply
+- ▁monster
+- iest
+- ▁об
+- 워
+- uer
+- ▁struggling
+- ған
+- ▁och
+- ▁뭔가
+- ▁geworden
+- ico
+- eli
+- ▁usted
+- ি
+- な
+- ▁graceful
+- ▁계속
+- ▁cour
+- ▁pou
+- ▁endless
+- ▁covering
+- rb
+- ▁adventures
+- ▁sue
+- ▁آن
+- tical
+- lk
+- ▁groote
+- ▁induced
+- ▁discourse
+- ▁York
+- ▁Ад
+- ▁beide
+- ▁painter
+- handel
+- ▁cad
+- 해서
+- ▁Don
+- ▁privilege
+- ۈ
+- ▁My
+- ▁sunlight
+- ▁til
+- னை
+- euse
+- ия
+- ▁hoping
+- jj
+- ▁qua
+- ▁apprehension
+- ▁og
+- ▁displayed
+- ▁glen
+- ▁tut
+- ▁norman
+- ۇ
+- ▁möchte
+- ▁costume
+- ▁materials
+- ▁historical
+- ▁byo
+- ▁scha
+- ▁villages
+- dzi
+- ▁ул
+- 种
+- ▁capture
+- ▁revelation
+- yne
+- merk
+- ▁raw
+- тым
+- ▁surrounding
+- ▁цяпер
+- ▁doubtful
+- стан
+- ▁injury
+- ▁jolly
+- ▁ersten
+- ▁Што
+- ▁kent
+- ▁feels
+- ▁aliaj
+- ▁contented
+- ▁ту
+- geben
+- ால்
+- اق
+- ▁happily
+- bonye
+- esp
+- ▁mutter
+- ▁nahm
+- ▁sunk
+- kommen
+- ▁mould
+- ▁bullet
+- ▁gross
+- odo
+- rina
+- 好
+- uje
+- uring
+- ▁University
+- sicht
+- ë
+- ▁schwer
+- ▁indifference
+- inde
+- ▁consisted
+- ▁pond
+- 금
+- 西
+- run
+- ▁revolt
+- ▁festival
+- ụ
+- ▁ample
+- ▁counted
+- rik
+- ▁mijne
+- 버
+- ssel
+- ▁yeux
+- igita
+- ول
+- ▁shouting
+- ▁preparations
+- ein
+- ▁delicious
+- வே
+- uren
+- ىنى
+- ▁printed
+- case
+- ▁altres
+- ▁lend
+- ▁mars
+- gura
+- 한테
+- eurs
+- ▁passengers
+- arbeit
+- ▁englishman
+- ▁Was
+- ี
+- phe
+- ▁newspapers
+- ē
+- ▁dried
+- ▁staat
+- ▁requires
+- сә
+- ▁mob
+- ▁hedge
+- ▁mono
+- pl
+- قى
+- ▁basis
+- ▁cara
+- ▁awakened
+- ގެ
+- ▁lent
+- ▁alien
+- gues
+- ▁pretended
+- ▁fred
+- ▁anxiously
+- 尔
+- ▁delle
+- ▁motionless
+- tete
+- ▁nyuma
+- esse
+- ▁educated
+- มา
+- ▁Sal
+- ▁metge
+- ն
+- ▁prominent
+- jun
+- ▁confessed
+- ▁trifle
+- ▁machte
+- ▁guerra
+- ▁ва
+- ▁아이
+- oy
+- ▁nda
+- ▁crimson
+- ▁havia
+- agenda
+- ▁emotions
+- ▁bomb
+- ▁namely
+- edifici
+- ▁reports
+- tief
+- ▁cub
+- ▁peaceful
+- ▁beggar
+- ▁bah
+- ▁걸
+- ▁combined
+- ▁torment
+- ▁sees
+- vier
+- ▁inquire
+- ▁moins
+- ▁attain
+- meister
+- ้า
+- ▁streams
+- hui
+- ▁shed
+- кан
+- ▁거지
+- zem
+- ▁careless
+- ws
+- ppe
+- ят
+- 下
+- ast
+- ▁andrew
+- ও
+- mmel
+- lands
+- ▁sana
+- ▁Zi
+- ▁blade
+- ▁loaded
+- mere
+- ▁шу
+- ▁recon
+- cou
+- gis
+- cinc
+- ▁pattern
+- คุณ
+- ▁despite
+- ▁pressing
+- ▁wid
+- ▁assistant
+- ▁partner
+- ▁joint
+- ней
+- ▁spin
+- ▁comrades
+- ▁Ac
+- ▁hideous
+- tail
+- ▁neue
+- ▁missing
+- lwa
+- ▁moderate
+- autre
+- ุ
+- ▁Et
+- 道
+- bridge
+- ▁toil
+- uro
+- ▁lecture
+- ▁Imana
+- ▁letting
+- ▁avoir
+- มี
+- ▁halt
+- ▁passionate
+- ា
+- ▁curiously
+- haus
+- ▁inevitable
+- ▁liebe
+- acions
+- gul
+- ▁questioned
+- ▁bargain
+- ▁якія
+- nez
+- ▁vivid
+- 상
+- ▁linen
+- wal
+- ▁frost
+- ▁admirable
+- ▁zeal
+- ddle
+- ▁Sam
+- gne
+- zug
+- ▁вер
+- eten
+- ▁leisure
+- usi
+- ▁allein
+- ▁upright
+- ▁import
+- ▁ashes
+- ▁december
+- ▁저
+- ▁velvet
+- bio
+- ▁colors
+- ▁disciples
+- ▁proposition
+- mata
+- ▁Cha
+- ▁corridor
+- ▁plays
+- ▁concert
+- nut
+- homme
+- ▁recorded
+- ▁heartily
+- ▁recognised
+- Al
+- ▁خود
+- ▁arrangements
+- ▁niece
+- ▁roger
+- aya
+- ▁অ
+- oi
+- ▁plane
+- guard
+- ▁germans
+- sol
+- ▁isabel
+- ▁без
+- ел
+- hill
+- ür
+- ▁restless
+- нне
+- от
+- ▁amazement
+- ▁Bir
+- ▁cliff
+- ▁liquid
+- tting
+- ▁disgrace
+- ▁introduction
+- ▁wages
+- ▁irregular
+- ▁bord
+- ▁shy
+- skin
+- ▁monument
+- most
+- ▁neglected
+- ▁darauf
+- vul
+- ▁fathers
+- ▁waved
+- အ
+- ▁claimed
+- ▁zit
+- ▁ladder
+- wyn
+- ▁occasional
+- rz
+- ▁footsteps
+- ites
+- ▁spy
+- ▁ал
+- ▁ĉiuj
+- ▁curtain
+- ظ
+- kat
+- ▁repair
+- ière
+- ▁neighborhood
+- ▁chu
+- rand
+- ▁hoog
+- rzy
+- ▁cho
+- ▁되
+- ▁oogenblik
+- ▁recalled
+- ▁limp
+- 我
+- ▁stake
+- ▁twilight
+- nsi
+- ▁embrace
+- hara
+- ▁rei
+- kna
+- ▁dug
+- mento
+- ▁없
+- quatre
+- ▁admiral
+- ▁Stadt
+- ▁sandy
+- shya
+- past
+- ▁geht
+- ▁lucky
+- lö
+- ▁cui
+- ▁simplicity
+- igt
+- ▁pont
+- 러
+- ▁followers
+- ▁guessed
+- ▁tranquil
+- ▁hon
+- ▁supplies
+- ▁thief
+- ▁reveal
+- ché
+- ▁sentiments
+- fri
+- xo
+- ச
+- ▁nka
+- ▁lloc
+- ▁How
+- ▁regards
+- ▁parole
+- ▁research
+- ▁Аднак
+- ▁anguish
+- ▁betrayed
+- ▁akk
+- ▁avuga
+- ▁amor
+- ▁harsh
+- pur
+- ે
+- tiv
+- 斯
+- ▁aya
+- ▁antaŭ
+- 운
+- ▁tragedy
+- ▁wander
+- ▁killing
+- ▁dense
+- pare
+- ▁admired
+- mate
+- ▁breathe
+- schrei
+- eco
+- reka
+- rib
+- vie
+- ▁ministers
+- ▁fortunately
+- L
+- hab
+- hang
+- tica
+- ▁نمی
+- ▁Santa
+- ▁wheat
+- ▁drift
+- ол
+- 식
+- tà
+- ▁nobles
+- ▁derived
+- ▁sem
+- ▁hinter
+- ▁entonces
+- tribu
+- ▁conceived
+- ▁pockets
+- ▁dif
+- ▁kommt
+- ▁prompt
+- ▁especial
+- λ
+- ▁inclination
+- ▁avant
+- asa
+- ▁hawk
+- ▁originally
+- ▁temperature
+- ▁moor
+- zna
+- idi
+- ▁те
+- дзя
+- ▁tables
+- ು
+- sem
+- amba
+- ר
+- ▁Que
+- ▁پر
+- ▁closing
+- ▁canvas
+- alo
+- ҙе
+- یت
+- ▁accounts
+- bble
+- ▁caution
+- ▁havis
+- ▁canoe
+- há
+- ▁Cu
+- ▁simon
+- ▁stores
+- ▁presents
+- ▁अ
+- ை
+- ▁deceived
+- ▁satan
+- ▁unfortunately
+- kuwa
+- ▁partie
+- ▁warriors
+- ▁것도
+- ▁create
+- ▁kal
+- ▁meiner
+- ▁jealousy
+- ட்ட
+- dru
+- ▁Tre
+- ▁어떻게
+- ▁voir
+- хо
+- ▁gallant
+- ▁étaient
+- ▁bells
+- ▁colony
+- obo
+- ▁behalf
+- ▁responded
+- ▁weep
+- bag
+- rau
+- ▁entitled
+- ▁immortal
+- ▁nerve
+- ▁peasant
+- oto
+- ▁ruins
+- raz
+- ▁riches
+- ▁corre
+- čč
+- ▁Ну
+- ▁destiny
+- ો
+- ▁assault
+- ▁statue
+- ▁Anda
+- ▁Ter
+- ▁ج
+- lings
+- ▁floating
+- ides
+- amour
+- ▁conquest
+- ▁Tio
+- ▁چه
+- ▁cheese
+- ▁convent
+- lash
+- ▁glasses
+- ▁wollen
+- ▁thirst
+- дә
+- ▁mole
+- ▁acknowledge
+- ▁qo
+- ▁எ
+- ▁trenta
+- ▁reverence
+- ян
+- ▁innocence
+- uye
+- indi
+- ▁dur
+- tisch
+- ▁cam
+- ▁mensch
+- ▁awkward
+- ▁над
+- ▁essay
+- হ
+- ▁добра
+- ukuri
+- ▁benshi
+- ▁ripe
+- ▁alexander
+- voj
+- ▁dabei
+- ▁pleasures
+- ▁surgeon
+- ▁selfish
+- ▁consult
+- ean
+- aries
+- ▁painting
+- ▁Amb
+- ▁بی
+- bl
+- ▁indicate
+- rom
+- ▁سر
+- ▁ஒரு
+- zira
+- ▁snake
+- ▁després
+- ▁wheels
+- ▁certainty
+- rira
+- ▁robin
+- 간
+- amos
+- path
+- ome
+- ▁tête
+- ▁pint
+- ▁acid
+- vr
+- ▁expert
+- ▁endeavour
+- stead
+- tors
+- ▁radio
+- rück
+- ▁juffrouw
+- ▁acknowledged
+- tò
+- ▁под
+- ▁january
+- bei
+- ▁kri
+- ху
+- ▁première
+- ▁obra
+- ▁کے
+- ▁When
+- nsen
+- ▁navy
+- ▁rief
+- ▁fires
+- ▁pis
+- ▁games
+- ▁kant
+- ▁Diumenge
+- ▁garrison
+- nica
+- ▁plaats
+- ▁vale
+- ▁Tri
+- ▁weit
+- ▁keeps
+- ▁planned
+- ▁forests
+- ங்க
+- aine
+- ▁expecting
+- kam
+- oma
+- ▁jail
+- dors
+- ▁americans
+- ▁vile
+- ▁То
+- ▁situated
+- ▁journal
+- ▁ibintu
+- rés
+- 于
+- ▁advised
+- ▁sultan
+- ▁plains
+- bung
+- ▁determine
+- ▁platja
+- ▁großen
+- vat
+- ▁raven
+- ▁시
+- ▁rond
+- ▁원래
+- или
+- ▁Са
+- ▁whither
+- рад
+- ▁julia
+- ▁amafaranga
+- ▁shower
+- pho
+- ▁forming
+- ን
+- ongo
+- piece
+- ▁heer
+- cin
+- ▁என்று
+- ▁recognition
+- ана
+- ugh
+- ▁admire
+- tain
+- пі
+- ▁jew
+- ▁Què
+- ▁earthly
+- ▁sla
+- lud
+- ▁shine
+- டி
+- ▁obscure
+- mittel
+- ▁sido
+- ▁conceive
+- ▁scream
+- ▁retire
+- vet
+- ще
+- ▁mutual
+- ▁boss
+- eld
+- ▁harmony
+- ▁drops
+- ▁hol
+- illo
+- ▁convert
+- ▁furnish
+- lip
+- ität
+- ▁aroused
+- ▁moeten
+- ▁yose
+- ▁wishing
+- ▁believing
+- 비
+- ash
+- ▁flour
+- ▁disorder
+- гы
+- uen
+- یان
+- ▁eta
+- ress
+- kolo
+- eks
+- ▁fatigue
+- ▁mansion
+- ▁meadow
+- menye
+- ▁convention
+- erie
+- ▁unwilling
+- ▁furious
+- ▁scarlet
+- ▁exile
+- ▁idol
+- ▁reproach
+- ▁있잖아
+- ▁wer
+- ▁spectacle
+- ▁violently
+- ▁tread
+- arre
+- ҡа
+- ngi
+- ▁grava
+- ▁stamp
+- ick
+- ▁warmth
+- 들이
+- sab
+- rou
+- ▁ross
+- ▁Г
+- 르
+- ção
+- The
+- ▁lief
+- ▁troop
+- ising
+- ▁Oni
+- ▁exquisite
+- ache
+- ▁monk
+- ▁mateix
+- ▁verge
+- ▁vilaĝo
+- eus
+- ▁س
+- voi
+- رى
+- duc
+- ▁가서
+- ▁warn
+- rab
+- ▁accompany
+- ative
+- ▁Eu
+- ▁encouraged
+- lies
+- ▁whistle
+- алі
+- pol
+- ī
+- lari
+- ги
+- ▁bond
+- ▁parting
+- ▁holland
+- zehn
+- ▁Paul
+- ▁lawn
+- ▁effective
+- bwira
+- general
+- ▁shakespeare
+- ▁erect
+- ▁два
+- dica
+- vio
+- ▁nicholas
+- ક
+- вала
+- ▁courts
+- rö
+- gehen
+- ▁butler
+- ▁sphere
+- ▁pensa
+- ▁league
+- ▁advantages
+- ▁Min
+- ė
+- ▁damn
+- shaw
+- zing
+- ▁tommy
+- ▁lasted
+- ▁gap
+- ▁satisfactory
+- ି
+- ▁Joan
+- ▁fund
+- ▁mast
+- нда
+- ▁tru
+- ▁attempts
+- uki
+- fulness
+- 以
+- ▁loyal
+- ▁corpse
+- ishi
+- ás
+- ▁cellar
+- знач
+- ▁eminent
+- ▁gasped
+- ▁mußte
+- ▁proceeding
+- ▁commune
+- ב
+- ▁convey
+- ▁fils
+- ▁uneasy
+- ▁assert
+- 문
+- ▁seats
+- ▁während
+- ▁okay
+- ▁geheel
+- ▁healthy
+- ▁button
+- ▁disaster
+- ▁딱
+- ▁media
+- ▁michael
+- pun
+- ▁demands
+- hor
+- ▁injured
+- 피
+- 家
+- west
+- ▁gerade
+- ▁개
+- ่
+- ▁Bra
+- krieg
+- ically
+- ▁ange
+- shuri
+- ▁records
+- ▁moeder
+- מ
+- ▁accomplish
+- ▁ellos
+- ▁danced
+- ▁criticism
+- ▁subtle
+- ▁emma
+- ▁Ber
+- ▁siege
+- ▁vit
+- ▁maj
+- lage
+- ▁institution
+- rad
+- ▁kommen
+- ologia
+- bles
+- vä
+- hung
+- ▁dost
+- ▁cathedral
+- ▁ciudad
+- ҙә
+- ▁кар
+- ▁sore
+- ▁monday
+- ▁cinq
+- ▁betray
+- ▁parto
+- ▁rocky
+- ▁breathed
+- ▁symbol
+- scher
+- kum
+- ů
+- 파
+- ▁bekannt
+- uɣ
+- ▁oui
+- ள
+- sä
+- مى
+- ĩ
+- hun
+- njye
+- uzi
+- ▁commit
+- ▁hoch
+- ried
+- ▁accord
+- ▁이거
+- ▁porch
+- ivo
+- ▁feathers
+- ▁motives
+- ▁mixture
+- ně
+- ▁persuade
+- ▁pose
+- ան
+- ▁tras
+- ▁bunch
+- ▁withdrew
+- ▁mevrouw
+- аш
+- ने
+- ▁gladly
+- ▁interval
+- 경
+- ▁twas
+- sal
+- ▁envy
+- ĝe
+- mpi
+- ▁spo
+- ▁усё
+- 으면
+- ▁mam
+- ▁poden
+- ▁hesitation
+- ▁rail
+- liness
+- ît
+- рым
+- ▁nova
+- ▁limits
+- ▁pacific
+- скай
+- ▁momento
+- ▁formation
+- rei
+- ▁کنم
+- ▁ninth
+- ▁Bal
+- ל
+- gado
+- estra
+- ▁maken
+- uel
+- star
+- ок
+- ▁mexican
+- sanga
+- ▁sally
+- ▁peril
+- reven
+- вод
+- ▁yea
+- ▁Sen
+- uv
+- ▁prosperity
+- ▁Ці
+- ▁chat
+- пы
+- fire
+- ▁wont
+- ▁mala
+- ий
+- ▁hither
+- ▁대
+- ▁faster
+- ▁breed
+- ▁otra
+- спе
+- ▁compr
+- سی
+- hana
+- ์
+- য়
+- ▁marvellous
+- ▁scar
+- ▁rigid
+- ▁polite
+- sko
+- bt
+- буд
+- ▁disagreeable
+- ▁undertake
+- ingham
+- ży
+- ▁sola
+- gging
+- dien
+- मा
+- lui
+- ege
+- ▁travelled
+- ▁crushed
+- eng
+- 래
+- 매
+- ▁deserve
+- ▁그래도
+- ▁lip
+- ▁varied
+- ▁temptation
+- tari
+- berry
+- ▁Dr
+- ▁agreement
+- yl
+- ▁pru
+- tial
+- aus
+- ▁wounds
+- że
+- কা
+- ▁이게
+- ڪ
+- ▁Han
+- ש
+- xed
+- ރު
+- ско
+- our
+- may
+- ▁있고
+- kuta
+- 名
+- ▁planted
+- rado
+- ▁greeks
+- ▁bars
+- ▁whoever
+- ▁akken
+- chy
+- tero
+- ency
+- ▁stare
+- fun
+- ▁depuis
+- tá
+- ▁print
+- ▁designed
+- κ
+- ɛa
+- tö
+- ▁проста
+- ▁gall
+- 市
+- ▁treaty
+- 으
+- aza
+- ifa
+- ▁landlord
+- ▁poc
+- ▁répondit
+- ▁interfere
+- ▁parallel
+- ▁chances
+- ந்து
+- ▁beings
+- ▁گ
+- 전
+- ual
+- ▁Th
+- ▁jackson
+- ޭ
+- ले
+- τ
+- انی
+- ▁advancing
+- ▁existing
+- ▁confirmed
+- haft
+- ▁Dissabte
+- zeug
+- ▁christianity
+- ▁stove
+- ▁app
+- 계
+- ▁dropping
+- ▁emily
+- stig
+- pens
+- vern
+- 점
+- ▁destined
+- ▁damage
+- nung
+- hard
+- ▁renewed
+- hood
+- ▁À
+- tima
+- äch
+- ▁예
+- ango
+- ▁zwar
+- ற
+- ▁rap
+- жу
+- ddy
+- ▁artificial
+- ▁jerry
+- ▁plusieurs
+- icht
+- ▁West
+- ▁canada
+- ▁frozen
+- ▁greeted
+- ▁gallery
+- ▁جا
+- ▁gor
+- ▁aveva
+- 거
+- ▁faced
+- ூ
+- ▁elbow
+- ▁comparatively
+- key
+- ▁discuss
+- ▁prof
+- ▁robe
+- eke
+- ▁Яны
+- ▁pulling
+- ▁elected
+- eze
+- aan
+- ▁offices
+- ▁kap
+- rne
+- ген
+- iled
+- ި
+- ▁Kar
+- ين
+- lee
+- নি
+- ▁abode
+- ▁dash
+- ▁preparation
+- ▁loudly
+- ▁riep
+- ▁Dar
+- ▁miracle
+- ▁reckless
+- ▁restore
+- ▁seves
+- awa
+- argent
+- шу
+- zak
+- ▁philosopher
+- iye
+- ▁entertained
+- borough
+- ▁quest
+- และ
+- 동
+- ▁attractive
+- graf
+- ▁mig
+- fus
+- ▁levi
+- яр
+- стра
+- pus
+- ρ
+- ▁additional
+- ▁slumber
+- ваў
+- клад
+- ias
+- ▁retorted
+- ▁wreck
+- ▁frontier
+- ▁adding
+- лё
+- ▁ฉัน
+- ণ
+- паў
+- ▁lovers
+- ▁тое
+- aquest
+- ▁stark
+- твор
+- gni
+- ▁exists
+- ▁instances
+- ▁discussed
+- figur
+- ▁duck
+- ▁он
+- ▁permet
+- ▁ky
+- hur
+- ▁todas
+- ▁sailing
+- ▁tenía
+- ▁associated
+- én
+- ▁harbour
+- ความ
+- gio
+- ▁descent
+- einander
+- ▁senator
+- ▁nan
+- ▁crest
+- ▁senate
+- erd
+- ▁eldest
+- को
+- ▁pris
+- nik
+- ▁blieb
+- ▁surprising
+- μ
+- ▁mère
+- colo
+- nc
+- 다가
+- ▁lords
+- 하게
+- ▁verses
+- ной
+- ▁drug
+- ▁rejoined
+- ▁dood
+- އް
+- quet
+- ▁eventually
+- ▁patch
+- ▁alten
+- agu
+- ▁personally
+- ▁punish
+- ▁Den
+- wij
+- ▁agitation
+- ▁ranch
+- ▁vriend
+- ▁icy
+- fur
+- ▁нават
+- elen
+- ▁fand
+- ▁hoop
+- ▁restaurant
+- ▁passages
+- ▁paradise
+- ▁haus
+- ▁اس
+- ▁pala
+- 成
+- ▁ardent
+- iche
+- band
+- ▁cage
+- ▁bulk
+- ▁cinquanta
+- 당
+- ▁ridge
+- ▁dale
+- ู
+- ▁attract
+- woman
+- ▁tube
+- bula
+- ogo
+- bach
+- fte
+- ▁tierra
+- ▁Fin
+- ▁ashore
+- ▁altered
+- ▁quien
+- ▁readers
+- ▁grip
+- '1'
+- ▁dwelt
+- ▁differ
+- ▁prudent
+- شت
+- ▁tenth
+- ோ
+- cult
+- ▁shells
+- nje
+- ▁acht
+- хі
+- ▁critic
+- ▁tlie
+- ▁stond
+- ▁бел
+- เ
+- ▁thumb
+- უ
+- kurikira
+- 三
+- মা
+- ▁pious
+- ▁santa
+- ▁secrets
+- loj
+- istes
+- ▁bonnet
+- ▁envelope
+- chan
+- цыі
+- ▁sincere
+- rod
+- eln
+- ▁glowing
+- iris
+- lern
+- ▁Bre
+- ▁Inter
+- ган
+- 선
+- ▁flies
+- ▁dank
+- ▁brutal
+- ells
+- ▁alt
+- ▁bata
+- π
+- ▁beds
+- bot
+- press
+- ▁consolation
+- trag
+- ibyo
+- ন্
+- rika
+- ń
+- ▁develop
+- ▁consist
+- ▁drowned
+- ▁Chi
+- ▁scott
+- ened
+- ▁secretly
+- ках
+- Ar
+- rim
+- ▁fünf
+- ▁madness
+- ▁Ten
+- ▁protested
+- ▁yielded
+- ▁sinking
+- ▁historia
+- ▁If
+- ▁problems
+- তা
+- ▁vingt
+- ▁witnesses
+- 었어
+- position
+- ▁protestant
+- ▁Land
+- part
+- lied
+- ▁landscape
+- ▁boxes
+- ▁recollection
+- ▁primitive
+- ск
+- ▁verw
+- nä
+- ▁ple
+- կ
+- ▁instruction
+- ▁pepper
+- ▁quaranta
+- ▁file
+- pass
+- ▁정
+- ▁kil
+- ▁shillings
+- ▁prejudice
+- ▁jongen
+- jwe
+- reu
+- 회
+- grad
+- ▁kor
+- kli
+- 用
+- ▁parson
+- ره
+- bbe
+- ▁سو
+- ▁clar
+- ٹ
+- dí
+- 后
+- ngira
+- ▁Яна
+- ▁baka
+- ▁personne
+- ▁chorus
+- pani
+- ▁longed
+- ▁بىر
+- mand
+- ▁cake
+- sty
+- ▁inspiration
+- ▁earliest
+- ▁mou
+- ▁perfume
+- ▁dell
+- ▁hostile
+- ▁py
+- ▁später
+- ▁elaborate
+- ▁lance
+- ▁iby
+- ▁rings
+- shyi
+- ▁twisted
+- ▁clasped
+- ▁spaniards
+- fil
+- ▁positively
+- gers
+- ▁yourselves
+- ▁cruelty
+- ধ
+- ▁essence
+- ▁disgust
+- dit
+- ▁justified
+- ▁ahora
+- ▁diu
+- ▁replaced
+- ▁sar
+- guer
+- ▁february
+- ▁neighbors
+- ▁tad
+- ▁meinen
+- ▁perception
+- ▁One
+- چى
+- ▁witch
+- ▁repent
+- gamba
+- ▁investigation
+- ▁upset
+- ▁inquiries
+- скі
+- him
+- ▁packed
+- ▁susan
+- ▁villain
+- class
+- mmi
+- kä
+- punkt
+- enza
+- ▁ти
+- ыт
+- ▁medio
+- ló
+- ▁Park
+- ▁disturb
+- ▁rever
+- ▁dave
+- ▁firing
+- ▁idiot
+- ▁shaken
+- 안
+- ▁promises
+- shaka
+- ▁decir
+- ▁hush
+- க்கு
+- ▁bé
+- 部
+- ▁harvest
+- cun
+- ▁걔
+- mili
+- ਕ
+- ▁хто
+- ▁quando
+- ▁errand
+- ▁beef
+- ▁consul
+- ▁memories
+- ▁wist
+- tim
+- ▁struggled
+- hil
+- pé
+- ▁inspector
+- gene
+- ەی
+- ▁luxury
+- 最
+- ▁Ба
+- ▁ausge
+- ▁obviously
+- ▁facing
+- ían
+- ▁strip
+- ▁lily
+- ▁வி
+- пан
+- ▁granda
+- ▁rou
+- зи
+- má
+- ▁complain
+- hole
+- ▁compliment
+- ▁preceding
+- ▁interpret
+- arm
+- ▁observations
+- ▁bending
+- kun
+- ▁swamp
+- excursió
+- ▁alchemist
+- ▁waving
+- гән
+- ரு
+- fühl
+- ages
+- ▁fragments
+- tom
+- ▁몇
+- ▁laten
+- anto
+- ▁openly
+- ▁kin
+- ▁ceiling
+- ▁shrill
+- ▁bout
+- ▁allí
+- rico
+- ▁realm
+- gor
+- ▁writes
+- ▁lots
+- ▁angle
+- iwa
+- abel
+- ▁conquered
+- 要
+- шә
+- ▁blown
+- maid
+- ▁Herr
+- eight
+- ▁boast
+- ▁arguments
+- кра
+- ▁arc
+- ▁Divendres
+- ▁judges
+- ▁establish
+- ▁wirklich
+- kulu
+- دار
+- ना
+- ▁boldly
+- ▁blanc
+- た
+- ▁habia
+- ▁version
+- ▁jou
+- ▁institutions
+- ebe
+- hus
+- ▁resources
+- ▁deer
+- ▁pencil
+- ül
+- ▁compare
+- rry
+- tick
+- ▁globe
+- ▁warrant
+- யா
+- aku
+- lles
+- stricken
+- کە
+- mita
+- ▁victor
+- ▁shan
+- tze
+- atory
+- ступ
+- ▁customs
+- জ
+- hir
+- 부터
+- валі
+- ▁인
+- ▁threatening
+- ▁neighbours
+- ▁greeting
+- تە
+- ▁mer
+- ▁dazu
+- mira
+- pia
+- dul
+- ▁grove
+- ▁gesicht
+- ▁È
+- ▁sweat
+- ಾ
+- ▁grandes
+- agira
+- ▁gossip
+- か
+- ▁میں
+- bari
+- ▁federal
+- ό
+- anda
+- schein
+- টি
+- ▁basin
+- ▁hans
+- ▁spr
+- ▁entertainment
+- tua
+- ▁heroic
+- ▁davon
+- ▁experiences
+- obu
+- rê
+- ര
+- ▁uku
+- ӓ
+- ր
+- ▁correspond
+- ▁sickness
+- ут
+- luc
+- ▁proceedings
+- ▁internal
+- тын
+- riko
+- ▁propos
+- ▁poets
+- pr
+- ▁Oku
+- ▁slim
+- 부
+- ▁springs
+- ▁pleasing
+- imana
+- ▁rag
+- ▁dislike
+- ▁abraham
+- ▁barbara
+- tina
+- لىق
+- mber
+- ▁agnes
+- ▁exclamation
+- ▁crimes
+- ▁Бо
+- ▁harder
+- ▁jacket
+- polis
+- ▁کر
+- ▁blowing
+- ▁observing
+- ▁introduce
+- 방
+- ▁molly
+- لو
+- ▁saxon
+- ▁puede
+- taka
+- ▁halted
+- gó
+- ▁backward
+- ▁sketch
+- өн
+- ▁juan
+- ▁newly
+- ▁torrent
+- cà
+- 히
+- leid
+- ▁той
+- ▁arrow
+- scal
+- нымі
+- wl
+- ön
+- ois
+- ыз
+- ▁respects
+- ▁Esta
+- bour
+- ▁Fue
+- ▁arrange
+- ▁plato
+- ▁legend
+- gaan
+- ▁champion
+- ▁beste
+- し
+- ▁colonies
+- ▁sha
+- ▁temporary
+- gir
+- ▁Tan
+- ▁lingvo
+- ▁vive
+- pal
+- dle
+- ▁dangers
+- ▁downstairs
+- ▁kwenye
+- roma
+- medi
+- anzi
+- ▁answering
+- ▁thanked
+- chester
+- ▁conference
+- ▁widely
+- azione
+- gol
+- ▁Med
+- ▁tenir
+- ▁industrial
+- ರ
+- ▁đ
+- ▁faintly
+- ▁tossed
+- ▁wal
+- ▁fru
+- ▁medium
+- ▁denken
+- ▁spared
+- مت
+- deck
+- ▁crazy
+- ▁boom
+- ▁corners
+- там
+- ▁학교
+- ▁dieu
+- dhi
+- pos
+- ▁terribly
+- ▁access
+- ▁tho
+- maal
+- ▁intercourse
+- ▁code
+- ▁nein
+- ▁attained
+- ▁payment
+- ▁werk
+- pac
+- ▁perpetual
+- ▁comrade
+- 事
+- unter
+- ierte
+- مە
+- ▁tracks
+- ▁checked
+- ▁johnny
+- ▁некалькі
+- pence
+- пра
+- шым
+- dress
+- ▁roi
+- ár
+- ▁jonge
+- bikorwa
+- ▁cash
+- فت
+- тай
+- ym
+- ாக
+- ▁dollar
+- ▁kubera
+- цэн
+- ▁rom
+- ▁rejoice
+- ▁Mariya
+- ▁mix
+- mala
+- ▁gaf
+- ▁fortnight
+- ▁crisis
+- ҥ
+- ▁artillery
+- шча
+- ▁hail
+- pet
+- ▁abo
+- 法
+- J
+- ▁шо
+- ▁tiger
+- fish
+- ▁signor
+- ▁honourable
+- stellung
+- avu
+- elijke
+- ▁horrid
+- ▁terrace
+- ▁seeming
+- ▁tat
+- oka
+- stor
+- ▁impatience
+- 교
+- ▁bind
+- ▁pad
+- ▁multaj
+- り
+- ▁nowhere
+- rug
+- ▁wasted
+- ▁testament
+- ▁mainly
+- ▁estan
+- ▁boil
+- ▁schnell
+- ▁sim
+- ▁gek
+- اف
+- ▁kya
+- ▁companies
+- ▁Che
+- tif
+- ▁catching
+- 과
+- ▁clung
+- kur
+- meɣ
+- ▁travers
+- став
+- natured
+- ▁improved
+- ▁Wer
+- 영
+- prob
+- kas
+- ▁vere
+- ▁vigorous
+- ▁fir
+- nh
+- ▁timid
+- も
+- ▁slaughter
+- yobo
+- ▁neighbour
+- cole
+- ▁бир
+- avi
+- ▁lowered
+- ▁highway
+- ▁país
+- 個
+- ▁excess
+- ▁rational
+- cup
+- griff
+- cant
+- ކު
+- lima
+- ▁estos
+- ▁severely
+- ▁affectionate
+- genommen
+- hé
+- rita
+- ▁disc
+- tivo
+- zá
+- ▁strictly
+- ▁advise
+- ▁pia
+- ▁baza
+- сти
+- ▁tomorrow
+- ▁solemnly
+- ▁defense
+- ▁kamer
+- ▁background
+- arme
+- ▁studio
+- ų
+- ▁mexico
+- ▁bot
+- rop
+- vien
+- ▁link
+- nek
+- ком
+- ▁gulf
+- ▁petition
+- ▁remembrance
+- ▁tremble
+- ària
+- gga
+- peri
+- ▁Ob
+- ▁día
+- ર
+- ▁Ati
+- cura
+- ▁unua
+- ▁nut
+- ▁спа
+- ▁sino
+- ock
+- র্
+- ▁amusing
+- ▁Ты
+- ▁judged
+- ▁Iki
+- ▁goddess
+- ▁seul
+- ▁economic
+- ▁cual
+- ▁collect
+- ▁ascend
+- eres
+- ony
+- ▁stil
+- ▁pomp
+- komeza
+- oo
+- вар
+- ▁kopf
+- ▁scandal
+- ▁Sud
+- ▁ville
+- ▁National
+- ▁abuse
+- cyo
+- ▁чым
+- ▁받
+- ▁defeated
+- istoj
+- ▁Ven
+- ន
+- ▁Мо
+- ▁부
+- 街
+- ▁suspicious
+- ▁lieu
+- ▁formidable
+- ▁supposing
+- ▁laat
+- ▁flora
+- ▁먹
+- ▁boot
+- orient
+- ▁executed
+- пи
+- ▁speedily
+- ▁lei
+- ▁mundo
+- ▁pony
+- iers
+- ▁먹고
+- ▁whereupon
+- әй
+- ்
+- rada
+- keeper
+- ▁leicht
+- ▁mouse
+- ndu
+- stock
+- ун
+- մ
+- vit
+- ▁slay
+- aĵoj
+- ▁guarded
+- ئى
+- ి
+- ▁thor
+- ▁survey
+- ▁Nta
+- ▁harbor
+- gala
+- In
+- ▁machinery
+- 위
+- ▁liquor
+- ▁schu
+- ▁generations
+- ▁endlich
+- از
+- 天
+- 갔
+- ▁provinces
+- ▁estava
+- ▁prevailed
+- ▁lifting
+- ▁rebel
+- anta
+- ▁staying
+- ▁holiday
+- ▁Col
+- ▁kiuj
+- ▁años
+- aud
+- kiza
+- ▁swim
+- ▁barrier
+- unda
+- ▁Ch
+- ▁deliberately
+- foto
+- ▁deposit
+- ▁Auch
+- ▁procure
+- ▁hinder
+- 'yes'
+- ▁thoughtful
+- ▁tempest
+- ▁bloody
+- tano
+- oir
+- ▁bolt
+- ▁learnt
+- open
+- ▁shared
+- mur
+- ▁Пры
+- ▁naval
+- ▁vivo
+- ▁ар
+- ▁declined
+- มัน
+- ▁confirm
+- वा
+- ҳа
+- ▁diamonds
+- mw
+- vê
+- ▁স
+- altra
+- ▁confederate
+- staan
+- 小
+- chin
+- ▁elfu
+- ▁museum
+- เขา
+- ific
+- ▁staircase
+- ▁olive
+- ▁hasn
+- ▁jimmy
+- хә
+- ▁holes
+- ame
+- ▁unity
+- ▁swell
+- 物
+- ▁protected
+- An
+- eme
+- 作
+- ▁bodily
+- court
+- ▁ufite
+- lab
+- ▁glancing
+- ▁vista
+- sible
+- reg
+- ▁wealthy
+- ▁Estas
+- wit
+- ▁significant
+- koj
+- made
+- ї
+- pers
+- ▁cable
+- ▁avoided
+- ▁amidst
+- ▁klein
+- ▁arab
+- ▁marsh
+- ▁willingly
+- ▁thereof
+- ▁dismissed
+- ు
+- ▁allah
+- ▁winding
+- ▁marvel
+- boo
+- iden
+- cate
+- sprechen
+- banga
+- ▁dalla
+- ▁significance
+- ▁groot
+- ▁aboard
+- bird
+- ▁client
+- ▁avail
+- 관
+- tien
+- ▁дзе
+- یه
+- χ
+- ▁threshold
+- ▁inward
+- yar
+- moj
+- ost
+- nent
+- ▁helping
+- ▁세
+- ▁shores
+- lde
+- kay
+- ത
+- ା
+- ▁Os
+- ▁kleinen
+- ▁edith
+- issa
+- kore
+- ▁swing
+- meze
+- ▁afforded
+- ▁xwe
+- ▁thankful
+- gani
+- ū
+- ▁placing
+- ▁disguise
+- mica
+- ▁desirable
+- ▁frankly
+- ▁tage
+- ▁Ort
+- ▁visits
+- ▁famille
+- ▁tenia
+- ▁ragged
+- gué
+- ksi
+- ▁deemed
+- ▁declaration
+- gari
+- ▁dy
+- gende
+- ▁edi
+- ե
+- ▁brows
+- ▁erste
+- ▁celui
+- ▁twi
+- ▁armen
+- ▁Uganda
+- ▁fabric
+- range
+- ▁babe
+- riye
+- ▁sou
+- ▁мае
+- ▁fragte
+- ▁نا
+- load
+- ▁kubona
+- fed
+- ▁autobús
+- ▁ann
+- ▁consented
+- ▁swore
+- ▁searched
+- ▁ape
+- pis
+- ત
+- ▁hearth
+- ▁commands
+- ▁bronze
+- ▁abundant
+- ▁goin
+- ▁pall
+- ▁headquarters
+- ▁barely
+- ▁wildly
+- aca
+- ▁devant
+- ▁Europa
+- ▁encountered
+- ▁seconds
+- main
+- spec
+- ▁telegram
+- ▁wilson
+- ім
+- ▁improvement
+- ެ
+- ▁creep
+- ਸ
+- gestellt
+- por
+- ▁boiling
+- pose
+- äng
+- ▁appreciate
+- ▁otros
+- ҙы
+- ▁prend
+- ordin
+- bbi
+- ▁olu
+- ▁wanjye
+- amazi
+- ӑ
+- maker
+- ▁frederick
+- itude
+- ▁engineer
+- gent
+- оп
+- ▁mme
+- ▁ojos
+- ▁web
+- 之
+- ្
+- ▁pens
+- ▁punished
+- ▁frightful
+- വ
+- dina
+- stained
+- ▁después
+- lag
+- 소
+- ▁awe
+- ▁oriental
+- urt
+- igheid
+- ▁jeder
+- ▁Über
+- ▁돈
+- ense
+- லை
+- ▁gru
+- ▁retained
+- ▁neen
+- ▁sens
+- ▁mama
+- ▁confident
+- ▁cerca
+- ▁debate
+- ▁두
+- rok
+- ▁nawe
+- ▁bene
+- 他
+- ▁remainder
+- ▁questo
+- ▁picturesque
+- toria
+- ▁wrist
+- stellen
+- ▁louise
+- jye
+- ▁volk
+- ▁filling
+- ▁offended
+- ▁hearty
+- ▁shallow
+- ▁modo
+- ▁prudence
+- geni
+- ▁samuel
+- emba
+- тары
+- ▁officials
+- ▁trouve
+- ▁commanding
+- вая
+- ▁alta
+- пол
+- ▁walks
+- dik
+- illes
+- რ
+- sburg
+- dur
+- ▁мала
+- kwiye
+- ▁suite
+- މަ
+- away
+- دىن
+- ▁woke
+- stel
+- ▁natur
+- ▁siempre
+- ▁affections
+- än
+- ▁induce
+- ▁어디
+- nner
+- cour
+- pli
+- ▁таго
+- pic
+- ▁brains
+- ड
+- ▁neighbor
+- ép
+- 山
+- alt
+- ▁Després
+- يى
+- ĉa
+- ▁vielleicht
+- ▁Та
+- ▁successor
+- ▁unsere
+- ▁flowing
+- ή
+- мов
+- ▁jet
+- erte
+- ʻ
+- ▁comprehend
+- ▁tempted
+- ▁childish
+- ñ
+- ▁addressing
+- kra
+- 方
+- ▁management
+- ▁incapable
+- fang
+- ▁coward
+- ▁liber
+- ▁conspicuous
+- glass
+- ھا
+- ▁날
+- ▁visiting
+- ▁sistema
+- گی
+- ▁되는
+- ता
+- ▁curtains
+- ▁earn
+- ▁möglich
+- ▁challenge
+- ვ
+- ▁Lia
+- ▁ambassador
+- ▁experiments
+- ▁purely
+- ▁yang
+- كى
+- ▁kick
+- ▁thorough
+- ▁heroes
+- ▁klar
+- ▁byinshi
+- ▁нам
+- ▁willen
+- ▁bestowed
+- ▁vya
+- ▁preacher
+- ▁شد
+- ị
+- ▁unbe
+- нам
+- ▁relatives
+- ĕ
+- ▁traces
+- ▁desolate
+- ĝ
+- ât
+- ▁diversaj
+- xe
+- ▁reaction
+- reb
+- ▁David
+- ssi
+- ın
+- ▁familia
+- entend
+- रा
+- ▁모
+- 去
+- ▁images
+- ого
+- itt
+- фе
+- vidi
+- singa
+- ▁eso
+- ար
+- ungs
+- kü
+- ▁pent
+- geven
+- rr
+- ▁radi
+- ▁bras
+- ▁стра
+- kara
+- legen
+- ▁Wal
+- ▁zéro
+- ά
+- called
+- ▁бол
+- ▁jours
+- піс
+- ▁гу
+- ▁үҙ
+- ▁artistic
+- zie
+- ▁enjoying
+- ▁reasoning
+- ▁сам
+- ▁wan
+- eṛ
+- ▁personality
+- ▁deserved
+- ▁ئې
+- ▁occupy
+- chair
+- gü
+- ▁Е
+- ▁measured
+- ▁favorable
+- ▁beim
+- ▁canon
+- ▁Ф
+- baar
+- rid
+- ▁harp
+- dwa
+- ligi
+- ▁fiery
+- ▁blanche
+- ▁expectation
+- kehr
+- ▁ас
+- ▁해야
+- 会
+- stin
+- haired
+- ▁greatness
+- meter
+- igung
+- ▁correspondence
+- rera
+- жо
+- ▁twin
+- ▁alliance
+- ▁sian
+- ▁Però
+- klo
+- ▁bitterness
+- ▁mabel
+- ▁harold
+- ware
+- ▁snapped
+- با
+- ▁extensive
+- ▁besser
+- ▁мя
+- ▁dusk
+- yez
+- ▁traitor
+- ▁sept
+- jas
+- ▁geist
+- ▁menos
+- ▁Això
+- ▁appropriate
+- ▁diana
+- ▁select
+- ▁considerably
+- ▁attorney
+- ▁verk
+- fahren
+- ▁hind
+- ▁scotch
+- ▁oxford
+- arch
+- Vo
+- ĝo
+- ▁calculated
+- ▁الْ
+- zal
+- ▁diverses
+- ▁Калі
+- ▁bump
+- لە
+- lica
+- ▁policeman
+- ▁seixanta
+- ▁alert
+- fla
+- ▁carr
+- mara
+- 겠지
+- ▁session
+- ści
+- ▁ви
+- teur
+- ▁shrewd
+- ots
+- ▁Iyo
+- ▁vault
+- ▁sufferings
+- sini
+- öl
+- ▁missus
+- onge
+- ▁uta
+- ▁Dijous
+- уа
+- ▁Ed
+- ▁manuscript
+- ▁representative
+- ▁پا
+- ▁kie
+- ▁ivan
+- ▁monkey
+- ▁parlour
+- imwe
+- som
+- ▁Jean
+- ▁phenomena
+- kor
+- ▁noche
+- ▁erra
+- ▁therein
+- sore
+- aires
+- ▁amazed
+- ▁servir
+- لار
+- halt
+- ▁від
+- տ
+- ▁lugar
+- ▁Ко
+- ▁tante
+- ▁behaviour
+- ▁clan
+- ying
+- udo
+- ला
+- ▁Sp
+- ▁seit
+- scribe
+- ▁revel
+- ester
+- ▁fright
+- mura
+- চ
+- ▁deprived
+- ▁quan
+- kari
+- 分
+- эн
+- ▁teachers
+- ▁cheerfully
+- ych
+- ▁prairie
+- ▁constructed
+- stat
+- ▁observer
+- dim
+- ▁gravity
+- 唔
+- ▁temples
+- ▁bestow
+- ña
+- aven
+- ရ
+- ▁robbed
+- cles
+- 카
+- ▁хо
+- ▁grin
+- awo
+- ▁whale
+- ▁dishes
+- ▁donne
+- вед
+- ▁sarah
+- ▁elsie
+- ▁graf
+- leigh
+- ▁fortress
+- ▁있는데
+- ▁surroundings
+- ▁гэтай
+- viv
+- ova
+- ไ
+- மா
+- ת
+- ▁limb
+- ▁باش
+- ▁terrified
+- rang
+- lation
+- ▁Tro
+- ▁parti
+- ▁goal
+- schuldig
+- ▁tiuj
+- ▁plum
+- ▁dreaming
+- 으니까
+- onne
+- 自
+- ▁schien
+- tenant
+- ▁lieb
+- даў
+- kula
+- ી
+- ▁injustice
+- ▁andern
+- ▁remedy
+- ▁saith
+- pid
+- ▁treasures
+- рә
+- reba
+- ▁asia
+- ▁negative
+- нов
+- '""'
+- ▁happening
+- oza
+- ▁actor
+- ▁frances
+- cloth
+- ләр
+- ▁rash
+- кты
+- ▁matt
+- ▁serving
+- kiwa
+- ▁impatiently
+- ▁uncomfortable
+- گو
+- ▁races
+- ▁lia
+- antoj
+- ▁dramatic
+- ▁fiction
+- ▁diversos
+- ▁dew
+- hod
+- maga
+- ▁ikintu
+- ▁comte
+- ▁uwe
+- ω
+- нні
+- ▁esteem
+- nah
+- ▁invented
+- ▁bod
+- ender
+- ▁rwose
+- ▁unjust
+- ▁atlantic
+- omo
+- ▁worried
+- ▁draught
+- ▁recollect
+- kya
+- ੀ
+- ▁haunted
+- ▁extend
+- ▁Bur
+- isten
+- ▁parece
+- point
+- undi
+- үү
+- ▁зна
+- ▁20
+- ▁throng
+- ▁autres
+- ▁chariot
+- ▁boven
+- ▁amy
+- გ
+- rien
+- franc
+- iq
+- بوو
+- ่า
+- irwa
+- ics
+- ▁confi
+- نے
+- 강
+- ▁nick
+- ander
+- ▁stato
+- ▁hause
+- tric
+- ▁bac
+- ▁Tanzania
+- ▁thrill
+- ▁مو
+- ▁hog
+- ▁viz
+- ▁배
+- erung
+- ▁hue
+- ▁cultivated
+- unga
+- 성
+- nti
+- তে
+- ▁Tamen
+- ۋا
+- ▁zelfs
+- ization
+- ▁què
+- ffen
+- ereza
+- ▁estar
+- ▁aza
+- насць
+- ▁aged
+- هی
+- ▁specially
+- ▁pupil
+- sanzwe
+- ▁sweetness
+- mission
+- ▁briefly
+- из
+- ▁Eine
+- 산
+- ▁які
+- pole
+- ▁angrily
+- hof
+- nud
+- ▁ganzen
+- ection
+- ▁kitty
+- tangira
+- nette
+- ▁fiercely
+- ▁solche
+- łem
+- ▁watu
+- ▁яму
+- ▁Paris
+- ▁France
+- ▁tame
+- 크
+- ▁augenblick
+- 여
+- rique
+- lev
+- ▁eighteenth
+- nem
+- ▁wissen
+- ▁공
+- 同
+- ▁traffic
+- ▁smallest
+- ▁doll
+- rier
+- ▁تە
+- ▁negroes
+- ▁hired
+- ▁charlotte
+- არ
+- ground
+- eaux
+- டை
+- นี้
+- ▁victims
+- வை
+- ▁cou
+- vic
+- ▁disappear
+- 발
+- ▁yake
+- ▁fortunes
+- emp
+- ▁rosy
+- ▁compte
+- ▁genera
+- tour
+- ▁barren
+- 南
+- ▁invent
+- wacht
+- ▁пла
+- ▁youthful
+- شی
+- tale
+- ▁rider
+- ▁دوست
+- ▁subsequent
+- ▁workers
+- ▁pavement
+- ▁blaze
+- ▁Teil
+- kaba
+- െ
+- bris
+- neuf
+- inc
+- ▁бай
+- сем
+- ▁erected
+- ▁pul
+- ▁faculty
+- ▁stirring
+- 本
+- ▁bewildered
+- ▁resemblance
+- Ma
+- ▁nieder
+- தா
+- ▁dreary
+- wydd
+- ▁overhead
+- ▁vera
+- ▁triste
+- ▁чалавек
+- א
+- ▁steht
+- ▁آیا
+- ▁turkey
+- ▁sabe
+- ▁victoria
+- ▁innumerable
+- ▁London
+- ▁madre
+- ▁celle
+- ▁commence
+- glich
+- ▁tribunal
+- nus
+- ▁swords
+- chter
+- ▁cargo
+- kri
+- ▁fain
+- ▁tracta
+- ▁keys
+- прав
+- ▁hurrying
+- ▁йо
+- ▁Ва
+- ktor
+- ▁veins
+- ство
+- ▁gihugu
+- tekereza
+- ▁unser
+- ▁doubts
+- ▁dessen
+- ▁Gre
+- ▁persona
+- ▁zona
+- ▁kuvuga
+- ▁reserved
+- ▁Many
+- ▁rebellion
+- ▁substantial
+- neɣ
+- ske
+- ▁inama
+- ▁кра
+- ▁Yu
+- oth
+- ▁skirt
+- bazo
+- ▁heraus
+- ▁complexion
+- imba
+- ▁City
+- 区
+- act
+- menya
+- ać
+- para
+- iĝo
+- lumin
+- mora
+- ▁stately
+- ▁forbidden
+- ▁axe
+- otto
+- asen
+- sobola
+- ۔
+- ▁literally
+- ▁needle
+- ең
+- кон
+- тыр
+- ▁abundance
+- roy
+- 州
+- ▁sermon
+- ▁keinen
+- rán
+- ▁dacht
+- ▁elephant
+- ▁centro
+- ▁sails
+- culture
+- ▁uburyo
+- ▁Tor
+- ▁އަ
+- ▁financial
+- اش
+- ▁yabo
+- atge
+- ▁doubted
+- rè
+- ▁buiten
+- ▁tank
+- ▁Neu
+- ▁cooking
+- ▁unit
+- ▁suited
+- сна
+- ▁floated
+- يا
+- ನ
+- ▁rows
+- bili
+- ▁hip
+- کی
+- ▁potatoes
+- ▁cavern
+- ▁fowl
+- éc
+- ▁halb
+- rius
+- ▁nancy
+- nyu
+- ▁Port
+- ▁reste
+- ▁nomo
+- υ
+- ▁그랬
+- tuma
+- tun
+- zzo
+- rag
+- ▁remorse
+- ▁relieve
+- ্যা
+- ischer
+- în
+- isation
+- ▁unnecessary
+- ▁hizo
+- stadt
+- тур
+- ນ
+- ▁illusion
+- ний
+- arra
+- ▁herbert
+- ▁nuit
+- ▁panic
+- ▁geb
+- 었
+- ▁assigned
+- 新
+- ▁bits
+- ▁separation
+- ▁это
+- bwe
+- ▁accent
+- ▁사람이
+- ▁bonne
+- ders
+- ▁거의
+- ▁steward
+- ▁hurriedly
+- ▁entertain
+- ▁Gen
+- ▁그치
+- ▁resta
+- equip
+- ▁quina
+- ân
+- 체
+- нем
+- ▁skies
+- uten
+- ▁comb
+- ▁echoed
+- ▁només
+- ▁bol
+- ▁regularly
+- nini
+- ▁mute
+- ▁resolve
+- nine
+- ▁hejuru
+- ▁될
+- ▁Tie
+- ▁amiable
+- ▁kimwe
+- enfant
+- ▁tots
+- ▁converted
+- ▁studying
+- ン
+- ▁tusschen
+- َّ
+- ▁towers
+- dus
+- ата
+- že
+- dou
+- pra
+- ddwa
+- praw
+- ▁我
+- ▁francisco
+- ▁onward
+- ▁kuva
+- ▁drie
+- ▁ни
+- ▁setanta
+- ына
+- ▁уз
+- ▁split
+- ▁Als
+- enti
+- loo
+- rò
+- ▁sire
+- posa
+- ▁بار
+- ▁rip
+- ▁gilt
+- ▁iri
+- ▁thyself
+- ▁mah
+- ▁assez
+- ▁skull
+- ▁preach
+- ▁traveller
+- ▁bills
+- ▁urge
+- ▁snap
+- 양
+- 會
+- agit
+- ▁Mil
+- ▁sup
+- ▁ph
+- ▁remembering
+- пад
+- ▁poems
+- بی
+- ▁mucho
+- beera
+- ▁cautiously
+- idade
+- cept
+- ▁보고
+- ▁rubbed
+- ▁Sur
+- ▁dwarf
+- ▁chains
+- ▁fundamental
+- ▁போ
+- ު
+- ▁beam
+- ▁invariably
+- ▁woo
+- ▁bernard
+- раў
+- equi
+- ▁mnie
+- 더라
+- jó
+- ▁doom
+- ▁shops
+- liegen
+- shot
+- ▁crash
+- uß
+- гьы
+- ▁같은데
+- ▁martha
+- ▁compass
+- ▁좋아
+- ching
+- stehen
+- ▁sob
+- ▁tú
+- ั
+- ▁сама
+- ▁muse
+- ▁exchanged
+- ulu
+- نه
+- ▁conf
+- sea
+- ▁brit
+- ▁gast
+- rir
+- ▁miriam
+- ▁Ang
+- ▁convince
+- ▁tenen
+- ୍
+- ब
+- ▁saviour
+- ▁muntu
+- iwe
+- ▁conveyed
+- ▁yiwen
+- ให้
+- even
+- ▁ancestors
+- ▁fried
+- ары
+- ▁Firefox
+- ▁encuentra
+- innen
+- ▁multe
+- seven
+- श
+- ▁encara
+- ▁jewel
+- ▁ئەو
+- ▁durante
+- ▁elect
+- ▁bands
+- خت
+- ▁thoughtfully
+- 오
+- ▁republican
+- yle
+- ▁merchants
+- ▁purchased
+- ▁bak
+- ▁trousers
+- ▁tuta
+- ▁courtesy
+- ▁weinig
+- char
+- ▁superstition
+- ▁vuitanta
+- ▁friday
+- ▁agents
+- ▁chicago
+- ▁tragic
+- ▁awhile
+- ▁span
+- ▁lowest
+- sili
+- limi
+- ▁avond
+- ೆ
+- ▁Gran
+- 근
+- odd
+- ▁invite
+- ▁kas
+- ▁sublime
+- ▁많
+- pier
+- ▁gehört
+- ▁swallowed
+- ▁ус
+- ல்ல
+- schi
+- лап
+- ▁answers
+- ▁warrior
+- ▁temperament
+- ▁neces
+- ▁deceive
+- nwa
+- ▁infinitely
+- ▁ac
+- ▁provision
+- uche
+- ▁fourteenth
+- ▁dealing
+- ▁alex
+- ▁maxim
+- ▁Nda
+- لَ
+- ▁opportunities
+- ▁zyn
+- ▁edition
+- ▁salon
+- 主
+- ▁unworthy
+- schap
+- ▁eagerness
+- ▁suspicions
+- ▁United
+- 작
+- wali
+- ▁swinging
+- ▁vacant
+- ▁catherine
+- ▁hesitate
+- ▁allem
+- യ
+- лет
+- cada
+- duka
+- ▁omdat
+- ayi
+- ▁ber
+- ▁mysteries
+- ▁failing
+- ezo
+- ▁buffalo
+- sley
+- ▁pueblo
+- ▁adult
+- 公
+- ▁recommend
+- ▁Quina
+- ▁voix
+- treten
+- ▁incredible
+- ▁wretch
+- sail
+- ▁video
+- child
+- what
+- ад
+- hut
+- ates
+- ▁hora
+- ▁rero
+- gos
+- ▁decay
+- ަ
+- Imana
+- ▁애들
+- கி
+- ▁dio
+- ▁baker
+- ▁grasped
+- ▁bureau
+- ▁levant
+- ▁unseen
+- ▁estimate
+- ▁alsof
+- uh
+- ▁맞
+- ▁charlie
+- ▁tenderly
+- ▁eer
+- ▁tire
+- ▁noranta
+- organiz
+- ▁insects
+- ▁advocate
+- 时
+- ▁جو
+- ▁clergyman
+- ▁petite
+- ▁hecho
+- baka
+- ▁pagan
+- ù
+- ▁것
+- lect
+- ▁glittering
+- ▁eternity
+- ▁knelt
+- ▁identity
+- ▁pupils
+- ▁climbing
+- ▁rim
+- aff
+- ▁jungle
+- ▁Une
+- ▁libro
+- ▁ringing
+- ▁publication
+- ▁bigger
+- ▁epi
+- ▁sprak
+- 北
+- кла
+- chant
+- ▁suspended
+- ▁spe
+- ▁연
+- လ
+- ▁cats
+- ▁fitting
+- ▁pag
+- ▁자기
+- amategeko
+- ▁conclude
+- ▁Mais
+- шка
+- ▁muscles
+- linda
+- 也
+- prov
+- ▁dependent
+- ▁screamed
+- ▁hace
+- ▁Vol
+- ▁drum
+- 来
+- ▁conce
+- 特
+- ▁specific
+- ▁deaf
+- ▁ক
+- wul
+- ▁Gal
+- ▁بۇ
+- ▁toda
+- siga
+- ▁zullen
+- lav
+- پی
+- ▁Х
+- fashe
+- ▁persisted
+- ▁attendant
+- ▁rey
+- 든
+- ▁Kom
+- ফ
+- ▁goose
+- ▁sources
+- ▁swan
+- ▁foster
+- ▁strongest
+- ▁tunnel
+- fic
+- iriza
+- hour
+- ▁އެ
+- awen
+- 十
+- ▁народ
+- ▁mejor
+- urwa
+- ▁eerste
+- ▁نە
+- core
+- ▁jewish
+- ▁coffin
+- ▁melted
+- anno
+- ▁sour
+- ▁quaint
+- φ
+- ▁exciting
+- ▁neben
+- ▁River
+- ▁autor
+- nnes
+- ▁dirt
+- ▁impressions
+- ▁알아
+- hle
+- munt
+- ▁daniel
+- ▁resort
+- ▁좋은
+- ▁occurrence
+- ▁Jes
+- ▁acquire
+- зу
+- ▁Pas
+- ▁everlasting
+- ▁rum
+- bis
+- ▁acute
+- gekommen
+- ▁vow
+- moi
+- ▁sheriff
+- amendement
+- kera
+- ▁kost
+- ▁adams
+- ▁improve
+- ▁proprietor
+- ▁کە
+- ▁tricks
+- ▁appar
+- ▁voort
+- ▁witnessed
+- ▁plague
+- ▁clothed
+- lden
+- гал
+- ▁rejected
+- odi
+- ایی
+- setzen
+- ▁football
+- ▁dock
+- ▁Bau
+- 得
+- ▁awfully
+- ▁chiefs
+- ▁poly
+- únic
+- 学
+- ▁gigantic
+- ▁amazing
+- tado
+- ▁awaiting
+- ▁offers
+- ibu
+- ▁반
+- ▁columns
+- ▁cars
+- ures
+- ▁Pla
+- ▁notable
+- اه
+- ▁일단
+- ▁Peter
+- ▁bella
+- lma
+- ▁magazine
+- ▁jardin
+- ▁morgan
+- ▁doctors
+- lick
+- ▁blossom
+- eja
+- ▁parlor
+- коў
+- ▁pushing
+- eww
+- ruf
+- ▁för
+- ▁righteousness
+- ન
+- ▁gens
+- ▁swimming
+- ▁banda
+- ▁banquet
+- ▁ब
+- gegeben
+- ▁wordt
+- erin
+- ▁reden
+- ▁4
+- ▁hasty
+- ▁discharge
+- مەن
+- ▁Dan
+- ▁fig
+- nit
+- ▁echter
+- lice
+- ▁dismay
+- ▁shocked
+- 심
+- ▁pobre
+- 定
+- اس
+- ▁ruler
+- ▁komunumo
+- mise
+- ▁했는데
+- geza
+- sinn
+- ▁strife
+- эт
+- lichkeit
+- '19'
+- rous
+- வு
+- ▁patent
+- anti
+- coloured
+- ▁동
+- ▁mod
+- ို
+- ▁bacon
+- yita
+- ▁clearing
+- lv
+- wand
+- phon
+- imo
+- hound
+- ▁ruled
+- fie
+- ▁Arm
+- ▁gloves
+- yin
+- angu
+- tlich
+- ▁radiant
+- nyarwanda
+- etti
+- ▁wu
+- ▁arrows
+- ▁explosion
+- ▁volta
+- heure
+- elde
+- ▁tub
+- 용
+- brun
+- ▁congregation
+- ▁집
+- ▁japan
+- ▁annual
+- ▁무슨
+- ▁destin
+- cident
+- ▁ша
+- ซ
+- ▁gan
+- ĝa
+- ▁Dies
+- ▁concept
+- tak
+- ▁yanjye
+- muntu
+- ▁plates
+- রা
+- ▁все
+- ▁raison
+- ũ
+- ት
+- ▁scheen
+- ▁probability
+- iens
+- ĉ
+- ▁Win
+- esco
+- sehen
+- 利
+- ▁pied
+- amo
+- utse
+- ▁effet
+- szy
+- ▁condemn
+- igne
+- ▁lodging
+- ▁brood
+- так
+- ▁egyptian
+- ▁escort
+- ▁provi
+- ▁Marc
+- ▁subsequently
+- ▁scripture
+- ул
+- ▁bonds
+- ▁tumult
+- ▁intentions
+- ▁shoe
+- ▁expensive
+- 감
+- lp
+- ▁viu
+- ▁vis
+- ▁rises
+- ▁requested
+- ▁ине
+- ▁yawe
+- tius
+- ▁quelle
+- ▁carved
+- палі
+- ular
+- পা
+- pira
+- ▁pump
+- লা
+- ▁mourning
+- ▁mule
+- utu
+- ough
+- ▁mechanical
+- ுக்கு
+- ▁kir
+- etto
+- ▁spoon
+- ▁indem
+- ▁sticks
+- ▁smiles
+- ▁assisted
+- vide
+- жи
+- ▁tools
+- 就
+- ▁mm
+- ис
+- dine
+- ▁winning
+- giving
+- ▁jug
+- ▁ŝ
+- ▁Bon
+- illon
+- ▁waiter
+- ▁economy
+- ▁venus
+- ▁abide
+- ambi
+- ▁squirrel
+- ▁sexual
+- ▁gehen
+- ▁bows
+- platz
+- ▁demanda
+- ▁numéro
+- ▁fuss
+- bile
+- seite
+- ▁rightly
+- lish
+- ▁нічога
+- θ
+- bü
+- ▁dismal
+- tische
+- ▁এ
+- رت
+- ▁hervor
+- onia
+- ▁vari
+- ▁oldest
+- ▁strive
+- ▁groß
+- وم
+- ▁pé
+- trees
+- ▁conjecture
+- ▁William
+- ▁kubi
+- ▁É
+- ▁প্র
+- ▁nad
+- east
+- ▁Tur
+- рат
+- ▁plough
+- ▁purity
+- ▁washing
+- ▁Rue
+- ▁seeds
+- lief
+- stal
+- ▁compassion
+- nego
+- ▁fantastic
+- ▁sama
+- ▁kugeza
+- ▁pastor
+- making
+- ▁decree
+- ▁clad
+- arn
+- ▁sac
+- ▁bass
+- kintu
+- ▁lamps
+- '`'
+- či
+- цэ
+- ▁mart
+- ▁County
+- ▁faults
+- cular
+- ▁بول
+- izza
+- ▁totally
+- ▁owl
+- kende
+- বে
+- pas
+- volu
+- тор
+- ssy
+- ▁Poste
+- zima
+- شە
+- ಕ
+- 천
+- ▁سا
+- ▁좋
+- ▁subdued
+- 水
+- ▁glacier
+- ▁explore
+- taf
+- ழ
+- anima
+- ▁conquer
+- ▁forgetting
+- ▁verte
+- າ
+- lop
+- ▁highness
+- ▁mede
+- нік
+- ▁functions
+- ▁stain
+- ▁Dimecres
+- guha
+- fia
+- дер
+- ▁cinc
+- ▁gedaan
+- ▁picking
+- ▁guided
+- assa
+- vig
+- ▁einge
+- ologie
+- ▁dean
+- ▁likeness
+- ▁ужо
+- 係
+- 였
+- ▁heathen
+- hom
+- ▁forbid
+- wara
+- ibindi
+- ▁happier
+- ▁Ak
+- gger
+- گا
+- ales
+- gaz
+- él
+- ▁Val
+- く
+- ▁latest
+- cir
+- ளி
+- лыҡ
+- kuu
+- ▁cob
+- اع
+- насці
+- 인데
+- фар
+- typ
+- aho
+- ruk
+- ▁deur
+- ▁heights
+- დ
+- embe
+- ▁iemand
+- cida
+- bound
+- ▁hommes
+- mpf
+- ▁allies
+- ▁Martí
+- ▁High
+- korera
+- ▁contrived
+- ▁abana
+- тыя
+- pil
+- hundert
+- ▁matthew
+- ▁monstrous
+- 县
+- ▁abend
+- ▁hielt
+- gation
+- ▁revenue
+- test
+- ▁sunny
+- ır
+- قی
+- estudi
+- ▁Nie
+- nej
+- ▁holl
+- ▁japanese
+- cido
+- ▁municipal
+- unge
+- ▁archbishop
+- ind
+- ▁examples
+- ▁يا
+- ▁carpenter
+- ▁okwe
+- ▁бул
+- ыць
+- ос
+- ▁bereits
+- ноў
+- ▁cuatro
+- ▁nineteenth
+- рак
+- ▁whe
+- ▁exhibition
+- ▁evolution
+- ishwa
+- gevoel
+- ▁gezicht
+- ▁dolor
+- ▁Camp
+- ▁douglas
+- inter
+- сця
+- ▁algo
+- ▁scared
+- see
+- ▁tidings
+- nov
+- polo
+- ▁enfin
+- aigua
+- ▁meinem
+- ▁fulfilled
+- ▁persian
+- цяг
+- ▁alternative
+- ▁thorn
+- وە
+- cord
+- ▁நி
+- ▁College
+- ைய
+- ▁könnte
+- ▁چی
+- ▁stump
+- ggle
+- বা
+- ▁près
+- ▁pleaded
+- ກ
+- ▁тым
+- diri
+- fé
+- ▁extract
+- ▁faculties
+- esprit
+- ▁voll
+- ▁unnatural
+- ▁시간
+- ▁gilbert
+- ▁다른
+- ▁haber
+- ▁particulars
+- 마
+- 명
+- idos
+- ▁naj
+- vingt
+- ▁schön
+- נ
+- ▁circuit
+- ▁organized
+- tiques
+- ▁producing
+- ancy
+- ▁vir
+- ▁representation
+- 국
+- ▁unique
+- ▁bidding
+- tam
+- ▁आ
+- vind
+- mag
+- Q
+- ▁await
+- ▁ки
+- ▁ele
+- etzt
+- garagaza
+- ▁notre
+- fangen
+- త
+- ▁bother
+- cil
+- ▁exhibited
+- kke
+- ▁rust
+- nama
+- аба
+- ▁quadr
+- igas
+- ▁месца
+- rust
+- ▁expenses
+- تم
+- ▁lump
+- ▁luego
+- ▁prevail
+- یە
+- ▁annoyed
+- ▁spake
+- மை
+- ▁savages
+- ▁Fre
+- ته
+- ▁Vall
+- ▁selection
+- ▁withdraw
+- ”.
+- ▁despise
+- inu
+- rg
+- ▁meadows
+- aten
+- puesto
+- ▁stad
+- enga
+- ▁Ger
+- ек
+- ▁ҡа
+- ▁dine
+- ▁dreaded
+- âme
+- ▁ikibazo
+- cic
+- ▁float
+- ▁crowned
+- ▁Pra
+- ▁odor
+- stelle
+- ▁marching
+- жым
+- ▁않
+- ▁lime
+- ▁casting
+- ramo
+- ▁polished
+- ▁akaba
+- ▁Bas
+- ▁dealt
+- ▁emerged
+- tol
+- ég
+- δ
+- edd
+- ▁Durant
+- ▁ober
+- rey
+- ĝis
+- ▁borrow
+- ▁sylvia
+- ▁blush
+- uwe
+- ▁tw
+- fect
+- ▁presume
+- ▁hafi
+- ▁humility
+- ▁también
+- دان
+- ▁sho
+- ▁parcel
+- ▁clergy
+- ▁agitated
+- ▁fashionable
+- ▁magistrate
+- ▁hinaus
+- شو
+- ण
+- ▁personage
+- ▁emphasis
+- ▁senior
+- teko
+- وان
+- ▁ox
+- дума
+- wegen
+- ▁원
+- ▁kirk
+- লে
+- gres
+- embo
+- ▁sweeping
+- чыць
+- ▁chef
+- ▁sung
+- ▁brig
+- ▁seines
+- ▁был
+- ரா
+- ▁yavuze
+- ▁Beth
+- ▁grade
+- ▁stamped
+- ▁infantry
+- ▁shriek
+- ▁twe
+- คน
+- له
+- lut
+- ▁constitute
+- ▁sting
+- ▁worldly
+- ▁pearl
+- ridge
+- ▁hospitality
+- ▁employer
+- ▁cursed
+- ▁vicar
+- gize
+- zier
+- onder
+- ▁balls
+- lagen
+- voca
+- ún
+- ▁hijo
+- ▁plank
+- liga
+- ▁déjà
+- ▁heures
+- ▁shouts
+- рын
+- ▁youngest
+- ▁gew
+- ▁verder
+- ▁bel
+- rya
+- ▁definition
+- dag
+- ەکە
+- ▁successfully
+- baye
+- ▁mondo
+- ▁massive
+- аваць
+- ensi
+- ▁mwi
+- ▁Ĉi
+- ында
+- ▁신
+- тү
+- ▁ascertain
+- ▁вось
+- ите
+- ▁liet
+- ▁Van
+- ▁relationship
+- cré
+- kker
+- ▁writings
+- acre
+- dian
+- hält
+- ▁pang
+- ▁ста
+- ▁Quel
+- dı
+- nait
+- umugabo
+- ▁peg
+- ▁approved
+- vos
+- 加
+- ▁bruce
+- ▁disturbance
+- ▁sentimental
+- hren
+- ▁차
+- det
+- ▁grupo
+- ▁invalid
+- ured
+- ▁vaguely
+- ▁wonderfully
+- tico
+- aquesta
+- shin
+- ▁tyrant
+- шке
+- ▁roared
+- ▁joyous
+- ▁spreken
+- ▁chart
+- ▁fee
+- ▁hood
+- ▁blake
+- কে
+- ▁ese
+- ▁engage
+- ▁player
+- hita
+- ▁peak
+- ▁conven
+- ▁imitation
+- nton
+- dó
+- ▁matin
+- under
+- ▁aunque
+- ▁tightly
+- ▁quant
+- ▁ہو
+- ▁eloquence
+- το
+- ▁sits
+- ote
+- ound
+- ▁colon
+- ▁في
+- ▁types
+- ▁asserted
+- ▁fifteenth
+- ▁fr
+- tio
+- ▁startling
+- ӗ
+- یں
+- ڕ
+- ▁dona
+- ▁chicken
+- ▁ĉefe
+- ▁هیچ
+- ▁arriving
+- ▁boiled
+- ▁devas
+- ▁complaint
+- ू
+- ▁marion
+- reɣ
+- ▁competition
+- force
+- ▁antes
+- କ
+- ▁gris
+- ▁barrel
+- oph
+- ▁passer
+- ▁spider
+- лас
+- soma
+- ▁luke
+- скіх
+- white
+- ▁welfare
+- ▁misschien
+- ijwe
+- rap
+- ▁кол
+- ރ
+- ▁Grand
+- ▁uses
+- ordre
+- ▁role
+- ▁دارد
+- ways
+- ▁saloon
+- چە
+- jin
+- ▁brigade
+- führen
+- ▁decline
+- ▁кон
+- ▁별로
+- ▁meetings
+- ▁Ara
+- how
+- ▁terme
+- ▁یه
+- road
+- ▁golf
+- ▁pasture
+- owa
+- ▁greece
+- ▁autre
+- kauf
+- ▁tas
+- ▁sonst
+- shinzwe
+- mwa
+- uter
+- ▁formula
+- ▁Arbeit
+- ▁circular
+- ▁encourage
+- ▁bamu
+- ум
+- ▁albert
+- rp
+- liv
+- lama
+- ▁plat
+- ▁scholar
+- ▁examining
+- роб
+- ▁awaited
+- ▁imperfect
+- ਿ
+- ▁сябе
+- jana
+- three
+- ▁wars
+- oke
+- ▁cert
+- ▁eerst
+- ▁utterance
+- ▁gus
+- ▁Tot
+- ▁plantation
+- ▁coloured
+- ▁oficial
+- ▁feminine
+- ▁whispering
+- ▁Für
+- іцца
+- haga
+- 행
+- egg
+- ▁داد
+- ▁Pen
+- eby
+- ▁bist
+- үн
+- ▁Mira
+- ▁valleys
+- ▁solomon
+- ▁무
+- ▁Berlin
+- ▁beautifully
+- ▁ethel
+- rama
+- ▁voy
+- hurst
+- ected
+- ▁teil
+- ▁episode
+- раз
+- ▁trova
+- licher
+- 日
+- gare
+- ▁foes
+- tare
+- ▁му
+- аг
+- ▁incidents
+- შ
+- ▁유
+- ▁tug
+- ▁raft
+- ▁Robert
+- ие
+- ▁raz
+- ▁estu
+- ▁patro
+- oba
+- ▁helmet
+- ▁restrain
+- 포
+- ▁wolves
+- ▁strand
+- ▁tropical
+- ▁bild
+- heden
+- ▁charms
+- иш
+- ▁стала
+- ▁exhibit
+- ▁medal
+- ▁lads
+- ▁mineral
+- ents
+- nyama
+- raba
+- office
+- ▁panel
+- या
+- ▁queda
+- ▁mirth
+- ران
+- tere
+- ▁luncheon
+- iment
+- ▁Spe
+- mpe
+- ▁إ
+- ▁attach
+- ▁sixteenth
+- zir
+- ▁bantu
+- ава
+- leta
+- ▁justly
+- ڵ
+- ▁excursion
+- ▁pear
+- ▁undertaking
+- ▁영화
+- uke
+- just
+- ɣen
+- ▁eloquent
+- 民
+- ís
+- дом
+- ▁кры
+- fit
+- kaj
+- ▁relate
+- ▁twelfth
+- ▁einfach
+- சு
+- ▁accordance
+- ▁verstand
+- chang
+- ▁chambre
+- ▁loin
+- ▁flowed
+- ▁verg
+- ▁genug
+- ▁butcher
+- ਲ
+- ▁zooals
+- wamu
+- ▁transfer
+- koko
+- komen
+- ▁pauvre
+- ▁arribar
+- ▁год
+- ▁방
+- ▁Bri
+- laf
+- venir
+- ▁stadt
+- ▁spots
+- ▁anthony
+- ▁와
+- ί
+- وو
+- ▁তিনি
+- шт
+- ден
+- gs
+- ▁earned
+- ▁esper
+- ێک
+- ▁mines
+- ▁yell
+- ▁roaring
+- gn
+- eron
+- ▁exercised
+- ▁environment
+- ▁snatched
+- teri
+- ▁justify
+- ▁carries
+- vill
+- ▁наш
+- ▁endeavoured
+- ▁ark
+- ▁raid
+- ▁kurz
+- ddu
+- 밖에
+- ▁juist
+- ▁rascal
+- ▁South
+- সা
+- ▁fy
+- ▁Cla
+- ▁oliver
+- onda
+- ▁factory
+- ▁spreading
+- ▁nigh
+- ▁mustn
+- 질
+- 德
+- ▁geh
+- ▁재밌
+- pfa
+- ▁agua
+- ▁umwana
+- ▁알
+- ▁Hey
+- ▁wink
+- ilo
+- ▁fueron
+- ▁echo
+- ▁partir
+- ▁cependant
+- sze
+- ▁rector
+- ▁abandi
+- ▁친구
+- ▁bafite
+- ▁Spiel
+- awe
+- لم
+- zze
+- ▁begon
+- ▁geven
+- ▁resentment
+- êr
+- ▁வ
+- ▁demon
+- rre
+- ▁horace
+- ▁geloof
+- ▁Но
+- ▁travail
+- ▁volumes
+- ▁approval
+- ▁constance
+- ▁depended
+- кән
+- bika
+- ▁donner
+- ▁중
+- ▁bois
+- ▁feu
+- ▁soixante
+- 미
+- ▁wari
+- ▁tau
+- meza
+- তি
+- ▁scrap
+- kali
+- lijke
+- 역
+- ▁birthday
+- ▁owed
+- ▁thicket
+- ▁trench
+- валася
+- ਨ
+- ref
+- essa
+- ▁Fel
+- ▁Unter
+- 所
+- wang
+- ▁М
+- ▁expressions
+- ▁дэ
+- yana
+- ▁kujya
+- шта
+- ▁soir
+- ▁amuse
+- ▁Quin
+- ▁talents
+- ▁singer
+- ▁kaum
+- ▁Para
+- ▁stud
+- ▁saa
+- 두
+- ▁cells
+- subi
+- ध
+- ▁woorden
+- ▁domain
+- вой
+- ▁donna
+- ▁coldly
+- ▁conform
+- ▁pluck
+- meester
+- ▁Ра
+- ▁baga
+- ▁prendre
+- rire
+- ▁olw
+- ▁demonstration
+- kto
+- stück
+- cock
+- ▁Ser
+- ▁token
+- cott
+- ▁charges
+- ▁manufacture
+- bred
+- yant
+- bali
+- ▁llama
+- ▁assertion
+- ▁دارم
+- kita
+- üb
+- ▁nchi
+- ▁dresses
+- nehmen
+- lij
+- ▁marshal
+- tique
+- verse
+- ัน
+- ▁feeding
+- ▁komt
+- ▁애
+- ▁foam
+- ▁grau
+- ▁trente
+- ain
+- ios
+- ял
+- раб
+- ▁guerre
+- ▁respected
+- ▁ayen
+- прац
+- ▁fools
+- temp
+- တ
+- ▁mädchen
+- ▁dad
+- ▁appearing
+- waka
+- ▁constru
+- ▁louder
+- ▁needn
+- ▁recommended
+- kol
+- ▁pathetic
+- ellen
+- ▁Hari
+- ▁mü
+- ologist
+- lement
+- ▁armour
+- ▁пакуль
+- ▁speeches
+- tron
+- ▁tents
+- cció
+- houses
+- ▁interpretation
+- 与
+- arri
+- كە
+- osi
+- ог
+- vou
+- ▁tribute
+- ▁sheets
+- tige
+- champ
+- മ
+- spect
+- ▁communicate
+- ▁chambers
+- ▁transform
+- ▁quella
+- τα
+- ust
+- mdash
+- ▁nei
+- ▁morir
+- ▁gregory
+- ▁ladyship
+- schiff
+- ▁Seg
+- ▁på
+- ▁Dat
+- ▁North
+- сан
+- ▁dora
+- ▁patients
+- ▁мин
+- ckte
+- сце
+- ▁decidedly
+- tali
+- ▁sache
+- ▁fist
+- utilitza
+- ▁arnold
+- wise
+- ▁baixa
+- ▁dome
+- ▁Op
+- koresha
+- aciones
+- rell
+- ▁punto
+- ▁nunca
+- ▁triumphant
+- ာ
+- кер
+- ▁Jahr
+- ▁Цяпер
+- ِي
+- ▁gemacht
+- fro
+- ▁genoeg
+- ▁tint
+- ▁virtuous
+- ▁creed
+- களை
+- ▁wits
+- gress
+- ▁чу
+- う
+- ▁tooth
+- ▁preaching
+- ▁continuous
+- ▁jaw
+- ▁Vila
+- ▁urgent
+- gebracht
+- ▁cows
+- ▁nome
+- zó
+- ▁wese
+- ▁dura
+- ▁guru
+- ▁10
+- ▁confer
+- ▁கு
+- sum
+- ▁kontraŭ
+- ▁differences
+- ▁chatter
+- ▁Во
+- ▁praying
+- ▁captive
+- garuka
+- enta
+- ▁chemical
+- noj
+- ▁دا
+- ▁candles
+- eixen
+- رم
+- ▁دی
+- ▁pren
+- ▁paths
+- ži
+- uve
+- ▁housekeeper
+- ▁brandy
+- cono
+- ▁jerk
+- ▁abstract
+- ▁General
+- ▁beer
+- baha
+- ▁pouring
+- aco
+- cá
+- find
+- ifying
+- ▁verdad
+- ree
+- ▁horns
+- ziehen
+- ▁loop
+- hou
+- renge
+- ▁habitual
+- ▁veces
+- ▁richmond
+- ▁riot
+- ▁appealed
+- ▁Yo
+- ▁effected
+- rle
+- bora
+- ▁menace
+- ▁stretching
+- ▁Cy
+- satz
+- ▁cabeza
+- lled
+- னி
+- િ
+- யில்
+- ▁fills
+- bby
+- ▁Af
+- ▁endured
+- ▁divorce
+- ▁impressive
+- ▁пере
+- ▁devote
+- ▁orator
+- bere
+- ▁switch
+- mek
+- ▁commons
+- power
+- ▁jag
+- ▁alongside
+- သ
+- fy
+- appel
+- ▁alguna
+- ▁patiently
+- ▁beaucoup
+- ▁bleef
+- ▁esther
+- ▁pow
+- quen
+- ▁пасля
+- ▁impress
+- ▁treball
+- ▁ambitious
+- idas
+- һә
+- ▁mason
+- ▁아직
+- ▁militar
+- dorf
+- tragen
+- ▁thereupon
+- 유
+- ▁pulse
+- ▁logical
+- ▁abge
+- ▁rachel
+- тры
+- ҡо
+- undu
+- ▁blanket
+- ▁eran
+- 里
+- lly
+- ect
+- ▁orchard
+- gero
+- 理
+- ট
+- ▁zat
+- ▁augment
+- ▁compound
+- ▁تر
+- ▁sustain
+- gebiet
+- ▁indignant
+- ▁stooped
+- ▁tact
+- ئی
+- ▁스
+- ▁submitted
+- ▁discern
+- ▁documents
+- rish
+- ase
+- ▁unconsciously
+- ▁travellers
+- ▁Castell
+- ▁Este
+- ranga
+- ▁yaba
+- 高
+- ambo
+- has
+- ▁해서
+- тат
+- ้
+- ▁frog
+- posició
+- née
+- ope
+- fasha
+- ▁loro
+- zorg
+- dora
+- ▁اون
+- ną
+- ▁yali
+- ▁initial
+- dog
+- ▁prose
+- ▁ĝis
+- eḥ
+- ▁wahr
+- ▁favourable
+- ▁serene
+- cara
+- ีย
+- іў
+- rated
+- ▁Dimarts
+- сла
+- парт
+- stern
+- zte
+- lated
+- ศ
+- zioni
+- ▁bart
+- ▁streak
+- ▁hebrew
+- ▁будуць
+- ▁wanneer
+- ▁dignified
+- ▁whereby
+- nelle
+- ▁shift
+- ▁ingenious
+- ەت
+- ▁Ку
+- kaza
+- ▁huit
+- ▁graves
+- ▁shrugged
+- wee
+- ▁glare
+- ▁animated
+- lub
+- ▁côté
+- langen
+- kir
+- ufu
+- gem
+- ▁stillness
+- ▁hunted
+- iku
+- ▁authors
+- ebb
+- ▁música
+- ▁ezi
+- tures
+- ▁scout
+- ▁continues
+- ▁adapted
+- ▁gente
+- ▁excite
+- ▁jaroj
+- ▁poble
+- ▁kelkaj
+- ங்கள்
+- gull
+- ▁ап
+- dium
+- ▁미
+- ▁morality
+- tiu
+- роў
+- ▁deliberate
+- ▁어떤
+- ▁nouvelle
+- ▁musket
+- olu
+- hale
+- ▁crowds
+- ▁finden
+- 터
+- kie
+- ▁nobility
+- 하니까
+- alla
+- ▁pearls
+- ▁cliffs
+- iver
+- وں
+- ▁flee
+- ▁Ez
+- ▁willow
+- ûn
+- ▁deepest
+- ▁sustained
+- ▁stehen
+- ṭṭ
+- schritt
+- ▁trains
+- nay
+- പ
+- ▁tute
+- یی
+- ▁prosper
+- ▁Az
+- ▁ras
+- ▁voz
+- яв
+- horse
+- ▁Ха
+- 직
+- ▁پێ
+- rono
+- ▁amen
+- ▁sci
+- bwi
+- larga
+- kubi
+- ▁polític
+- ▁présent
+- ▁miller
+- ▁Ankaŭ
+- ▁sabbath
+- ▁너가
+- ▁readiness
+- дыр
+- ▁frighten
+- ilia
+- 노
+- ▁hereafter
+- ▁பு
+- ▁statt
+- ▁algun
+- uy
+- ▁oars
+- ▁familie
+- espera
+- ស
+- ▁complained
+- daw
+- registr
+- ▁harmless
+- ىسى
+- ქ
+- নে
+- ▁suicide
+- rons
+- ▁komm
+- kis
+- кры
+- ▁despised
+- kus
+- ▁luther
+- ▁인제
+- bee
+- ▁கா
+- ▁carbon
+- viol
+- ▁eva
+- lande
+- ssin
+- ▁bona
+- ▁sealed
+- chus
+- न्
+- ▁buttons
+- sul
+- paro
+- ▁proves
+- ▁continua
+- into
+- ▁strained
+- ateur
+- ނު
+- ான்
+- ▁welcomed
+- ▁parla
+- cchi
+- leɣ
+- ▁oppose
+- rme
+- ▁papier
+- furi
+- ▁ўсе
+- ▁ernst
+- ▁almighty
+- glio
+- blu
+- ▁aufge
+- ät
+- ▁dip
+- шо
+- ▁Ту
+- zien
+- ▁Cre
+- 克
+- ▁outline
+- ээ
+- ▁goat
+- ▁artists
+- ▁punch
+- ▁kat
+- 개
+- ▁bees
+- ▁primary
+- ▁mois
+- ▁হ
+- ▁ہے
+- ▁오늘
+- ▁lhe
+- нак
+- زا
+- kti
+- મ
+- ▁possibilities
+- ▁cai
+- rol
+- ▁dernier
+- ▁monks
+- нд
+- ▁Nya
+- ސް
+- ▁underneath
+- ▁sympathetic
+- late
+- ▁superiority
+- ▁rejoiced
+- ੇ
+- cade
+- etta
+- iff
+- ▁convict
+- then
+- ▁damned
+- ▁sombre
+- uu
+- ▁premi
+- ruka
+- ebla
+- bier
+- ▁lazy
+- ▁Bor
+- ▁apostle
+- ▁Ant
+- بر
+- kreis
+- ன்ற
+- چا
+- ▁angles
+- ▁promising
+- ▁embarrassment
+- dell
+- ▁sworn
+- boden
+- ▁radical
+- ▁baa
+- ▁gener
+- ▁encouragement
+- nuncia
+- ▁blick
+- ▁translation
+- kop
+- ▁zeg
+- shop
+- ▁violin
+- اپ
+- gone
+- ฉัน
+- 호
+- ▁activities
+- ▁thirteenth
+- lust
+- ▁argue
+- culo
+- ▁stimme
+- ก็
+- ▁clue
+- ▁Juli
+- tez
+- 보다
+- ▁dispose
+- ▁herz
+- ▁그건
+- ▁bog
+- ▁consum
+- ▁Ils
+- ▁rencontre
+- try
+- ▁leute
+- بو
+- 教
+- ▁herald
+- فی
+- sión
+- ▁comfortably
+- ▁señora
+- tering
+- ▁groupe
+- cient
+- ▁솔직히
+- ев
+- ▁apples
+- ▁babi
+- flo
+- ▁के
+- nemen
+- ▁downward
+- ▁attending
+- łu
+- ▁wußte
+- ▁slice
+- ▁flashing
+- উ
+- ▁az
+- coming
+- ▁selling
+- ▁conte
+- ladi
+- 拉
+- ಲ
+- ▁비
+- ▁domina
+- ▁mona
+- ▁clinging
+- gora
+- ▁الم
+- ▁hunters
+- ▁distrust
+- ▁kugirango
+- ▁battery
+- ▁fearing
+- ▁holiness
+- estre
+- grund
+- ▁dira
+- ▁wambaye
+- ▁observa
+- ▁buli
+- క
+- ▁ascended
+- gesetzt
+- ggs
+- ▁rural
+- ▁mga
+- ▁invasion
+- ▁дә
+- jem
+- ▁compact
+- ▁armor
+- ▁creeping
+- ▁fué
+- эй
+- 하면
+- ▁प
+- ▁pleasantly
+- ▁gabriel
+- ▁vita
+- ▁Cer
+- guru
+- ĝi
+- 及
+- ▁preached
+- ▁administra
+- eto
+- лай
+- 왔
+- lov
+- ▁dart
+- ҳ
+- ▁yr
+- yim
+- ии
+- ▁honestly
+- ▁realised
+- ▁vrij
+- ▁엄마
+- ▁workmen
+- வர்
+- ▁dive
+- camp
+- ▁sprung
+- comb
+- 업
+- ▁detect
+- ▁cluster
+- овы
+- ▁practised
+- zahl
+- nahme
+- bos
+- pres
+- ▁missionary
+- ▁gordon
+- ▁crust
+- ▁reverse
+- ▁breathless
+- ▁logic
+- ਤ
+- tani
+- ▁designs
+- өр
+- ddling
+- ▁maud
+- ▁saber
+- ▁kye
+- ▁Martin
+- ▁reap
+- ▁blazing
+- ania
+- ▁max
+- ក
+- ▁mujer
+- ▁yella
+- рен
+- houd
+- ▁manifestation
+- ▁shaft
+- て
+- ▁abbey
+- ▁요즘
+- 被
+- iles
+- ▁Rus
+- ▁mehrere
+- ▁canadian
+- ▁sets
+- ▁lingered
+- glia
+- ▁weighed
+- ▁iba
+- ▁tamen
+- ▁liable
+- ▁Hier
+- lein
+- ▁piled
+- sobanu
+- ▁roughly
+- ▁си
+- ▁farmers
+- zim
+- ▁complicated
+- йын
+- ▁ensuite
+- dder
+- ▁ela
+- ▁possessions
+- ▁мар
+- ▁languages
+- 前
+- kata
+- limu
+- ▁attacks
+- ެއް
+- ▁attachment
+- ▁praised
+- bij
+- ▁Тут
+- ▁acres
+- ubwi
+- ▁wiped
+- лык
+- ▁Там
+- ▁Its
+- tino
+- flam
+- arde
+- renga
+- ▁schoon
+- ▁petty
+- yaka
+- sieg
+- ▁robes
+- ▁Ban
+- rage
+- plant
+- ngen
+- үл
+- ▁embargo
+- ▁Umugabo
+- ▁kati
+- ред
+- hart
+- ▁hacia
+- dol
+- ▁Bro
+- ▁imyaka
+- ▁merciful
+- ▁twentieth
+- ыў
+- ▁castell
+- rijk
+- autres
+- ccion
+- ▁거기서
+- ▁admission
+- ның
+- ▁freed
+- trau
+- ▁philadelphia
+- ▁eigentlich
+- ▁سے
+- таш
+- ▁wickedness
+- ец
+- rog
+- ં
+- ▁puff
+- 食
+- tul
+- эл
+- ▁desperately
+- نە
+- 된
+- ▁hose
+- ▁ghastly
+- ▁avaient
+- нат
+- enge
+- ▁anni
+- ▁marine
+- ▁Cas
+- ҭа
+- angwa
+- meye
+- нә
+- ▁proclaimed
+- evi
+- ▁있지
+- nts
+- ▁gag
+- ന
+- ения
+- ▁ditch
+- ▁frail
+- ▁ebb
+- رد
+- rot
+- zes
+- ▁morris
+- ու
+- ▁upwards
+- iser
+- ▁veu
+- ▁slechts
+- ▁einzige
+- ▁Film
+- ▁reporter
+- ▁йы
+- ▁pap
+- 티
+- ously
+- vivi
+- nych
+- ▁wag
+- ▁და
+- baut
+- ▁5
+- enne
+- ▁tailor
+- ▁yacht
+- ▁tutti
+- ▁exalted
+- ▁frenchman
+- ▁drill
+- ▁soit
+- ▁bunny
+- üh
+- ав
+- sitz
+- ▁دست
+- таў
+- bing
+- ▁accuse
+- ▁lame
+- lucht
+- ▁garment
+- ències
+- ▁uzi
+- ▁offensive
+- osos
+- ▁mach
+- ▁vicious
+- baro
+- ▁vot
+- ▁piety
+- emos
+- nwen
+- enca
+- ▁flush
+- tava
+- ▁Ḥ
+- ▁allowing
+- ▁resigned
+- corn
+- کار
+- ▁destination
+- ▁stealing
+- geführt
+- nier
+- ▁dicho
+- ▁alfred
+- ▁conqueror
+- ▁eĉ
+- ▁exploit
+- ▁peer
+- ▁가는
+- ▁assent
+- git
+- ▁goede
+- ▁aye
+- mouth
+- year
+- brid
+- свет
+- ▁pinch
+- ▁Ay
+- beck
+- ▁obtaining
+- bereit
+- һ
+- ▁consulted
+- ▁retour
+- ję
+- 던데
+- ▁lakes
+- ▁tarde
+- ▁malice
+- ▁Té
+- ▁graham
+- clav
+- ▁annie
+- dá
+- ▁Вось
+- ▁هە
+- list
+- ▁hield
+- matic
+- ▁allan
+- ▁serait
+- folg
+- ▁genial
+- ▁ankoraŭ
+- ட்டு
+- ▁ultimate
+- spekt
+- ▁muto
+- ▁gwe
+- ▁nabi
+- ▁comprend
+- ▁как
+- ▁gest
+- haye
+- inta
+- ▁visions
+- ▁groaned
+- ▁excellence
+- ▁growled
+- lha
+- ▁Auto
+- މ
+- ▁알바
+- bber
+- ught
+- ▁elderly
+- berries
+- ▁continu
+- ▁wol
+- ▁tran
+- ▁instructed
+- py
+- ▁positions
+- ulation
+- ▁bravely
+- forme
+- ▁versuch
+- ▁bitte
+- ▁zouden
+- ర
+- uɣal
+- ▁strengthen
+- ▁Oh
+- kov
+- စ
+- ▁맨날
+- histoire
+- ▁donkey
+- lers
+- ▁haut
+- ▁gezegd
+- ▁achieved
+- ▁kure
+- гор
+- abord
+- ▁rede
+- tritt
+- ▁plea
+- بەر
+- ▁suffice
+- ▁gem
+- ▁knot
+- gl
+- ▁vicinity
+- ▁seasons
+- гын
+- ▁condu
+- 겠다
+- ▁American
+- ▁gleaming
+- 度
+- ▁warmly
+- ხ
+- ▁asks
+- ▁inheritance
+- ▁dol
+- ▁shawl
+- ▁enfants
+- ▁aṭas
+- ▁mae
+- ▁guitar
+- ▁desirous
+- ▁본
+- ▁essa
+- ▁significa
+- lok
+- nger
+- த்தை
+- stieg
+- ▁fiel
+- ▁corresponding
+- ▁enthusiastic
+- ▁segui
+- sell
+- ▁ministry
+- byi
+- àn
+- ▁pouvait
+- тара
+- chal
+- ▁бит
+- ▁defiance
+- buch
+- ▁frère
+- ▁twain
+- ▁perd
+- ▁junge
+- ▁ash
+- ▁endeavor
+- ▁diferents
+- ▁خیلی
+- kumi
+- ▁tray
+- ▁viene
+- ação
+- ▁berg
+- ▁astonishing
+- eller
+- ▁steer
+- quis
+- ▁людзі
+- ▁sorrows
+- ▁peggy
+- ▁frown
+- ▁thro
+- ▁continual
+- ▁mbili
+- ▁constable
+- rup
+- ▁adapt
+- ▁drain
+- খ
+- cit
+- ▁westward
+- keneye
+- ▁transaction
+- ▁pitiful
+- ▁representatives
+- ▁loan
+- ▁gust
+- itud
+- kina
+- ▁weten
+- ▁commonplace
+- ▁recognise
+- ▁grund
+- guin
+- ▁segon
+- ▁seventeenth
+- ▁russ
+- mez
+- ĉe
+- ▁applause
+- ▁quantities
+- lid
+- ▁tough
+- தே
+- ▁Ни
+- ▁bishops
+- ▁Sha
+- ▁handful
+- ▁hazard
+- ▁espèce
+- ▁keeper
+- ▁праз
+- ▁prophets
+- ▁гар
+- ▁wort
+- ▁blankets
+- gled
+- pression
+- ▁рабіць
+- ▁curt
+- ▁waarom
+- ▁Afrika
+- ▁oo
+- ▁않아
+- لي
+- ▁charley
+- ▁berlin
+- lah
+- ▁tutto
+- nard
+- ▁affirm
+- ▁sacrament
+- ▁waking
+- ▁allowance
+- ▁cort
+- ṛa
+- ▁mum
+- ▁meisje
+- ▁Nel
+- ▁wonders
+- ▁souvent
+- ▁thieves
+- 내
+- 인가
+- ▁attempting
+- ▁sechs
+- ▁allusion
+- ▁embraced
+- ▁лә
+- ▁pledge
+- ного
+- wart
+- ▁ئەم
+- rren
+- ▁venu
+- tanu
+- шчы
+- obli
+- ▁admiring
+- ▁rejoicing
+- ▁lawrence
+- ▁compromise
+- ▁symptoms
+- ▁accusation
+- ▁theories
+- ▁Parti
+- ▁contemplation
+- ▁kuza
+- ▁apron
+- भ
+- غان
+- ▁lid
+- ject
+- ▁bate
+- ▁mantle
+- زان
+- schw
+- ▁behaved
+- rato
+- ▁horseback
+- ▁놀
+- ▁approve
+- ▁nouveau
+- ▁deine
+- bourg
+- produkt
+- ▁creator
+- ್
+- ▁meals
+- ட்
+- 케
+- ▁thursday
+- ▁lawful
+- рон
+- ▁vrai
+- bly
+- فَ
+- ເ
+- ▁gent
+- ▁civilized
+- ▁orphan
+- ▁carelessly
+- ▁duc
+- ▁traced
+- ▁sparkling
+- ▁hostess
+- ▁transparent
+- ▁Mag
+- ▁loko
+- ▁letzte
+- ▁casual
+- ▁darted
+- ▁trumpet
+- ▁african
+- ▁machines
+- bringen
+- cated
+- raw
+- deel
+- ري
+- ▁کنید
+- із
+- ▁manhood
+- ▁таму
+- schau
+- ्या
+- ވަ
+- ▁preĝejo
+- kind
+- ▁ceux
+- ode
+- ▁quoted
+- riz
+- ҟа
+- ▁austria
+- ▁repeating
+- grave
+- شن
+- 외
+- ▁stuart
+- duct
+- vita
+- ფ
+- ▁frock
+- ▁mira
+- лаш
+- cation
+- imbu
+- 海
+- لىك
+- ▁notions
+- ▁ornament
+- ▁perdu
+- ▁transferred
+- ▁Ɛ
+- ▁drifted
+- ▁influences
+- ▁borrowed
+- ▁attraction
+- Р
+- мәй
+- ▁spi
+- ▁behave
+- ▁assassin
+- ▁bursting
+- iy
+- ▁dusty
+- ▁싶어
+- ▁communicated
+- ▁uri
+- qq
+- ▁laden
+- ▁opge
+- (
+- ▁skeleton
+- تون
+- ▁або
+- ▁declaring
+- ▁psalm
+- ▁inscription
+- дж
+- dera
+- கள
+- ▁namen
+- ▁perished
+- eid
+- impa
+- wich
+- mpli
+- ▁mississippi
+- ▁hanged
+- ▁cog
+- ▁jene
+- скія
+- ▁location
+- ▁irre
+- ހަ
+- ▁meditation
+- ▁sullen
+- ▁보니까
+- tian
+- ▁executive
+- ▁spectacles
+- ▁biza
+- ્
+- ▁consistent
+- dau
+- ▁yett
+- და
+- ▁announcement
+- ▁hop
+- өл
+- ▁하면
+- ▁sadness
+- ."
+- ში
+- ▁prosperous
+- ▁Wan
+- сть
+- مه
+- ▁passe
+- ді
+- ▁confidential
+- ▁daran
+- dition
+- ▁respecting
+- ▁أن
+- bû
+- liest
+- 디
+- lika
+- ▁Ariko
+- ▁Oli
+- igna
+- ▁rouge
+- ▁glances
+- ▁hagati
+- ▁фа
+- સ
+- ándose
+- ▁customary
+- ▁adjoining
+- yamba
+- ▁babies
+- ▁squadron
+- année
+- ▁wasser
+- ▁இந்த
+- ▁Ул
+- ▁가고
+- ▁extending
+- thought
+- ▁distressed
+- ▁sentences
+- ▁setzte
+- ▁evils
+- scheiden
+- ▁margin
+- ▁returns
+- ▁circles
+- ▁Sy
+- 文
+- கை
+- god
+- ▁우리가
+- plan
+- аюць
+- ▁вас
+- ▁legion
+- ▁treason
+- ный
+- ▁lleva
+- ▁apartments
+- тен
+- mother
+- ீ
+- ▁거는
+- ▁قو
+- ▁German
+- iba
+- ▁갔다
+- ▁publik
+- ▁fam
+- ▁whatsoever
+- ▁warfare
+- ▁plunder
+- ▁plaça
+- hend
+- ▁errors
+- ஸ்
+- ▁política
+- ۋە
+- ▁wiser
+- moni
+- ▁sleepy
+- ▁governed
+- ▁lieve
+- ▁marrying
+- પ
+- тәр
+- wesen
+- ▁essentially
+- ▁milieu
+- ▁weigh
+- ▁chaque
+- ▁includes
+- ▁monta
+- ▁snatch
+- nci
+- ▁juice
+- tec
+- гад
+- ▁avenge
+- ncy
+- ▁produ
+- ▁tien
+- 这
+- ▁ubuzima
+- ▁maître
+- کردن
+- kend
+- ▁hunne
+- sí
+- 키
+- fungu
+- 았
+- ј
+- ▁gardener
+- сту
+- ▁didst
+- ▁camel
+- ▁salva
+- ▁refined
+- ▁pete
+- ▁vue
+- ores
+- кар
+- फ
+- ▁lodgings
+- ▁caso
+- zed
+- ▁meg
+- ▁nacional
+- won
+- ▁será
+- ▁balloon
+- ▁quasi
+- ▁delicacy
+- play
+- stairs
+- 통
+- ▁nem
+- ▁decisive
+- ▁Del
+- ailleurs
+- 明
+- ▁dainty
+- pata
+- ▁maakte
+- れ
+- ▁hurled
+- pó
+- ▁Stra
+- ▁pew
+- 安
+- ▁Mark
+- gren
+- ▁aange
+- ▁serves
+- ▁quoi
+- ▁hombres
+- ▁tongues
+- zera
+- ▁Га
+- 약
+- ▁questioning
+- ▁pip
+- cena
+- ბ
+- вым
+- ▁ticket
+- ▁chanced
+- ▁plunge
+- ▁profond
+- кая
+- ▁vehicle
+- ▁dominion
+- pad
+- acco
+- ▁ridicule
+- ▁contributed
+- ▁paces
+- ▁stalk
+- 길
+- ▁splendour
+- uḍ
+- gina
+- ▁beter
+- licht
+- ▁ה
+- rif
+- ▁homeward
+- Aba
+- ந
+- ▁avez
+- ▁소
+- ▁haw
+- 亚
+- ▁있을
+- ▁baggage
+- ▁generosity
+- ▁merits
+- ▁shapes
+- ▁laatste
+- ▁Està
+- amento
+- 하지
+- spel
+- aji
+- ▁apostles
+- ▁henceforth
+- handed
+- ▁defended
+- ▁grit
+- proc
+- cry
+- lou
+- ▁Ɣ
+- shyira
+- with
+- ные
+- ▁liking
+- acce
+- straße
+- issant
+- ▁converse
+- ▁ricevis
+- กับ
+- ɣur
+- ▁vine
+- ▁пя
+- ▁corruption
+- abandi
+- ▁bira
+- ▁persecution
+- ▁phenomenon
+- ▁gravel
+- கு
+- ▁quivering
+- ▁dicht
+- ▁Ag
+- цый
+- ▁affliction
+- ▁rimwe
+- ctive
+- ק
+- ▁gesagt
+- ▁anticipated
+- ▁duel
+- klu
+- ▁democratic
+- ▁battles
+- ▁Air
+- 能
+- ▁bluff
+- ▁gef
+- cita
+- зя
+- ▁pension
+- ▁shelf
+- ▁dam
+- pak
+- lino
+- ▁blunder
+- цу
+- ע
+- ▁Most
+- ▁argued
+- ▁imposed
+- ▁supernatural
+- ▁umbrella
+- ▁uwa
+- wacu
+- ▁ўз
+- twara
+- ▁discretion
+- ▁dimly
+- cents
+- extrem
+- lini
+- ▁cosas
+- jahr
+- ▁hinge
+- ভ
+- ▁cot
+- ▁eyebrows
+- ▁eleanor
+- ▁falsehood
+- ▁daher
+- ▁Mary
+- defini
+- amt
+- laub
+- ▁Ҡа
+- ▁abound
+- ▁grandeur
+- ▁secondly
+- mod
+- ▁bez
+- ▁ahubwo
+- аваны
+- ▁gene
+- fes
+- rca
+- 에는
+- lisa
+- eder
+- ▁cora
+- ▁trifling
+- ▁neighbouring
+- ▁precise
+- яўляецца
+- ▁obligation
+- ▁litter
+- ▁blik
+- phone
+- ▁줄
+- anyi
+- ▁acaba
+- ▁stray
+- ▁fuel
+- ▁bottles
+- vey
+- ▁sais
+- ▁같
+- ▁resurrection
+- ▁venerable
+- estima
+- ▁provincial
+- ▁venice
+- versa
+- ▁만나
+- ▁consumed
+- ▁translated
+- ▁teu
+- ނ
+- nywa
+- ▁cove
+- ór
+- ▁vegetables
+- ▁attributed
+- coat
+- ▁hob
+- ▁apparatus
+- '200'
+- ▁philosophers
+- ▁traveling
+- ▁beth
+- ▁hiyo
+- ▁lined
+- ائ
+- ▁quoth
+- ▁kumenya
+- ▁extravagant
+- ▁harness
+- zion
+- ▁breach
+- ▁counting
+- ▁zulk
+- ▁sternly
+- дэ
+- lec
+- turo
+- ▁fuhr
+- кава
+- ▁socrates
+- ▁phase
+- mä
+- yaga
+- ▁نیست
+- ▁naught
+- ▁attendants
+- ▁gelijk
+- ▁programa
+- ▁separa
+- கோ
+- ▁jaws
+- ▁descending
+- ▁pharaoh
+- ▁intensity
+- riza
+- тр
+- rée
+- aran
+- ▁prolonged
+- 음
+- ▁worm
+- denken
+- ▁technical
+- dden
+- дин
+- ▁owners
+- ▁encontr
+- ▁strikes
+- ▁disregard
+- ▁puerta
+- ondo
+- ▁гэтыя
+- ▁accomplishment
+- mount
+- ▁مع
+- ▁pels
+- ▁weird
+- ▁fragment
+- ▁jasper
+- ▁sincerity
+- legt
+- лам
+- றி
+- qe
+- сын
+- ▁rumor
+- cian
+- 구나
+- wir
+- ▁fum
+- ▁gahunda
+- ▁proudly
+- cyaha
+- love
+- vision
+- ▁pretence
+- ▁whim
+- ประ
+- ▁maggie
+- ił
+- cier
+- ▁буду
+- arian
+- ný
+- ▁vy
+- ▁سی
+- ▁estates
+- ▁kettle
+- ▁stations
+- ▁cough
+- ▁людзей
+- ▁Nun
+- ▁mwana
+- ▁사람들
+- ▁باید
+- ranye
+- ▁زۆر
+- 而
+- ▁mocht
+- inen
+- ▁rwego
+- த்தில்
+- ▁Monsieur
+- ▁latitude
+- ▁Ji
+- ▁miracles
+- ▁gorgeous
+- မ
+- ▁closet
+- ▁solely
+- ▁toast
+- še
+- яс
+- гла
+- ebi
+- lch
+- print
+- ře
+- ▁кү
+- zelve
+- ▁partit
+- ▁basa
+- ▁divisions
+- ▁peasants
+- ▁camino
+- avais
+- ▁mouths
+- ío
+- ▁naye
+- ▁mystic
+- gambo
+- ▁ای
+- shaped
+- ▁Ent
+- wandel
+- ็
+- ▁leagues
+- ▁família
+- ню
+- ▁distributed
+- rite
+- cò
+- ску
+- ▁instinctively
+- ▁orleans
+- mid
+- arms
+- ▁alley
+- Л
+- ▁woord
+- моў
+- ).
+- ▁energetic
+- ▁terrific
+- ▁사실
+- ▁그걸
+- ▁Qu
+- ▁lumber
+- ▁suppressed
+- ▁honorable
+- ▁focus
+- ▁eut
+- 후
+- үр
+- ▁kö
+- ▁reduce
+- пар
+- ▁schau
+- ▁jahre
+- 목
+- нен
+- 由
+- ▁lettre
+- ▁말이야
+- konstru
+- ▁georgia
+- ▁brightly
+- раш
+- ▁Мин
+- ▁ça
+- La
+- ▁George
+- ngera
+- gier
+- ถ
+- ▁slid
+- ając
+- ▁consisting
+- ес
+- سم
+- ێت
+- ▁달
+- laden
+- terra
+- ▁circus
+- ▁cotxe
+- ffi
+- kha
+- йым
+- ▁loaf
+- ▁analysis
+- ▁libre
+- सा
+- ▁bake
+- ▁sports
+- ▁package
+- ▁buena
+- ▁zimmer
+- mari
+- 其
+- ▁Quant
+- ▁나오
+- gleich
+- ▁Just
+- bizi
+- ▁joc
+- marca
+- cze
+- ▁insect
+- ▁paw
+- rida
+- ▁kennen
+- च
+- ▁divin
+- ▁sicher
+- ▁пред
+- urukundo
+- ▁exterior
+- èr
+- front
+- ▁شده
+- ▁doute
+- cción
+- nett
+- ▁fireplace
+- ▁interruption
+- ▁тады
+- ▁dearly
+- ▁flatter
+- ▁zaidi
+- écria
+- дын
+- шь
+- блі
+- вет
+- ▁domini
+- ▁такі
+- ▁hatch
+- ▁franklin
+- 세
+- 한데
+- ▁oppressed
+- ▁privileges
+- уп
+- ▁veure
+- lion
+- ▁oud
+- ап
+- ▁tira
+- makuru
+- ▁tafel
+- ▁clare
+- ▁mothers
+- ▁reins
+- ▁alguns
+- ▁voilà
+- ▁digging
+- ▁kvar
+- tojn
+- ▁пара
+- foje
+- ▁spending
+- зна
+- ▁Kiel
+- ▁pelo
+- ▁occurs
+- об
+- ager
+- ▁objected
+- ▁produces
+- ▁Ал
+- lton
+- ▁brace
+- ▁dopo
+- reta
+- ▁하지
+- ▁laut
+- ▁systems
+- lassen
+- fas
+- ibility
+- ै
+- ▁jumping
+- slag
+- ▁간
+- ▁helm
+- rew
+- ดี
+- esi
+- niej
+- ▁merri
+- blick
+- фи
+- ▁connu
+- ▁successive
+- regel
+- ▁tastes
+- వ
+- き
+- ▁pioneer
+- ▁brightness
+- ▁factor
+- ▁summons
+- лыш
+- tawa
+- 져
+- ▁stan
+- ▁nuko
+- chung
+- ▁kro
+- ▁scoundrel
+- ▁warum
+- ▁ёй
+- дні
+- ▁blocks
+- vina
+- ▁dragging
+- roll
+- ▁excel
+- ▁kazi
+- ▁treasury
+- če
+- ▁pill
+- gelijk
+- ▁sledge
+- ▁skins
+- quí
+- thorpe
+- ▁cooked
+- itwa
+- utsch
+- ▁equality
+- ▁hark
+- uzu
+- ▁refusal
+- ▁twist
+- ▁краіны
+- ▁concerns
+- bó
+- ים
+- likuwa
+- उ
+- ▁appreciation
+- 者
+- ▁portions
+- пада
+- bru
+- ▁excessive
+- จาก
+- ▁disco
+- ▁products
+- ▁ஏ
+- ▁reis
+- ▁بد
+- ▁distribu
+- ▁restraint
+- ▁하는데
+- word
+- ▁irresistible
+- shima
+- ▁letzten
+- ▁Tal
+- nym
+- ▁faithfully
+- ▁cyose
+- ▁kamen
+- ք
+- ▁jeden
+- ▁curve
+- ▁cómo
+- ▁sheer
+- ▁carta
+- ▁Ek
+- ▁haupt
+- ▁hopeful
+- ▁automobile
+- ▁Not
+- ▁jeanne
+- ▁moves
+- 來
+- yitibwa
+- ▁rack
+- doj
+- ▁tuesday
+- ัง
+- ▁수도
+- 날
+- aq
+- ▁করে
+- ▁Dilluns
+- ▁caravan
+- ▁spray
+- ▁நா
+- ança
+- ▁cinco
+- ▁Kandi
+- 걸
+- ▁zog
+- bracht
+- pte
+- iten
+- ▁shudder
+- ில்
+- ▁rely
+- ▁spit
+- ▁simpl
+- мет
+- ▁jimmie
+- ▁device
+- fekt
+- ▁bliss
+- ▁ведае
+- turn
+- ▁germ
+- iana
+- 무
+- ▁neuen
+- ▁leon
+- ▁inventor
+- igan
+- ▁заб
+- ык
+- ▁ص
+- ▁granite
+- eko
+- ಂ
+- פ
+- 合
+- ▁sands
+- ▁grinned
+- ▁Bel
+- ▁ziel
+- ▁verloren
+- ▁gush
+- ▁몰라
+- தை
+- ▁seemingly
+- kta
+- ▁obedient
+- ▁pedra
+- ▁accurate
+- ▁orden
+- ▁threat
+- ▁northward
+- colored
+- ▁groom
+- fet
+- ▁bags
+- ramu
+- ▁prote
+- tangiye
+- ▁cakes
+- ▁এবং
+- ходзіць
+- 박
+- лын
+- ▁zie
+- ▁bridle
+- ▁Ev
+- ▁mock
+- ▁gau
+- oda
+- ḥa
+- ▁alto
+- ▁treachery
+- sun
+- cina
+- truc
+- ம்ப
+- чных
+- ▁Pan
+- ▁choked
+- ▁Anem
+- ▁illustrious
+- ▁Hor
+- ▁withdrawn
+- ▁hermit
+- ▁kurya
+- ▁stubborn
+- ▁visto
+- به
+- ▁loyalty
+- ём
+- faction
+- ▁equipment
+- ▁sohn
+- pane
+- manyi
+- ▁wezen
+- baki
+- cule
+- ют
+- ▁sever
+- ▁haughty
+- ▁Two
+- ▁screw
+- เธอ
+- wil
+- ▁leaping
+- ▁substitute
+- ▁tavern
+- ▁casi
+- ния
+- ▁doctrines
+- ▁spoiled
+- ▁Fra
+- ҭ
+- ▁compel
+- ▁aren
+- որ
+- він
+- ▁districts
+- دم
+- gine
+- uq
+- ▁kneeling
+- ▁renew
+- rma
+- ▁slide
+- cola
+- ▁Far
+- rimu
+- ▁aught
+- neh
+- ▁nightingale
+- стро
+- autobús
+- ▁peculiarly
+- ▁doit
+- ▁placid
+- ▁Alle
+- ▁characteristics
+- ▁rotten
+- ▁kicked
+- ▁Те
+- kuzi
+- ▁trivial
+- impe
+- ண்ட
+- ▁detected
+- ▁comer
+- ছে
+- ▁loath
+- city
+- ▁sauce
+- ود
+- 와
+- ▁adorn
+- ▁ibibazo
+- ս
+- әр
+- ▁discharged
+- ▁soothing
+- vuze
+- ▁trials
+- ▁ми
+- ▁properties
+- lite
+- ▁조
+- ▁Non
+- ashaka
+- ▁organs
+- ▁pierced
+- ▁stockings
+- ▁tasted
+- buk
+- ▁crois
+- rach
+- ח
+- ▁expressing
+- ▁finn
+- ▁reflections
+- ▁hull
+- ▁zult
+- ▁편
+- ▁Dur
+- ▁classic
+- ▁Roman
+- ▁unkind
+- ▁tiam
+- ▁맛있
+- ▁hymn
+- ▁До
+- ▁destaca
+- ▁disposal
+- hoo
+- ▁هست
+- nova
+- taro
+- ▁edgar
+- ▁feat
+- rgen
+- ▁dret
+- ▁strove
+- כ
+- 초
+- riya
+- ▁ledge
+- grade
+- ▁بىلەن
+- ppi
+- ▁splash
+- ▁мог
+- ▁femmes
+- ograph
+- ▁compris
+- ▁працяг
+- Uni
+- ▁honesty
+- ▁homoj
+- ▁architecture
+- ый
+- ▁적
+- ваецца
+- ▁superintendent
+- ▁overwhelmed
+- ▁tonight
+- ▁fiend
+- ▁crab
+- ▁finer
+- iller
+- ▁diris
+- fini
+- ▁Grund
+- ▁annoyance
+- ldi
+- ▁प्र
+- ▁intently
+- ▁stride
+- ▁Ah
+- eki
+- ▁cradle
+- ▁preci
+- ▁vede
+- ▁rainbow
+- enz
+- ▁begann
+- ▁dined
+- ▁Schw
+- avoir
+- ▁нешта
+- ▁benjamin
+- 面
+- ▁dirigi
+- ▁дене
+- mming
+- mple
+- ppel
+- ▁Wenn
+- ▁eric
+- ۋ
+- ▁anterior
+- ▁questa
+- ▁trenches
+- ▁crush
+- وت
+- ር
+- ▁efficient
+- ▁roast
+- 님
+- anye
+- ▁momentary
+- ▁sect
+- แล้ว
+- ▁fitz
+- ▁clubs
+- ▁duly
+- ▁membre
+- тро
+- 立
+- ▁soap
+- rc
+- ▁isle
+- jer
+- ▁akazi
+- ▁baptism
+- chron
+- dores
+- ▁salary
+- ▁아니라
+- uso
+- 巴
+- ▁adequate
+- ▁Ass
+- 싸
+- 意
+- ▁чо
+- ▁campo
+- tono
+- pä
+- ▁Plan
+- ▁coil
+- ▁sleeve
+- ▁brengen
+- ▁extension
+- ▁pirates
+- ુ
+- rov
+- ▁инде
+- ▁schlecht
+- labor
+- gr
+- ▁convenience
+- ▁programme
+- ▁gorge
+- idades
+- ▁gerald
+- tausend
+- ▁penalty
+- مُ
+- ▁பா
+- fol
+- dha
+- 색
+- ▁gill
+- ▁conductor
+- ▁nails
+- рг
+- rima
+- ▁logs
+- ▁borders
+- ▁misma
+- ▁ghosts
+- bé
+- firma
+- bahn
+- ए
+- ▁cordial
+- ▁righteous
+- кой
+- ▁keenly
+- ▁вялікі
+- pple
+- ▁candid
+- ▁recruit
+- ର
+- ▁bug
+- ▁isaac
+- لەر
+- eɛ
+- chief
+- ▁stepping
+- wijze
+- ▁тры
+- médi
+- ▁layer
+- third
+- ▁drawer
+- ▁inspection
+- leb
+- ів
+- ▁natürlich
+- ▁lemon
+- ▁Пасля
+- lines
+- ▁James
+- ▁severity
+- ாள்
+- тка
+- ▁oba
+- ▁denk
+- rh
+- trop
+- ▁polish
+- ▁delayed
+- ▁procured
+- ▁mau
+- கொண்ட
+- gemaakt
+- ▁oven
+- dienst
+- 온
+- 发
+- ▁বি
+- pê
+- ▁hörte
+- clou
+- тэр
+- чым
+- ▁тор
+- ▁완전
+- ▁crop
+- dung
+- geschichte
+- ▁prices
+- ▁causing
+- ▁darf
+- ▁어제
+- orna
+- 任
+- alter
+- ös
+- ▁nari
+- 性
+- ook
+- ется
+- ▁chip
+- och
+- ▁jed
+- ▁intrigue
+- ▁label
+- altres
+- ndwa
+- ▁hamlet
+- ▁persist
+- ▁expedient
+- bita
+- fertig
+- hawe
+- grand
+- ▁wisely
+- ▁byiza
+- ▁lighter
+- ▁potent
+- ▁carrera
+- ▁dagger
+- posta
+- ▁patriot
+- ▁intolerable
+- ▁그래가지고
+- ▁vexed
+- tini
+- ▁naam
+- tempe
+- ▁sham
+- ▁Jahre
+- ▁weariness
+- ▁countrymen
+- ▁elevation
+- ▁vegetable
+- еи
+- aged
+- ▁했어
+- яць
+- remo
+- kni
+- ▁Kinder
+- ▁antwoord
+- ваюць
+- ▁ekster
+- ▁다시
+- ▁ҡара
+- ▁ribbon
+- eter
+- ▁bred
+- ▁realise
+- ▁poll
+- ▁numero
+- ение
+- ▁Seine
+- ख
+- lett
+- fond
+- genomen
+- ▁storms
+- trat
+- 했어
+- што
+- ▁groan
+- င်
+- chos
+- мін
+- ▁neb
+- ▁taxes
+- llar
+- kiye
+- ▁socialist
+- ▁verdict
+- ▁Kwa
+- ▁influenced
+- ▁receipt
+- ▁forthwith
+- ▁clark
+- ▁satin
+- kombe
+- ▁eleventh
+- ancia
+- ▁адна
+- itch
+- ▁schooner
+- bh
+- aquell
+- gab
+- ▁sway
+- ▁wednesday
+- 보고
+- ▁attendance
+- ▁ру
+- ัก
+- ▁denen
+- ▁roland
+- ▁traditions
+- مل
+- ▁visage
+- ▁shattered
+- ambu
+- ▁jenny
+- bou
+- ▁spectators
+- ▁novels
+- had
+- ▁reprit
+- ▁indulgence
+- ▁philosophical
+- ▁soldat
+- iras
+- ▁entry
+- 体
+- ▁wales
+- raad
+- fran
+- ▁connect
+- ▁drunken
+- bett
+- ▁unexpectedly
+- ▁battalion
+- ާ
+- ▁zijt
+- ▁locks
+- ppy
+- ▁Nova
+- ▁devils
+- 싶
+- ▁пре
+- ▁sneer
+- ▁relating
+- itz
+- ▁pickwick
+- ▁resulted
+- ▁général
+- ▁blijven
+- cula
+- న
+- menti
+- вацца
+- ▁brad
+- ▁wawe
+- መ
+- ▁schöne
+- ▁Ces
+- زن
+- ▁embarrassed
+- ▁pant
+- ▁genau
+- வர
+- هُ
+- ▁finance
+- ▁sacrifices
+- ▁altre
+- ▁figura
+- het
+- abana
+- ▁입
+- ▁hasten
+- аецца
+- ▁panting
+- ubwo
+- ດ
+- ▁foremost
+- mpel
+- ▁chum
+- ▁noisy
+- ünde
+- ▁babiri
+- ▁kutoka
+- ▁먹어
+- fik
+- ▁submission
+- ▁depression
+- dolph
+- denial
+- pè
+- activ
+- ▁inhabited
+- ▁상
+- ▁exemple
+- rock
+- ▁bil
+- eria
+- wall
+- ▁forgiven
+- ▁성
+- ವ
+- cato
+- ▁droit
+- imiento
+- ▁Dy
+- だ
+- badde
+- ▁periods
+- ▁fluid
+- ▁peine
+- ▁Mat
+- ark
+- пам
+- ▁grievous
+- ▁pry
+- Est
+- ▁breaks
+- ▁athos
+- 属
+- nies
+- ▁wagen
+- ކަ
+- ▁imprisonment
+- ▁tearing
+- ީ
+- ▁habla
+- ▁stumbled
+- ▁fancies
+- ▁specimen
+- ▁davant
+- ▁submarine
+- ▁blessings
+- uwa
+- ▁emergency
+- ▁refrain
+- ▁kort
+- eira
+- ▁galley
+- ▁med
+- ▁grieved
+- last
+- ▁shadowy
+- ▁figli
+- ▁distin
+- ▁subjected
+- தான்
+- ▁pursuing
+- 올
+- ▁Muri
+- ▁볼
+- ▁identical
+- یک
+- ▁grab
+- 해가지고
+- ▁cups
+- ▁lydia
+- ▁mieux
+- ▁target
+- ▁assign
+- ▁equivalent
+- ▁bullets
+- ない
+- ▁buying
+- wright
+- ▁riu
+- ▁torch
+- ▁massa
+- ▁phrases
+- ▁dachte
+- ▁breadth
+- ▁fairies
+- ފަ
+- school
+- lette
+- রে
+- ▁rumour
+- ▁antwortete
+- ▁shrank
+- ▁communion
+- ▁û
+- ▁unite
+- شک
+- ▁geben
+- მა
+- 차
+- ▁distribution
+- ▁insane
+- ▁darkened
+- ერ
+- ▁nabo
+- ▁cylinder
+- ▁ultimately
+- ючы
+- ▁mound
+- ▁cedar
+- ▁Australia
+- ständig
+- ▁uncommon
+- ▁transformed
+- ▁gasp
+- மே
+- bala
+- öpf
+- ▁venir
+- ▁dienst
+- lacht
+- hot
+- 린
+- cía
+- ▁들어가
+- ▁appearances
+- ▁striving
+- ▁bounds
+- ▁così
+- ▁ок
+- ▁vois
+- ▁junior
+- ▁crisp
+- ង
+- xes
+- ▁micro
+- ▁poles
+- ▁crooked
+- ▁circum
+- dron
+- reden
+- ▁hanze
+- ▁currently
+- ▁bland
+- вал
+- 갖고
+- bril
+- ▁wengi
+- ▁rapport
+- ▁خوب
+- ▁роз
+- voll
+- schrift
+- ▁spark
+- ▁alia
+- ▁reed
+- ▁hounds
+- iciro
+- kraft
+- ▁colonial
+- ▁manifested
+- ▁tiel
+- ნ
+- ▁kaiser
+- ވެ
+- hura
+- ▁llarg
+- abyo
+- ▁ravine
+- ▁partially
+- ▁lark
+- ▁falta
+- ▁மா
+- ▁plume
+- ▁feverish
+- ▁noir
+- elte
+- ąc
+- giu
+- ▁rwo
+- ▁faisait
+- ım
+- нең
+- ▁reaches
+- ▁ibikorwa
+- igh
+- ாய்
+- ▁gesp
+- ▁damsel
+- ▁адказ
+- ▁superb
+- ▁Well
+- ▁Dia
+- dini
+- ▁oxen
+- 现
+- ould
+- ▁zacht
+- ▁etwa
+- 力
+- ▁lacking
+- ▁southward
+- ▁furiously
+- uo
+- дна
+- ▁victorious
+- siz
+- ydi
+- ▁sollten
+- ▁reject
+- ▁arte
+- aña
+- чна
+- ▁lever
+- ▁petits
+- ▁Las
+- ▁impamvu
+- ▁phantom
+- ▁술
+- ות
+- ▁mourn
+- ▁shades
+- rino
+- ▁Actualment
+- ▁explaining
+- ▁quote
+- ▁processes
+- ▁divert
+- ▁iom
+- 代
+- face
+- ▁uli
+- ▁tornar
+- ▁constitutional
+- ▁engines
+- пры
+- ▁rubber
+- ▁bron
+- tischen
+- ▁już
+- ▁pyramid
+- bwenge
+- ▁مح
+- ыс
+- ▁lewis
+- ▁furnace
+- ▁mugihe
+- ▁gekomen
+- ▁stripped
+- हा
+- ▁banker
+- ▁edel
+- kia
+- ▁stress
+- ▁verde
+- ▁himmel
+- ▁indulge
+- ▁coral
+- 울
+- ▁shrine
+- trice
+- ▁manda
+- ▁rounded
+- sept
+- ▁profess
+- ▁vraag
+- ▁foliage
+- ▁extremity
+- ▁gaining
+- “.
+- ▁ornaments
+- ▁conversion
+- ▁nap
+- 있
+- ▁tisa
+- mov
+- ▁byari
+- ▁dentro
+- jou
+- ▁выступ
+- ▁robbers
+- aquestes
+- ▁Mac
+- ან
+- bund
+- ▁ekzistas
+- ▁contradict
+- empre
+- ▁joys
+- ހު
+- tada
+- ▁preference
+- yor
+- эш
+- ▁plötzlich
+- ▁assuredly
+- ▁legislature
+- ▁clam
+- ▁hof
+- ▁obstacle
+- nyura
+- ▁wald
+- ▁outfit
+- vad
+- formed
+- ▁roam
+- ▁earthquake
+- ▁rags
+- ▁sanction
+- tang
+- ▁staggered
+- ogen
+- ▁defined
+- ▁Hol
+- ▁interposed
+- ▁sara
+- ▁gezien
+- ▁seule
+- ▁covers
+- ▁pest
+- ▁wakati
+- 프
+- arte
+- ▁confounded
+- ▁settlers
+- ▁wondrous
+- ▁obstinate
+- ция
+- ▁conventional
+- gesellschaft
+- ▁ہے۔
+- rere
+- след
+- ▁darüber
+- 하면은
+- ▁contradiction
+- ▁봤
+- ▁travels
+- ま
+- ▁horsemen
+- refu
+- ▁lions
+- wd
+- ▁Cel
+- ▁basil
+- ▁deposited
+- ▁negoci
+- 복
+- ▁sped
+- fanya
+- sted
+- 嘅
+- 次
+- ▁davis
+- ▁ropes
+- ▁führte
+- դ
+- ▁archer
+- ▁ĉiu
+- ition
+- ่ง
+- ▁flint
+- ▁repentance
+- 하면서
+- gisha
+- ▁Star
+- ▁konata
+- হা
+- embre
+- ров
+- ▁temporal
+- 추
+- ▁twinkle
+- ▁fertile
+- ▁suck
+- ▁urban
+- ▁hats
+- 준
+- եր
+- tou
+- ▁assented
+- ▁maka
+- пля
+- ▁indispensable
+- ▁Perezida
+- ▁begging
+- cado
+- chamber
+- aid
+- ▁rubbing
+- ▁blossoms
+- ▁swam
+- bian
+- ▁robinson
+- aime
+- ▁loneliness
+- kuri
+- ▁areas
+- ▁reside
+- folge
+- 外
+- ▁gaat
+- hér
+- ▁announce
+- ▁imprisoned
+- ütz
+- ▁gingen
+- ▁seizing
+- ▁discoveries
+- ▁sec
+- фа
+- ▁uneasiness
+- ▁зусім
+- ▁guten
+- ▁rattle
+- ின்
+- პ
+- ▁gesehen
+- ▁Yesu
+- ▁romp
+- ▁아니면
+- ▁tiun
+- ▁afar
+- دن
+- ▁tener
+- ▁touches
+- ▁strings
+- ▁forgiveness
+- ▁courtyard
+- ▁reigned
+- ıl
+- racy
+- је
+- rara
+- hö
+- ▁Charles
+- yinza
+- ▁survive
+- ▁зараз
+- ▁physic
+- بِ
+- тып
+- read
+- vý
+- ▁suspense
+- ▁aloft
+- ште
+- кала
+- ▁سەر
+- emu
+- ▁kunst
+- ▁trunks
+- ▁있었
+- ਹ
+- β
+- нікі
+- 過
+- ▁beams
+- ▁concentrated
+- ▁costly
+- ▁plead
+- ോ
+- ▁remnant
+- krank
+- ▁wereld
+- ▁Wy
+- kamp
+- ▁inspire
+- ène
+- body
+- ئو
+- leu
+- ▁trim
+- ▁robbery
+- chas
+- ▁discussing
+- ฟ
+- ▁mentre
+- ▁hoarse
+- ▁freight
+- ▁جي
+- зда
+- tual
+- ▁Nur
+- neu
+- ▁organism
+- ▁determina
+- ▁дзень
+- ▁edges
+- ▁قىل
+- ▁void
+- rana
+- ▁sacrificed
+- ▁advertisement
+- ▁players
+- komeje
+- ▁repaired
+- скага
+- île
+- ▁adorned
+- ▁bored
+- ткі
+- gabe
+- ▁statements
+- ▁holmes
+- ▁이러
+- ▁intimacy
+- crea
+- ▁disdain
+- 하다
+- ehrt
+- pá
+- ▁kara
+- вых
+- ▁penetrate
+- mac
+- ▁cac
+- ▁και
+- داری
+- trä
+- ▁objections
+- ▁clatter
+- antic
+- anche
+- ▁vocal
+- ▁repu
+- lateral
+- ▁adventurer
+- ▁مە
+- ▁cecilia
+- 재
+- kala
+- ▁desp
+- சா
+- ▁yw
+- schlag
+- yama
+- ▁siaj
+- ု
+- gelegen
+- மாக
+- ▁doomed
+- ▁crude
+- ▁verme
+- ▁rejoin
+- dow
+- ▁waistcoat
+- ێکی
+- eqq
+- ทํา
+- mies
+- ▁World
+- ًا
+- ▁سال
+- issait
+- ▁softened
+- ான
+- ŭ
+- ▁vainly
+- cip
+- ▁dafür
+- 면서
+- ▁grimly
+- ▁partake
+- ▁públic
+- чак
+- klop
+- нос
+- 位
+- ▁hardy
+- ▁خواهم
+- ▁ernest
+- ▁Ny
+- assi
+- ▁خا
+- ▁pike
+- ▁kitten
+- ▁crawled
+- ▁ami
+- ▁conclusions
+- 가지구
+- ▁igual
+- ▁nursery
+- ▁steeds
+- ▁pipes
+- ggi
+- ▁Pal
+- ▁wagons
+- ▁cured
+- ncia
+- kanya
+- legi
+- mill
+- ▁pleading
+- ▁과
+- ▁inyuma
+- morph
+- ម
+- 짜
+- 급
+- ▁quarante
+- ▁barber
+- ▁slopes
+- ▁llibre
+- tli
+- tah
+- ▁freund
+- ▁presenta
+- ▁abe
+- ▁greet
+- ▁hinein
+- વ
+- ▁splendor
+- ▁remarkably
+- 었는데
+- ▁bucket
+- ▁flot
+- ▁zooveel
+- ▁flows
+- kubwa
+- ▁neighboring
+- ▁የ
+- ▁painfully
+- ▁zer
+- ▁recur
+- ▁jaar
+- ▁molts
+- ▁muerte
+- impo
+- ス
+- ބަ
+- ▁persones
+- ▁offend
+- ▁akka
+- acord
+- ▁legitimate
+- ▁migra
+- combe
+- ние
+- 력
+- кт
+- ▁subordinate
+- ▁Are
+- ▁cherished
+- nzo
+- ▁engel
+- ▁rye
+- ▁né
+- ▁elevated
+- mula
+- ▁valiant
+- ▁кор
+- ▁funds
+- muka
+- ▁signifas
+- ▁scenery
+- escu
+- quel
+- ▁lá
+- ▁ĉirkaŭ
+- ▁sobbing
+- ▁Leta
+- ▁summon
+- тоў
+- ▁coolly
+- mena
+- ▁proclamation
+- ▁Haupt
+- ▁signature
+- ▁preceded
+- ▁marius
+- pio
+- tika
+- ▁vegades
+- ▁bazi
+- ▁bribe
+- thu
+- ▁Europe
+- ▁springing
+- ▁recovery
+- ▁kissing
+- going
+- ▁sinner
+- ▁cautious
+- rist
+- md
+- ▁yielding
+- ც
+- ▁beech
+- cious
+- ▁êtes
+- ▁carne
+- ▁paragraph
+- тура
+- ▁fosse
+- ▁instal
+- ▁completa
+- aves
+- वि
+- ▁guarantee
+- ▁click
+- ▁fos
+- ▁lawyers
+- zähl
+- మ
+- ▁때문에
+- ▁rugged
+- igihugu
+- ▁esa
+- nął
+- ▁bati
+- ▁tourist
+- sima
+- 時
+- ▁чаго
+- ▁garb
+- bwiye
+- ▁controlled
+- zungu
+- ▁unusually
+- schloß
+- ▁watchful
+- ▁caps
+- ▁boards
+- ▁dua
+- ▁respective
+- фор
+- ▁cinquante
+- ▁abrupt
+- थ
+- nomen
+- ▁passenger
+- ▁zitten
+- ▁fearless
+- ▁món
+- ▁herrn
+- ▁nit
+- gericht
+- wadde
+- ▁tumbled
+- ▁tanta
+- igte
+- iere
+- organ
+- ▁Then
+- ▁typical
+- ▁possesses
+- ▁definitely
+- ▁tama
+- ismus
+- bald
+- ▁Iyi
+- mbere
+- rut
+- ▁salute
+- 入
+- volle
+- ▁seis
+- ▁lava
+- ▁aujourd
+- ▁staan
+- ▁refusing
+- реш
+- ▁fathom
+- houden
+- ▁공부
+- 四
+- ▁majestic
+- chie
+- plu
+- ▁sólo
+- ▁specimens
+- ▁instincts
+- ▁reckoned
+- ▁이번에
+- ▁highland
+- ßen
+- ▁Ale
+- ▁Tour
+- ▁knocking
+- ▁wallace
+- ▁hector
+- ▁mwiza
+- ▁intensely
+- ▁resume
+- ▁stormy
+- ▁corporal
+- ▁otras
+- ▁sensations
+- ▁pitt
+- ▁lament
+- ▁paddle
+- ▁convi
+- ▁você
+- ▁tyranny
+- ▁gaily
+- тыш
+- ▁isolated
+- impose
+- ▁maintenant
+- ▁cupboard
+- ▁hir
+- ▁nod
+- ▁regal
+- ▁befindet
+- ▁removal
+- ▁lw
+- 이나
+- جی
+- ält
+- illas
+- ▁ang
+- vra
+- 站
+- ▁여기
+- rmi
+- ▁regent
+- ▁telèfon
+- ▁ivory
+- ▁abbot
+- ▁beforehand
+- nac
+- schaffen
+- ▁Bay
+- ▁되는데
+- әү
+- ▁kisses
+- مَ
+- ▁кара
+- ccio
+- ▁disk
+- ▁lapse
+- য়ে
+- wandi
+- ▁hii
+- ▁chronicle
+- baho
+- زى
+- ▁babylon
+- ▁chancellor
+- jur
+- мовы
+- achtig
+- ▁proofs
+- mile
+- ▁survived
+- کو
+- amara
+- меж
+- ▁speck
+- ▁welcher
+- йшлі
+- gegangen
+- ▁shuddered
+- lb
+- ▁telescope
+- ▁marque
+- ▁кил
+- ▁ibya
+- ▁backwards
+- name
+- ▁proclaim
+- gali
+- ▁earnestness
+- labora
+- stoff
+- ▁corrupt
+- ▁rusange
+- fach
+- ▁shrink
+- ▁luz
+- ▁лепш
+- druck
+- ▁uncertainty
+- ▁herds
+- umukobwa
+- ▁cricket
+- ▁shabby
+- pion
+- لِ
+- ▁تۆ
+- suit
+- cart
+- دو
+- ▁bana
+- ▁overwhelming
+- ▁resign
+- ▁estado
+- beth
+- ▁früh
+- ▁amaso
+- գ
+- rec
+- ▁moto
+- ▁Бер
+- ▁resignation
+- ules
+- ิน
+- ▁dama
+- origine
+- ▁rapidity
+- kono
+- ▁modesty
+- ▁права
+- вання
+- ▁array
+- ▁partial
+- ▁사람들이
+- ▁debts
+- garten
+- ▁Wat
+- ▁prepara
+- gend
+- ▁bleeding
+- ▁inflicted
+- ▁Pere
+- ▁angst
+- hne
+- ▁muito
+- щи
+- église
+- ▁exclusive
+- зо
+- او
+- kee
+- iron
+- дра
+- ▁longue
+- 士
+- ▁shaped
+- spirited
+- ▁exercises
+- ▁sincerely
+- িক
+- ▁mà
+- вялі
+- ▁amateur
+- ыч
+- ▁hudson
+- alina
+- ▁slew
+- koze
+- 그
+- енә
+- ▁glück
+- ën
+- ▁pep
+- ใจ
+- ▁erhalten
+- ▁puritan
+- ▁제일
+- ▁cling
+- ▁Us
+- ▁쓰
+- halb
+- ending
+- ышт
+- อย่าง
+- ▁decidi
+- almente
+- से
+- stellt
+- ▁fino
+- ▁Jen
+- ▁тип
+- introdu
+- ▁cris
+- ▁scholars
+- قان
+- mada
+- ntes
+- ▁occhi
+- ▁mogelijk
+- stunde
+- ▁Other
+- گەن
+- ▁perilous
+- ▁그렇
+- zam
+- ▁igihugu
+- رس
+- qqim
+- tempered
+- ▁protector
+- あ
+- ▁sembla
+- ▁loft
+- ▁nag
+- ▁Sub
+- ▁명
+- ▁ankle
+- ▁kabiri
+- ▁biscuit
+- ▁diameter
+- ▁vapor
+- ckt
+- қәа
+- zette
+- ▁presque
+- ່
+- ▁grotesque
+- ▁mules
+- ▁architect
+- ▁packet
+- ▁rol
+- ▁пі
+- ▁Dis
+- ▁predi
+- ▁algunos
+- ▁ваш
+- ▁specio
+- ▁habían
+- sang
+- ▁caroline
+- équipe
+- ▁May
+- αν
+- ▁lone
+- 씩
+- ▁deem
+- кө
+- ▁meisten
+- tib
+- écri
+- ▁heißt
+- лт
+- ▁его
+- ತ
+- ▁wien
+- angle
+- ▁volas
+- ▁Mc
+- ▁eastward
+- 东
+- ▁kru
+- ▁smote
+- baix
+- ться
+- ▁roofs
+- ▁strode
+- 心
+- ▁dreadfully
+- год
+- ▁conspiracy
+- mato
+- ▁wegen
+- ▁augustus
+- аль
+- tide
+- ▁fishes
+- ▁mature
+- ▁East
+- 二
+- छ
+- preservation
+- ەکان
+- ece
+- ԥ
+- ▁Josep
+- ść
+- нуў
+- dauer
+- ▁nervously
+- ▁divinity
+- ▁часу
+- rack
+- ▁mug
+- ▁regiono
+- ▁rebels
+- ▁essere
+- yat
+- ▁viola
+- ▁vino
+- ▁endowed
+- ▁torre
+- ▁deserves
+- jí
+- ▁ending
+- ▁cœur
+- ▁tops
+- ▁beware
+- ▁sections
+- ▁fiddle
+- ifies
+- nian
+- ▁gele
+- ▁inasmuch
+- ▁espi
+- ▁ubwa
+- gada
+- ▁planning
+- 此
+- ▁বা
+- ▁kindred
+- лым
+- ▁interference
+- ▁heaps
+- пера
+- ▁brachte
+- ▁hesitating
+- 相
+- ▁pitched
+- ▁bounded
+- ▁usage
+- ▁wizard
+- ▁governess
+- sinzi
+- lernejo
+- mbra
+- ▁kinderen
+- ▁deceased
+- 등
+- ▁biggest
+- ▁flaming
+- ▁ammunition
+- ▁Louis
+- ▁pea
+- ▁schemes
+- 레
+- ▁ceremonies
+- ▁petr
+- ▁neutral
+- rong
+- ▁viva
+- ▁فر
+- ▁zaak
+- jyana
+- bearer
+- ▁passar
+- gwe
+- ▁envoy
+- garagara
+- ▁Су
+- ▁smoked
+- ▁걔가
+- ▁laboris
+- มาก
+- 属下的一个
+- 되
+- ▁folds
+- ▁elk
+- 빠
+- ▁representing
+- ד
+- ▁sentinel
+- ▁King
+- લ
+- ▁curled
+- ▁Eki
+- ез
+- ▁catastrophe
+- ▁왜냐면
+- tili
+- ▁bamwe
+- กัน
+- หา
+- ▁undertaken
+- hmm
+- ▁Ol
+- лег
+- 집
+- ▁indien
+- ▁English
+- ▁infernal
+- ▁Kur
+- sió
+- ▁parade
+- mobil
+- amerika
+- ▁matches
+- gelegt
+- ▁여
+- Ba
+- ▁creu
+- uwen
+- ulated
+- ▁hire
+- ▁competi
+- cco
+- ▁boughs
+- vies
+- nywe
+- 因
+- ▁rifles
+- vera
+- ▁heed
+- ▁اور
+- ▁aha
+- tics
+- ▁podía
+- ▁lieber
+- ▁یا
+- тал
+- ▁tram
+- ▁junto
+- ▁demi
+- ▁wichtig
+- ং
+- ވެސް
+- ▁camera
+- ▁sheltered
+- ▁jud
+- ▁saß
+- ▁massacre
+- ▁fisher
+- ttes
+- ▁bijna
+- лер
+- нер
+- ▁wand
+- ▁troy
+- ▁curs
+- ▁yankee
+- स्
+- ▁далей
+- ▁boarding
+- ▁chocolate
+- ▁antonio
+- 적으로
+- гар
+- mania
+- 手
+- tsinda
+- escola
+- ▁strengthened
+- ▁Jun
+- ▁perceiving
+- ▁necklace
+- ▁furent
+- من
+- ▁underground
+- ▁lavish
+- аваў
+- дык
+- ▁nuestra
+- ▁exceeding
+- leh
+- ▁evenings
+- ▁nosotros
+- mí
+- ▁importante
+- ▁represents
+- ▁vieux
+- ▁ohio
+- цаў
+- ▁rebuke
+- ліва
+- ▁Spa
+- nous
+- tuur
+- unu
+- ▁imbaraga
+- ▁너는
+- 주고
+- мага
+- ስ
+- flower
+- ▁wig
+- еит
+- ▁precipice
+- rose
+- ▁sunrise
+- іс
+- ▁inherited
+- ▁ஒ
+- ▁sogar
+- worn
+- ▁continuing
+- ▁lungs
+- সে
+- წ
+- 코
+- ▁perché
+- ▁Schul
+- ▁ferry
+- ▁burial
+- spann
+- ▁rulers
+- ▁relacion
+- ▁gallop
+- ▁promote
+- говор
+- ▁imana
+- oest
+- tanto
+- аю
+- ▁früher
+- ▁plateau
+- ▁gradual
+- ▁정말
+- ▁stool
+- kamer
+- ▁adieu
+- ▁pots
+- szt
+- ▁melt
+- ▁amigo
+- ษ
+- ▁του
+- ▁слова
+- іць
+- ▁expressive
+- ▁pir
+- ▁mall
+- stick
+- ▁zon
+- выя
+- ать
+- ύ
+- ▁나가
+- ▁charter
+- اً
+- ▁cass
+- ké
+- ambul
+- ▁speculation
+- lug
+- ▁occasioned
+- ्य
+- cible
+- ifi
+- ▁trait
+- ▁burns
+- ▁alteration
+- ▁dagen
+- ▁taylor
+- ிற
+- ▁passive
+- ▁план
+- ▁inflict
+- ▁apology
+- ▁portal
+- ▁peel
+- kozi
+- ciu
+- univers
+- ▁갑자기
+- ruḥ
+- ▁plump
+- ▁kennedy
+- ▁wert
+- ši
+- ▁rapture
+- 台
+- ▁judging
+- ▁donné
+- ም
+- uzuye
+- ▁habitation
+- ▁bower
+- ▁vein
+- ▁wunder
+- ிருந்த
+- ▁snug
+- ▁Sein
+- ▁revived
+- ▁handen
+- ▁garde
+- hali
+- ▁whistling
+- ▁sobbed
+- guye
+- ▁nemen
+- ▁copies
+- ▁banner
+- esser
+- மி
+- ▁critics
+- ▁Hall
+- люб
+- ڈ
+- ▁perplexed
+- 청
+- ▁mechanism
+- ▁mio
+- ▁nuts
+- ▁turkish
+- سه
+- ships
+- ▁lingering
+- shore
+- ▁що
+- ▁gufata
+- ▁barre
+- bino
+- raf
+- ▁solar
+- ▁prostrate
+- ▁contrari
+- ▁outrage
+- ▁enclosed
+- ▁күр
+- யே
+- ▁guidance
+- ▁irr
+- ▁suggestions
+- lyn
+- ▁vond
+- ▁nostrils
+- mico
+- ▁cielo
+- ▁nowadays
+- бір
+- ▁dezen
+- ார்கள்
+- اک
+- moc
+- ▁perquè
+- кана
+- ண்டு
+- ▁эс
+- ▁launch
+- ▁multi
+- Argent
+- ▁кеше
+- ▁saul
+- tile
+- ▁Flor
+- ▁Tim
+- ▁Parlament
+- นา
+- ▁répond
+- quo
+- zicht
+- boro
+- ▁shorter
+- akt
+- herr
+- ▁marian
+- 연
+- 모
+- ▁están
+- ▁draws
+- ▁theater
+- ▁traveled
+- oti
+- ▁nobleman
+- ▁karl
+- ▁arriba
+- iam
+- fili
+- ▁misfortunes
+- خوا
+- ▁besonders
+- ▁Ibyo
+- мак
+- itza
+- ▁shots
+- ▁coroner
+- зам
+- ▁harriet
+- ছিল
+- ▁bachelor
+- Co
+- ▁ĉu
+- ▁muchos
+- 見
+- ▁worte
+- پە
+- की
+- naga
+- füg
+- ▁carl
+- زار
+- ▁sancho
+- ▁prophecy
+- rance
+- ▁supp
+- eft
+- 속
+- ▁States
+- dded
+- ▁없는
+- ▁stages
+- ▁umuryango
+- ▁vigor
+- ▁abaturage
+- ▁stammered
+- ues
+- ▁lou
+- ழு
+- ະ
+- इ
+- હ
+- 可以
+- ▁considerations
+- ▁pillars
+- ovitch
+- بت
+- ▁tus
+- ▁tense
+- ▁nekk
+- ▁hamilton
+- பெ
+- ebwa
+- ▁escaping
+- iel
+- ▁penetrated
+- ▁першы
+- ▁boca
+- anne
+- ▁Amerika
+- ▁incense
+- ▁luft
+- ▁zik
+- ▁Bet
+- സ
+- ▁grond
+- ▁glücklich
+- ▁plaster
+- ▁fork
+- ▁elkander
+- iber
+- ▁varying
+- cess
+- ▁repetition
+- organisation
+- ▁pode
+- ▁ware
+- ▁oyo
+- με
+- って
+- ▁impart
+- ▁전에
+- lif
+- هن
+- ▁dialect
+- fici
+- ▁días
+- fellow
+- ▁foar
+- fait
+- ▁uur
+- roth
+- ▁arg
+- ▁dimensions
+- ▁ssente
+- дым
+- ▁costa
+- ▁ведаю
+- ▁jupiter
+- ▁sculpture
+- ▁leva
+- மான
+- ▁gad
+- ▁yoke
+- ▁bavuga
+- afrika
+- ▁buen
+- قۇ
+- مو
+- action
+- ▁zone
+- szcz
+- վ
+- แต่
+- ރި
+- ▁cue
+- เล
+- ▁verily
+- ▁pigs
+- ىدۇ
+- ס
+- ▁democracy
+- loopen
+- ▁entrar
+- ▁rats
+- holm
+- ▁erg
+- ▁spears
+- ▁cracked
+- ▁타
+- ▁Tag
+- tato
+- ▁Det
+- ▁forlorn
+- voor
+- sein
+- ▁என்ற
+- ら
+- ruhande
+- ▁minder
+- ▁cavalier
+- ▁proceeds
+- ▁houden
+- ▁substances
+- 全
+- ▁slap
+- ▁traditional
+- ▁trou
+- ▁mooi
+- ▁revis
+- ▁passionately
+- alde
+- dressed
+- ▁acquaintances
+- ▁politely
+- änd
+- ▁año
+- 這
+- operate
+- ▁decoration
+- ▁pines
+- boko
+- ник
+- ▁oscar
+- ▁elf
+- ▁செ
+- ▁myth
+- ury
+- fati
+- ▁unver
+- дум
+- ▁lighting
+- ▁manos
+- lange
+- ▁fulfil
+- торы
+- ▁Pot
+- ▁sewing
+- ▁shilling
+- က
+- ▁encouraging
+- 현
+- ▁Uru
+- ▁rescued
+- ssar
+- ▁scrub
+- ▁Hab
+- ▁sulla
+- altro
+- bad
+- enco
+- mmed
+- ▁ieder
+- ▁pela
+- ▁남
+- ▁zegt
+- ▁delusion
+- kup
+- ▁pourquoi
+- ▁amas
+- ▁fuera
+- ▁mournful
+- tait
+- ▁afflicted
+- гэ
+- '000'
+- ▁Sus
+- ▁lifetime
+- ▁refreshment
+- ▁melody
+- ▁Africa
+- ▁regulate
+- ▁hive
+- ▁Twa
+- ▁diseases
+- trud
+- ixe
+- control
+- ▁oog
+- ▁resolute
+- ҵ
+- ергә
+- ▁meta
+- ▁Fou
+- ▁morn
+- ▁espa
+- ▁hosts
+- ▁puts
+- ▁oblige
+- ▁chor
+- ▁نه
+- heer
+- ▁exempl
+- ▁glee
+- ▁intens
+- ▁cousins
+- ად
+- ▁Yagize
+- ▁umugore
+- ▁August
+- ▁elders
+- ប
+- ▁problema
+- gö
+- ▁hijos
+- kleed
+- ▁irgend
+- த்தி
+- ▁Cada
+- ▁percy
+- marin
+- ▁moist
+- ▁Fort
+- ▁veteran
+- ▁englishmen
+- ▁Why
+- ▁carriages
+- ▁compliments
+- ▁força
+- ▁lado
+- ▁supporting
+- ▁obstant
+- ▁beseech
+- ▁antagonist
+- ▁tipus
+- ▁Ми
+- tê
+- ▁robber
+- grün
+- ▁mistakes
+- ▁petersburg
+- ▁countless
+- дня
+- 야지
+- ނެ
+- ș
+- ▁혼자
+- acha
+- rito
+- ▁weitere
+- ▁Hy
+- ▁leo
+- vir
+- rud
+- ▁eccentric
+- ▁expose
+- ▁valid
+- ▁cran
+- беҙ
+- ▁nyinshi
+- డ
+- ▁loi
+- ▁unaware
+- kota
+- ▁honoured
+- ▁quitted
+- ukira
+- ρα
+- льны
+- élé
+- hä
+- ▁neɣ
+- ▁conseil
+- ŝa
+- مون
+- غۇ
+- ▁propriety
+- ▁screaming
+- eq
+- ▁regretted
+- tira
+- ▁brighter
+- bec
+- clos
+- ffy
+- ప
+- ▁chalk
+- ifica
+- ▁journalist
+- lieu
+- ▁conti
+- honneur
+- سب
+- sio
+- ▁handled
+- ▁jaro
+- ▁amaze
+- ▁monotonous
+- क्
+- ▁covenant
+- avons
+- thra
+- ▁그거는
+- ▁adds
+- abili
+- ▁suits
+- brand
+- ▁excellency
+- ▁herzen
+- ే
+- ▁luxurious
+- gabo
+- ува
+- ▁건데
+- lada
+- ▁professed
+- ມ
+- 거나
+- ▁decorated
+- cultura
+- ksa
+- ▁fade
+- teg
+- shwa
+- ▁famine
+- ேன்
+- serva
+- vez
+- ณ
+- ▁geoffrey
+- ▁bail
+- ▁resp
+- rinda
+- ▁Gemeinde
+- heb
+- ▁lili
+- rf
+- ▁adhere
+- ▁swarm
+- Sa
+- ▁retreated
+- ▁coats
+- ം
+- ում
+- ables
+- ▁blot
+- think
+- fli
+- ▁morte
+- ▁relaxed
+- ▁eugene
+- льныя
+- ▁Haus
+- hinga
+- ▁parc
+- ▁nehmen
+- ▁parle
+- ޑ
+- ▁assumption
+- 처럼
+- ▁candidate
+- ▁waarin
+- ը
+- ▁carro
+- ▁snuff
+- ▁headed
+- ▁fraud
+- ▁objective
+- ▁piercing
+- ▁mantel
+- ▁confine
+- ▁matron
+- ances
+- ▁bleiben
+- ▁Club
+- ▁од
+- ▁restrained
+- ▁Tak
+- ément
+- 内
+- ▁grieve
+- ▁mills
+- ▁tries
+- ▁inconvenience
+- xu
+- uck
+- ▁caballero
+- ёр
+- funa
+- rás
+- وه
+- 눈
+- ▁celestial
+- ▁fli
+- kiko
+- ▁eût
+- harmoni
+- ▁admirer
+- ▁괜찮
+- ▁weib
+- ▁homage
+- ▁exposure
+- reḍ
+- lingen
+- ▁obstacles
+- ▁Het
+- ▁janet
+- дай
+- нар
+- ▁게임
+- ▁dodge
+- كَ
+- ▁afge
+- ▁bloß
+- ▁despatch
+- raient
+- ▁messages
+- ▁hermano
+- enberg
+- ▁agency
+- ▁bajo
+- loch
+- ▁humiliation
+- ▁dés
+- ▁lastly
+- ▁perse
+- ▁Sport
+- ▁descendants
+- ▁Ме
+- ▁statues
+- ▁lizzie
+- self
+- ▁skirts
+- ▁peered
+- ▁nascuda
+- ▁suburb
+- ▁apollo
+- eṭṭ
+- ▁manier
+- нуць
+- afi
+- ▁describes
+- ▁solicit
+- weis
+- ▁сёння
+- клі
+- ▁ink
+- ▁bubble
+- ▁በ
+- ▁izan
+- ▁gato
+- һын
+- ã
+- ▁ntiba
+- ▁export
+- ▁beaver
+- ▁alikuwa
+- ▁implement
+- ▁wad
+- wald
+- ▁gebe
+- ৈ
+- ▁Bell
+- sabye
+- луч
+- ▁egy
+- өй
+- llu
+- ▁sends
+- oid
+- wyd
+- ▁umwami
+- ▁senza
+- ▁sistem
+- стр
+- ▁manly
+- ▁scr
+- anomena
+- ▁Gri
+- ▁shun
+- ▁schuld
+- ▁homo
+- 马
+- ije
+- ▁precaution
+- ocht
+- stamm
+- ▁vase
+- ▁blut
+- ▁Mer
+- Or
+- ▁blushed
+- ▁працы
+- nischen
+- ніка
+- یر
+- ▁appreciated
+- ▁withered
+- duk
+- done
+- ▁är
+- ▁threaten
+- ▁exclusively
+- ▁هر
+- ém
+- ▁uitge
+- ▁disadvantage
+- 책
+- 通
+- ▁blau
+- ▁politeness
+- ▁gutanga
+- idde
+- шан
+- ▁여행
+- ▁imaginary
+- thon
+- wani
+- ▁Molt
+- нал
+- ▁banished
+- ▁dames
+- ▁proverb
+- гру
+- ▁draft
+- dix
+- ▁commun
+- خر
+- ▁varieties
+- ▁Ph
+- ▁emerald
+- าน
+- ▁nziza
+- ▁neatly
+- лов
+- ▁mildred
+- taking
+- ▁needless
+- லா
+- ▁wholesome
+- ▁sig
+- ▁jej
+- ▁fable
+- ▁اینجا
+- ▁chemin
+- ▁monastery
+- ▁legislation
+- ▁fein
+- ▁topic
+- shimira
+- ▁заста
+- enc
+- ▁orchestra
+- cast
+- ▁brisk
+- ▁turf
+- ▁leonard
+- fried
+- ▁poe
+- ▁заўсёды
+- fallen
+- ▁eigenen
+- ▁replace
+- prä
+- lius
+- ▁kibazo
+- poj
+- lek
+- batu
+- чны
+- лес
+- ▁inclu
+- cci
+- ▁statesman
+- ▁supposition
+- ▁succeeding
+- ▁때는
+- она
+- ▁în
+- ҳәа
+- ▁okazis
+- ▁займа
+- ▁resembled
+- ▁reared
+- glan
+- ▁maidens
+- ▁joyful
+- ciò
+- ▁instru
+- ለ
+- чо
+- ▁беларускай
+- vallen
+- ▁comforted
+- ▁neer
+- zone
+- ▁pirate
+- baum
+- ▁바로
+- ▁grans
+- ▁membres
+- tunga
+- ▁sounding
+- gha
+- aci
+- ڕێ
+- ▁tempt
+- ▁밥
+- ▁destroying
+- влак
+- ▁rewarded
+- ▁그럴
+- stände
+- bwo
+- ▁defect
+- ▁되지
+- ▁huts
+- ▁models
+- ▁clumsy
+- ▁rusty
+- indirimbo
+- ▁hü
+- ▁형
+- äß
+- อง
+- ▁anticipation
+- ▁eigene
+- ▁மு
+- gewe
+- stral
+- ▁tess
+- ▁Vull
+- ▁grie
+- ▁jungen
+- ▁Бу
+- ▁Studi
+- feh
+- лод
+- ▁carlyle
+- ▁frowned
+- ▁restoration
+- ▁pretext
+- ▁sali
+- ▁nile
+- ▁sorrowful
+- anna
+- ▁cuanto
+- eros
+- maro
+- usu
+- ▁constituted
+- ▁Will
+- ▁geest
+- ތަ
+- જ
+- ▁insignificant
+- ▁아예
+- ▁subsist
+- ▁nuevo
+- ތ
+- ▁prospects
+- አ
+- గ
+- ▁contribute
+- richten
+- ပ
+- ▁بعد
+- wend
+- ▁Both
+- aje
+- ▁proportions
+- jal
+- ▁brushed
+- subiza
+- ▁Мне
+- schule
+- ▁horrors
+- ▁henri
+- upe
+- ▁santo
+- ▁knives
+- ▁گفت
+- ▁issues
+- ▁patrol
+- ▁rad
+- فر
+- oga
+- ▁sid
+- ▁Where
+- ▁freude
+- ▁weeds
+- ▁vielen
+- ▁rogue
+- ▁associations
+- ତ
+- 토
+- ▁worthless
+- ▁sponge
+- ▁vegetation
+- чыцца
+- ▁solve
+- ▁glas
+- ▁oedd
+- lap
+- ▁сабе
+- ▁threats
+- maß
+- ▁confronted
+- লি
+- mü
+- бор
+- ▁менш
+- ▁repeatedly
+- teye
+- jf
+- ▁hap
+- ▁dazzling
+- ▁endurance
+- ▁conservative
+- ▁pudding
+- metre
+- trekken
+- át
+- ▁би
+- ട
+- ▁konstruaĵo
+- brook
+- ▁beauties
+- ▁posted
+- ▁fray
+- ▁heated
+- ▁moustache
+- ▁asylum
+- кен
+- ▁Усё
+- ▁gebracht
+- ▁trovi
+- ते
+- ĝu
+- ▁integr
+- ▁Union
+- eti
+- ▁cyrus
+- ου
+- ▁வெ
+- ▁biba
+- ivity
+- ▁babo
+- ▁somme
+- ▁Welt
+- ▁accidents
+- enburg
+- ▁teba
+- ▁posterior
+- ▁nchini
+- ڕا
+- ić
+- ▁tennis
+- 설
+- ▁canvi
+- зем
+- ufa
+- ▁murray
+- sighted
+- ▁клас
+- ▁jeff
+- sozi
+- ▁crops
+- sters
+- nius
+- jä
+- ▁ilin
+- ▁judicial
+- ▁일본
+- rump
+- kl
+- ▁yon
+- teria
+- ▁Cam
+- entre
+- mates
+- ει
+- ్
+- ▁Раз
+- ▁gaul
+- ▁license
+- ghi
+- ரை
+- 지만
+- ▁conscientious
+- ▁facile
+- yongera
+- ▁Fall
+- ▁hawa
+- dessus
+- ▁howard
+- ises
+- ▁humbly
+- ▁ĉiam
+- ▁complica
+- kundi
+- ▁enlightened
+- green
+- tracht
+- ▁westminster
+- ▁jugador
+- rije
+- ▁Aka
+- ▁exaggerated
+- ▁pulpit
+- ▁reverend
+- ▁hack
+- ▁Ky
+- चा
+- ▁turks
+- ▁ик
+- ▁camin
+- wort
+- ▁fascinating
+- ▁academy
+- ▁strait
+- ▁nell
+- ▁cye
+- ▁format
+- folk
+- ▁publicly
+- ▁kubwa
+- ▁kiam
+- ▁Beni
+- zag
+- ôn
+- ▁indulged
+- ▁pretending
+- ▁treacherous
+- Z
+- ▁rapp
+- ozi
+- ▁valet
+- wol
+- ▁Max
+- ▁이케
+- ▁choses
+- kanye
+- ını
+- ▁dispers
+- limo
+- ▁kumi
+- jong
+- ▁fantas
+- ள்
+- ▁fits
+- gano
+- ▁batteries
+- ▁xa
+- ▁ĉefa
+- ▁dingen
+- ▁secondary
+- ▁pilgrim
+- ត
+- fier
+- ▁calamity
+- ▁thrice
+- ▁slipping
+- ▁lain
+- ▁sowie
+- تان
+- ▁attentive
+- жыць
+- ▁fem
+- ▁surf
+- ġ
+- ▁elapsed
+- ▁보면
+- ▁umunsi
+- ▁costat
+- เก
+- erf
+- ements
+- ▁Form
+- 학년
+- iyor
+- ▁northwest
+- ▁scratch
+- در
+- wig
+- ▁wot
+- ▁moan
+- ▁Yella
+- ▁Roc
+- ▁eindelijk
+- ozo
+- гляд
+- ▁vices
+- ▁actors
+- ▁contemporary
+- lhe
+- byaha
+- ▁posts
+- чэ
+- ▁frantic
+- ▁irritated
+- ▁rouse
+- uva
+- mist
+- ▁prolong
+- ▁Он
+- ɗ
+- ▁landlady
+- ▁insolent
+- ▁Был
+- origen
+- нес
+- ▁brazil
+- ▁vragen
+- tako
+- ▁moll
+- ▁skilful
+- rako
+- tub
+- ▁gait
+- ▁undertook
+- ▁vigour
+- لىرى
+- macht
+- ▁assuming
+- 려고
+- ▁warren
+- ப்பு
+- associa
+- ▁epoch
+- نده
+- ▁peep
+- ifer
+- бач
+- ▁puzzle
+- ▁compa
+- ▁말고
+- ▁Cham
+- ▁nail
+- ▁불
+- ▁sales
+- ▁detailed
+- رب
+- ▁Miss
+- luck
+- ▁disappearance
+- pack
+- tü
+- ▁popularity
+- ێر
+- ەر
+- ▁Esto
+- ït
+- ▁servei
+- kunde
+- ▁espe
+- ▁respectful
+- ▁texas
+- ▁vincent
+- ▁naples
+- ▁tutor
+- ▁unos
+- storm
+- ވާ
+- lg
+- ▁diverse
+- ▁fenster
+- ckle
+- цыю
+- 五
+- σε
+- ▁termed
+- یل
+- ▁condescend
+- ▁miraculous
+- ▁parler
+- mě
+- ▁berth
+- mour
+- fact
+- зар
+- ▁Zo
+- ▁Hand
+- ▁poetic
+- ▁chestnut
+- ▁richly
+- ▁Sse
+- ▁historic
+- fru
+- ▁leer
+- irira
+- ▁discovering
+- kwam
+- raum
+- dak
+- ▁aimed
+- лены
+- ▁receives
+- rani
+- ▁indica
+- ▁contracted
+- گه
+- ▁claws
+- ▁peaks
+- ▁salut
+- ▁됐
+- ▁cultivation
+- ▁ecclesiastical
+- ▁모르
+- ▁sieht
+- tjes
+- schloss
+- ▁seulement
+- 이라
+- ▁foreigners
+- ▁ridden
+- เรา
+- ▁pom
+- rade
+- ▁mocking
+- ▁تع
+- reichen
+- liza
+- tip
+- tsch
+- ▁accuracy
+- ▁Yes
+- ಗ
+- ▁austrian
+- 表
+- ▁csar
+- wanga
+- yas
+- grown
+- ▁Tam
+- hí
+- ▁sieben
+- rood
+- ▁faltered
+- ▁dates
+- imp
+- ん
+- ▁champagne
+- ▁corona
+- ▁tartar
+- ières
+- ▁entr
+- ▁See
+- dolf
+- ▁devient
+- ாத
+- ях
+- ▁lodged
+- üg
+- ո
+- ▁언니
+- ▁glauben
+- ▁majoria
+- ▁missouri
+- état
+- ▁daddy
+- ▁linger
+- ▁dins
+- ▁generals
+- ▁truck
+- ▁scias
+- ▁bust
+- ▁sophia
+- unque
+- ▁herinner
+- St
+- ▁scriptures
+- chem
+- ▁umutima
+- ▁consol
+- ▁absolut
+- даць
+- ▁Ihre
+- ganya
+- ▁enters
+- том
+- lach
+- ▁diminished
+- ▁bilan
+- ▁vinden
+- ▁நீ
+- ▁sciences
+- tangaza
+- ▁sinister
+- ▁ogni
+- ▁그것도
+- ް
+- ▁cuerpo
+- ▁apparition
+- ▁nelson
+- ▁bobby
+- က်
+- nten
+- ką
+- wun
+- geboren
+- ▁drake
+- ▁allied
+- ▁izina
+- ▁Ses
+- ▁fidelity
+- ▁sample
+- cional
+- こ
+- ▁bestimmt
+- tata
+- ▁wrap
+- rangi
+- ▁aided
+- ▁offspring
+- ▁rhyme
+- ▁Süd
+- େ
+- ையும்
+- ▁whirled
+- buza
+- ▁генә
+- ▁Island
+- ▁Aber
+- ผู้
+- hama
+- anoj
+- ▁herd
+- นั้น
+- ▁hoo
+- ▁labours
+- ▁lebens
+- ▁phone
+- ▁compete
+- trieb
+- ▁profitable
+- ▁restrict
+- ▁singularly
+- ▁sweetly
+- ரிய
+- ван
+- ▁helena
+- ▁სა
+- ▁practise
+- ▁brake
+- ▁아마
+- ▁plen
+- dez
+- ĝoj
+- ▁shivering
+- ▁sigui
+- bwiza
+- ▁Canada
+- ▁succ
+- ▁tedious
+- laid
+- ▁있다
+- pod
+- muziki
+- ▁morals
+- ▁joining
+- ▁differently
+- ▁customers
+- ▁cement
+- ▁چند
+- ▁preliminary
+- انە
+- ▁strolled
+- ▁zose
+- кат
+- ovna
+- ▁сказаць
+- чай
+- ▁Kristo
+- ▁rustle
+- barra
+- aḍ
+- ▁primo
+- ▁روز
+- ▁courteous
+- льным
+- ою
+- teɣ
+- ▁moyen
+- obra
+- ӡ
+- ▁stellen
+- 종
+- ▁sweetest
+- ▁Uyu
+- riɣ
+- ▁barge
+- ▁concession
+- जा
+- ▁regiments
+- ▁Torre
+- ▁hag
+- 阿
+- ▁carolina
+- maak
+- hnen
+- ▁Это
+- ých
+- iez
+- ganda
+- laufen
+- ış
+- ▁cantonada
+- ▁Red
+- κα
+- œuvre
+- ਜ
+- observ
+- deɣ
+- стар
+- ▁huo
+- ▁acceptance
+- ▁могуць
+- ligen
+- ▁힘들
+- ▁вам
+- 공
+- ▁komp
+- ▁municipi
+- ▁environ
+- ▁bertram
+- بل
+- ardi
+- ▁hoorde
+- ల
+- ierung
+- ▁foundations
+- ▁languid
+- 라는
+- covered
+- ▁gentil
+- ▁sor
+- ▁каля
+- ▁юҡ
+- بار
+- মি
+- ▁privately
+- ▁backed
+- gezogen
+- ▁애들이
+- udu
+- ▁Menschen
+- respect
+- ▁gum
+- ▁loom
+- ▁heath
+- date
+- ▁ensued
+- dura
+- lose
+- ▁swayed
+- ▁quinze
+- ▁يې
+- ▁tended
+- әк
+- ▁prejudices
+- ▁pros
+- huit
+- ▁lull
+- ่อ
+- tria
+- ▁unreasonable
+- ▁removing
+- ▁equipped
+- ▁framed
+- ▁vanish
+- nish
+- ▁crawl
+- ▁rustic
+- ▁behavior
+- ▁lloyd
+- ▁attire
+- ▁settling
+- pita
+- girl
+- ▁sotto
+- 喺
+- ▁fragrant
+- ▁Ebi
+- ktion
+- arc
+- ▁모르겠어
+- ▁bicycle
+- ▁coachman
+- ▁Shi
+- ▁адно
+- ▁grub
+- nico
+- found
+- thy
+- ▁sturdy
+- ▁Rw
+- ▁такое
+- dec
+- ▁abans
+- نت
+- walt
+- ització
+- ▁Frau
+- ▁gedanken
+- லி
+- ▁roc
+- klar
+- ▁comic
+- ская
+- ▁agricultural
+- esque
+- ▁lesser
+- lè
+- ▁zy
+- ▁corporation
+- ▁hannah
+- ▁Tett
+- 长
+- ▁commencement
+- ▁klas
+- ▁yelled
+- ▁exertion
+- ssé
+- μα
+- час
+- ▁historio
+- ▁viewed
+- ▁ideals
+- ▁England
+- ય
+- ▁devout
+- ▁rut
+- ▁brim
+- ▁kio
+- evo
+- ▁instruis
+- ▁نو
+- ▁laboratory
+- burgo
+- ▁primero
+- ▁بیا
+- ▁gelaat
+- ▁Niba
+- ppo
+- ▁blackness
+- ▁extinguished
+- ▁liar
+- ▁цяжка
+- gion
+- ▁liefde
+- ▁milli
+- gambi
+- ▁links
+- ▁crater
+- ▁pennsylvania
+- viel
+- ▁porthos
+- ▁augustine
+- bira
+- ▁routine
+- чык
+- ▁haul
+- luk
+- paper
+- ▁gelukkig
+- ಸ
+- альна
+- ▁әле
+- ▁pai
+- ▁Unu
+- еү
+- 对
+- ▁cecil
+- ▁sweetheart
+- ▁peering
+- sema
+- ▁Scha
+- asent
+- ▁bandage
+- kunze
+- اج
+- tay
+- koloni
+- kers
+- ▁burglar
+- ▁tengo
+- ень
+- nato
+- ▁ehe
+- ▁repos
+- ▁bertha
+- etan
+- ▁verkis
+- ▁okuba
+- ▁clenched
+- öhn
+- fusion
+- ivi
+- ▁Ei
+- ssu
+- ▁ئىش
+- кае
+- ▁sagt
+- ▁simultaneously
+- ▁perch
+- များ
+- ன்ன
+- อยู่
+- ▁dla
+- ▁쪽
+- ▁orderly
+- ▁spade
+- ▁ancora
+- ▁jesuit
+- важ
+- ▁raging
+- ▁carrier
+- ▁drifting
+- ສ
+- ▁flocks
+- ▁boon
+- ▁contribu
+- ность
+- ▁turi
+- ixen
+- маг
+- gire
+- ▁renounce
+- kwata
+- ▁Dans
+- ▁displeasure
+- ▁maids
+- riga
+- ▁vertical
+- rongo
+- ▁superficial
+- ▁fling
+- ▁detachment
+- ▁Fri
+- ▁deren
+- abanya
+- fica
+- melo
+- ▁nearby
+- ▁생각을
+- tops
+- ▁lingvoj
+- ▁fascination
+- رف
+- 하잖아
+- bürger
+- ▁industrious
+- дат
+- پتۇ
+- ▁accounted
+- ▁imitate
+- ▁distracted
+- froid
+- magnet
+- ▁gye
+- ▁Frage
+- ▁gotten
+- ▁comedy
+- bano
+- ▁hailed
+- ▁register
+- ▁saluted
+- эння
+- ▁exit
+- 따
+- cn
+- /
+- ▁actress
+- okra
+- һеҙ
+- kunga
+- ▁кошт
+- '16'
+- été
+- سان
+- gwi
+- ▁weniger
+- ▁gebruik
+- über
+- ▁Spi
+- ▁якая
+- ▁mains
+- ▁recib
+- ▁penn
+- 난
+- aplica
+- ▁dein
+- ▁hinted
+- '30'
+- ▁camps
+- miz
+- cava
+- goma
+- ား
+- ujo
+- ງ
+- ▁troublesome
+- ť
+- ▁był
+- ▁열
+- media
+- 만원
+- ▁precipitate
+- ▁Sel
+- ተ
+- ▁accommodation
+- ▁daisy
+- ▁statute
+- ▁cod
+- ▁fugitive
+- lige
+- ▁wha
+- ▁dolly
+- dore
+- ▁Bis
+- саҡ
+- ellement
+- ▁gratified
+- ▁geheim
+- embla
+- ▁stair
+- ▁pies
+- цкі
+- kire
+- ▁Recht
+- ▁breeding
+- ▁començar
+- ▁gladness
+- ink
+- ▁ddim
+- ▁patches
+- ▁shameful
+- teddu
+- ▁Cro
+- rufen
+- corre
+- ▁identified
+- croft
+- üß
+- hagaze
+- 번
+- 金
+- ää
+- ▁precipe
+- ▁bam
+- ▁hija
+- ▁weaker
+- کرد
+- ▁booth
+- ▁unbroken
+- 為
+- ▁clump
+- same
+- arle
+- ▁shiver
+- ndash
+- аб
+- ▁Frank
+- llah
+- about
+- ▁Mr
+- ▁Char
+- kere
+- ▁chaplain
+- ▁enchanted
+- ▁nodding
+- timo
+- служ
+- ▁nachdem
+- ▁companionship
+- igeze
+- weit
+- дам
+- ▁detached
+- ▁grammar
+- sized
+- ▁uzata
+- شان
+- री
+- tende
+- ▁considera
+- ▁fritz
+- ▁corri
+- ▁który
+- 출
+- ganira
+- dank
+- ต้อง
+- 女
+- ▁dripping
+- fast
+- ▁kuti
+- 갈
+- Һ
+- ▁richtig
+- ▁securing
+- ▁relish
+- ▁curls
+- ▁ĉar
+- hili
+- ด้วย
+- ▁claude
+- ▁lasting
+- uzo
+- ▁رفت
+- ач
+- หล
+- ніч
+- ▁muddy
+- ▁Gil
+- ▁educa
+- ▁controversy
+- zet
+- ▁profits
+- goro
+- 순
+- ká
+- تۇ
+- schieden
+- ranger
+- ▁trouver
+- ுவ
+- ір
+- sign
+- ▁estaban
+- 知
+- ▁crouched
+- 글
+- مس
+- ▁julian
+- krib
+- ▁streaming
+- ▁athens
+- zzle
+- ▁tylko
+- ▁instruct
+- ▁Richard
+- ▁Problem
+- umuntu
+- ▁ransom
+- ▁오빠
+- ▁males
+- ▁writ
+- ▁stitch
+- πο
+- gezi
+- ती
+- ▁cowardly
+- foli
+- ▁lyon
+- ▁syria
+- ▁verle
+- ▁discouraged
+- ▁nommé
+- әт
+- feli
+- ▁arro
+- ▁yamu
+- ▁metro
+- fant
+- اء
+- ▁publi
+- orde
+- ގައި
+- ▁przed
+- ენ
+- ▁அவர்
+- trap
+- 임
+- ▁feind
+- umvise
+- ▁whipped
+- ▁exert
+- ▁hoofs
+- ▁생각해
+- ▁dungeon
+- ▁என்ன
+- праў
+- ▁aqui
+- ▁faris
+- ĉi
+- 줬
+- ▁그러면은
+- laba
+- nça
+- ▁alexandr
+- cend
+- ▁tortured
+- ▁coneix
+- ▁түгел
+- ▁언제
+- ▁бала
+- ähr
+- ▁familio
+- ▁memorable
+- ▁kenn
+- hagi
+- ▁toilet
+- ▁니가
+- ▁resource
+- rane
+- ▁molto
+- ▁duration
+- ▁туды
+- पा
+- ▁manor
+- ตัว
+- ▁construct
+- craft
+- ▁clamor
+- ▁commend
+- dah
+- ensemble
+- ▁palabras
+- кру
+- ▁comin
+- 無
+- rain
+- ▁deliverance
+- Н
+- 새
+- makers
+- ▁banc
+- borg
+- ▁perdi
+- ▁entertaining
+- ▁surrendered
+- ▁urging
+- iĝi
+- ▁তার
+- ▁arises
+- ▁athenian
+- ▁moisture
+- ▁noodig
+- geno
+- னர்
+- ▁amar
+- ▁chickens
+- verein
+- ▁footing
+- lı
+- ▁Roma
+- adores
+- rila
+- ▁persuasion
+- ▁stag
+- ▁eri
+- ▁howling
+- ▁prit
+- ▁Tar
+- cata
+- ▁ascertained
+- کۆ
+- ރަ
+- чка
+- eared
+- isant
+- ▁uttering
+- ▁همه
+- ▁invest
+- ಮ
+- асць
+- rill
+- ▁gros
+- hwe
+- tanze
+- ▁spontaneous
+- ▁consultation
+- teeken
+- ▁Kra
+- ctic
+- 엔
+- ancien
+- ▁derive
+- imento
+- лад
+- ø
+- ▁jaren
+- ٍ
+- ▁electricity
+- ▁biz
+- ▁husbands
+- ▁Hill
+- ▁sorte
+- ▁rive
+- ▁undergo
+- ല
+- ▁sparrow
+- nage
+- ▁걔네
+- хи
+- kirche
+- ▁boundary
+- ӧр
+- ▁snare
+- ubuzima
+- ▁oppression
+- ▁camels
+- ▁паказ
+- wain
+- ▁Bat
+- ▁뭐지
+- ▁birch
+- شى
+- نَ
+- ▁ubwoba
+- ▁historian
+- ▁pati
+- änder
+- ਮ
+- ▁façana
+- ่น
+- ▁nuestro
+- bow
+- 身
+- ▁집에
+- பி
+- ▁magician
+- ▁되고
+- ▁Universitat
+- ▁pillar
+- iah
+- 先
+- urus
+- ▁roy
+- ała
+- ▁reel
+- ▁glaube
+- 하기
+- ▁bury
+- ▁chairman
+- ▁heroine
+- ือ
+- ▁anew
+- ▁rook
+- leben
+- ▁prodigious
+- ▁rho
+- ▁canoes
+- ▁зноў
+- politik
+- wana
+- ▁Eb
+- сен
+- subira
+- rauch
+- rę
+- ▁outlaw
+- ▁allerdings
+- فا
+- ▁forcing
+- ▁prick
+- ლი
+- ▁남자
+- eilt
+- ▁mechanic
+- ▁desolation
+- ▁imposing
+- ▁Mur
+- ▁regi
+- ▁canton
+- ▁yahweh
+- ▁mazi
+- home
+- ▁cours
+- یەک
+- ▁describing
+- ыла
+- mog
+- 灣
+- vine
+- ▁define
+- ▁nicely
+- ▁poisoned
+- idée
+- aï
+- ދަ
+- ▁এই
+- phan
+- schaften
+- ▁seeks
+- ▁females
+- ▁merrily
+- вяз
+- activitat
+- ▁poz
+- ▁menshi
+- ▁gelang
+- kling
+- ▁accepting
+- ▁thrilling
+- ▁ihres
+- ▁execute
+- ▁ebenso
+- there
+- تو
+- ▁Und
+- ▁movie
+- ▁degli
+- agala
+- ▁clasp
+- ▁dernière
+- ▁heavier
+- aliment
+- ▁listener
+- tagon
+- ▁inevitably
+- ▁correspondent
+- мент
+- ▁gleamed
+- ▁reluctantly
+- ▁баш
+- ▁retra
+- ▁pronto
+- poko
+- ugi
+- ▁адчу
+- kutu
+- ▁afore
+- ▁hurl
+- ▁esteemed
+- кія
+- ія
+- ▁perchance
+- ▁decade
+- ▁regain
+- pera
+- ▁complaints
+- ▁morton
+- turm
+- ▁agafar
+- ▁Tt
+- ▁celebra
+- ▁herein
+- ▁Gas
+- schließen
+- ▁sums
+- ▁collecting
+- dier
+- wert
+- ▁clement
+- ▁lucien
+- ▁Ana
+- ▁bustle
+- estat
+- xon
+- fau
+- ▁hände
+- ▁Ober
+- ው
+- ▁resembling
+- ผ
+- ▁dispense
+- ▁organic
+- ▁volgen
+- ▁resisted
+- 原
+- ▁stationed
+- ▁venne
+- ▁believes
+- 했던
+- ▁aristocratic
+- ▁région
+- ▁intending
+- ટ
+- 삼
+- ▁mike
+- ▁paste
+- نم
+- ▁California
+- shoboye
+- விட
+- ▁milit
+- ▁pedro
+- ▁ecstasy
+- ▁wears
+- cine
+- ▁oben
+- ▁Bla
+- ▁revolutionary
+- '!..'
+- pher
+- liwa
+- ▁тө
+- ▁australia
+- ▁usurp
+- èn
+- ▁salle
+- ▁sorely
+- rette
+- тру
+- ▁requisite
+- ▁sneak
+- rta
+- ▁compensation
+- ▁worshipped
+- vura
+- ▁semble
+- ▁lois
+- ▁despairing
+- ▁surround
+- ▁alcohol
+- ▁Did
+- ▁awaken
+- gesprochen
+- ▁perplexity
+- ▁colli
+- ▁shu
+- ▁abyss
+- ▁jehová
+- ▁portuguese
+- 란
+- wachsen
+- هم
+- ▁notorious
+- lê
+- ▁roche
+- lulu
+- hwa
+- нды
+- ▁cooper
+- ▁howl
+- ▁aramis
+- ▁associate
+- جا
+- ▁kentucky
+- ▁settlements
+- efk
+- ▁headache
+- ▁último
+- ▁egyptians
+- ▁prim
+- ▁luggage
+- 比
+- ▁palms
+- impu
+- ద
+- ▁epistle
+- interes
+- ▁ominous
+- ырға
+- ▁nav
+- ▁discreet
+- 基
+- ▁shivered
+- ▁muffled
+- ▁morto
+- τι
+- ним
+- ▁ciel
+- ▁detained
+- сцю
+- ▁fisherman
+- ٽ
+- ▁demons
+- 주는
+- staat
+- ▁particles
+- ▁dedicated
+- ▁quedar
+- авана
+- ▁serie
+- ်
+- graduate
+- ▁musing
+- ▁зрабіць
+- дуць
+- ▁starving
+- ▁publisher
+- nneɣ
+- ▁grocer
+- ▁жыццё
+- ▁thereafter
+- 起
+- ▁disgusted
+- ▁trading
+- здо
+- шты
+- ▁двух
+- ▁ferdinand
+- zda
+- ▁exceed
+- 친
+- ▁circulation
+- ▁kunt
+- ência
+- ▁아니고
+- อะไร
+- ▁한번
+- treffen
+- afu
+- gefallen
+- ▁Super
+- ▁involuntarily
+- ▁thirsty
+- вать
+- ará
+- ▁talks
+- ән
+- situ
+- ▁Durch
+- ▁hardened
+- bí
+- ▁cork
+- spread
+- ▁canst
+- emente
+- ▁British
+- ▁unequal
+- ▁flourish
+- erta
+- ▁эр
+- ▁sujet
+- ▁봤어
+- ымі
+- ▁fama
+- schal
+- ▁ahantu
+- ▁joue
+- ▁hals
+- cere
+- ▁prompted
+- uï
+- ޓ
+- kio
+- ▁situa
+- ▁muchas
+- ▁memorial
+- ▁Albert
+- ▁balina
+- ▁costs
+- ▁direkt
+- ▁empress
+- меш
+- xton
+- ▁straat
+- ▁depressed
+- лат
+- уд
+- ков
+- ▁racing
+- ▁cherry
+- ▁corazón
+- ▁crag
+- Pa
+- った
+- gte
+- ▁erschien
+- плы
+- ▁labors
+- ▁bedside
+- mio
+- ວ
+- ▁wiggily
+- ▁siete
+- ▁meek
+- ▁vigorously
+- erton
+- ▁sleeves
+- ئے
+- ▁икән
+- ▁jacques
+- ijuru
+- radi
+- ▁төш
+- ▁Mad
+- ▁China
+- ▁باز
+- ▁burke
+- ▁horizontal
+- пыт
+- ▁briskly
+- rial
+- abri
+- чнай
+- ້
+- ▁primarily
+- ▁chuckled
+- ▁curved
+- сьці
+- ার
+- '15'
+- baw
+- tätig
+- ▁genus
+- ▁sinners
+- ▁choir
+- asu
+- ့
+- į
+- ▁trabajo
+- ▁batter
+- ▁luminous
+- ▁darker
+- ▁moore
+- ▁giants
+- ▁motions
+- ▁hateful
+- ō
+- ħ
+- mers
+- نش
+- koro
+- ▁hears
+- rush
+- ▁sekva
+- visi
+- ▁glove
+- ▁voluntary
+- ▁losses
+- وار
+- ▁Von
+- ▁mechanically
+- ève
+- ▁prova
+- ▁Ken
+- nika
+- ▁Ĉe
+- ▁energies
+- ▁veni
+- ▁technique
+- ▁fellowship
+- ▁escrit
+- ▁miß
+- faransa
+- ▁edmund
+- 겠
+- ώ
+- half
+- ▁riddle
+- avez
+- ▁나왔
+- mäßig
+- роз
+- ▁ziet
+- 렸
+- péri
+- ▁blunt
+- ▁titles
+- ▁williams
+- usse
+- ▁recovering
+- ▁Ntu
+- ▁само
+- է
+- ▁rolls
+- шын
+- ▁pail
+- ▁founder
+- ▁ripple
+- halen
+- ▁marcus
+- ▁rwanda
+- ▁excepting
+- ▁civilisation
+- geschlossen
+- angiza
+- ▁була
+- quill
+- dealer
+- rigo
+- ▁schwarz
+- ▁dickens
+- ▁erhielt
+- ligt
+- tega
+- ▁snowy
+- ▁그서
+- ▁pronounce
+- tré
+- ▁entgegen
+- нія
+- maz
+- ▁salad
+- ▁já
+- лей
+- ▁moth
+- gaba
+- ▁feuer
+- ▁அந்த
+- خە
+- зм
+- ваўся
+- ?!
+- ▁picnic
+- esc
+- world
+- ▁traduk
+- ▁constantinople
+- ▁així
+- ▁archie
+- нава
+- ▁clown
+- маш
+- ▁ooit
+- gesteld
+- ▁Яго
+- поль
+- ▁charmed
+- йшла
+- ▁Madrid
+- чныя
+- ▁peru
+- ▁opponent
+- wahr
+- ▁locality
+- ▁steed
+- ▁sobs
+- мый
+- ▁maso
+- ами
+- بە
+- ▁owen
+- ▁हो
+- рға
+- gena
+- ▁superstitious
+- ▁altri
+- ▁Fla
+- விட்ட
+- ▁Pour
+- ▁infidel
+- ▁kilo
+- ▁felix
+- ▁respondió
+- ▁flattered
+- loca
+- ▁mining
+- ацца
+- ▁stored
+- estructura
+- ▁dazed
+- erei
+- oze
+- ▁nieuwe
+- ▁aaron
+- யாக
+- ▁anecdote
+- ▁diary
+- ▁가면
+- ▁Let
+- ▁humorous
+- ▁syllable
+- miento
+- pfu
+- gown
+- ▁shovel
+- tamente
+- lob
+- зь
+- ▁teaches
+- މާ
+- ▁tart
+- ▁administered
+- ▁ignored
+- ▁balcony
+- ош
+- ▁rubbish
+- ▁State
+- ▁finishing
+- тем
+- ▁dissolved
+- ẓẓ
+- ▁criminals
+- ▁nza
+- 布
+- ▁cambridge
+- koni
+- ▁Kor
+- gib
+- част
+- ión
+- ▁contemplated
+- ▁hubiera
+- ▁Tes
+- ▁swelling
+- ▁tons
+- ▁út
+- ▁Quer
+- ▁duncan
+- ▁strap
+- ifuza
+- ▁headlong
+- ▁seele
+- ▁possessing
+- aille
+- ▁dormi
+- éra
+- вуч
+- 常
+- ▁injure
+- ▁neder
+- ▁orthodox
+- ▁chemist
+- ▁thickly
+- ▁worms
+- ▁delante
+- ▁parrot
+- ▁emotional
+- ð
+- ▁expectations
+- ▁beans
+- litz
+- іх
+- ▁esqui
+- ▁Lake
+- ▁trent
+- ▁ties
+- hak
+- ყ
+- һе
+- ▁tips
+- ▁Pat
+- inshuti
+- ▁futur
+- ▁свой
+- 애
+- eerd
+- anze
+- ▁declara
+- '00'
+- ▁resemble
+- ▁داشت
+- әлә
+- ▁Jan
+- валіся
+- ▁bridegroom
+- ▁많아
+- ▁bowing
+- ▁vuba
+- ▁شا
+- Ho
+- iano
+- ދު
+- ▁principally
+- rsi
+- ▁curate
+- ▁plentiful
+- ▁clemens
+- ▁Wasser
+- ваю
+- ride
+- ▁fawn
+- ▁iminsi
+- ▁imirimo
+- óż
+- দি
+- nach
+- ▁embroidered
+- wende
+- ▁inquiring
+- ▁canyon
+- ▁effectually
+- ▁holt
+- gung
+- ▁glowed
+- ▁secrecy
+- ▁Thomas
+- ▁nueva
+- жэ
+- wał
+- ▁lac
+- servant
+- εί
+- shyize
+- ▁blamed
+- ▁Jam
+- ▁jego
+- بێ
+- annon
+- ▁physicians
+- ▁barbarous
+- ▁shifted
+- ахь
+- ▁vienna
+- ▁accidental
+- ▁explica
+- ▁participate
+- ptic
+- 该
+- ▁mortis
+- ▁overcoat
+- ▁miser
+- ĥ
+- ▁starve
+- ▁anzu
+- bih
+- lida
+- ▁expand
+- ographic
+- grid
+- ▁eliminat
+- бер
+- ezza
+- ▁scanty
+- jk
+- ▁caballo
+- யும்
+- rega
+- ▁parola
+- bry
+- ▁procedure
+- ▁trug
+- kaa
+- ▁ئۆ
+- ▁Jag
+- ▁persistent
+- ▁Arab
+- ▁representa
+- ▁кожны
+- ▁Kirche
+- ▁arising
+- plaats
+- ▁alarming
+- ▁solchen
+- ront
+- ▁struggles
+- ▁steadfast
+- ▁martyr
+- пал
+- hoff
+- ▁Dw
+- ▁Of
+- ▁parker
+- дө
+- edged
+- kiem
+- ρι
+- tana
+- ▁suoi
+- ▁peal
+- িত
+- bore
+- ▁baronet
+- ธ
+- ەن
+- ับ
+- bilitat
+- quent
+- ▁Posteriorment
+- ▁conversa
+- ▁joven
+- ▁reine
+- ▁fing
+- ▁운동
+- 格
+- baga
+- ▁continental
+- vyo
+- เท
+- ▁goats
+- org
+- ▁catholics
+- ▁declares
+- ▁quint
+- ▁devices
+- ▁camí
+- ▁Berg
+- big
+- ▁signify
+- ▁composure
+- parent
+- ▁stew
+- ▁insight
+- ▁Obu
+- ▁क
+- Y
+- ям
+- æ
+- ▁yacu
+- imodoka
+- ▁Ĝia
+- στ
+- ▁dus
+- simu
+- ▁arisen
+- بین
+- ير
+- пат
+- گى
+- ▁dismiss
+- чыта
+- ▁marches
+- ▁beak
+- ▁آنها
+- dige
+- ▁alleged
+- ष
+- ஜ
+- ▁internacional
+- ▁peach
+- ▁wedi
+- rine
+- свят
+- ▁competent
+- کاری
+- ▁Internet
+- ▁moltes
+- cushion
+- اند
+- ▁compose
+- ▁Verein
+- waardig
+- enia
+- ▁profoundly
+- ▁paso
+- ▁spotted
+- ▁needful
+- ▁gehe
+- نس
+- ▁ruddy
+- adors
+- ▁pilgrimage
+- ▁elektr
+- ▁soften
+- ▁loveliness
+- ckel
+- ▁hooren
+- ▁crane
+- である
+- ▁mówi
+- ▁trusting
+- ▁übrigen
+- க்கும்
+- builder
+- profund
+- mpaka
+- ▁kneel
+- ▁gewiß
+- ைக்
+- ▁hai
+- bild
+- ила
+- ▁Their
+- buka
+- éis
+- ▁autour
+- ▁censure
+- 플
+- پا
+- ▁Dio
+- wanda
+- ▁haya
+- һи
+- ▁overlooked
+- ▁novelty
+- iad
+- ▁honored
+- ▁kugenda
+- vl
+- ▁Plu
+- ▁Sil
+- شر
+- wechsel
+- ▁français
+- ▁chime
+- ▁doubled
+- ▁polar
+- ▁wanderer
+- ել
+- ▁inviting
+- undo
+- ▁Kal
+- ▁messengers
+- ▁funciona
+- ▁bato
+- ▁handling
+- ▁darin
+- бай
+- ތު
+- ▁kila
+- ▁propri
+- ப்பட
+- ▁watches
+- ▁mash
+- ▁petites
+- urira
+- ▁feudal
+- esten
+- ▁magnitude
+- ▁turtle
+- кс
+- ▁enforce
+- ▁devait
+- гер
+- વા
+- ▁reuni
+- تۇر
+- àtic
+- ruta
+- ▁emptied
+- ▁insensible
+- mino
+- யை
+- ▁Jahren
+- ▁launched
+- lex
+- ▁membro
+- xx
+- ত্র
+- ▁праца
+- бал
+- ▁humming
+- fres
+- ydd
+- 花
+- ▁scowl
+- yam
+- ▁teams
+- ▁manière
+- ▁ginger
+- nyije
+- 可
+- ▁hello
+- issement
+- ▁uproar
+- ▁ocup
+- ▁bibli
+- yó
+- tempo
+- 建
+- kili
+- ▁lacked
+- mack
+- gelegenheit
+- ▁console
+- tora
+- gestalt
+- стая
+- ▁Pau
+- 써
+- ▁spacious
+- ood
+- ▁youngster
+- ▁involve
+- bbling
+- ▁clutched
+- yak
+- 或
+- ▁patriotism
+- іўся
+- ți
+- ▁гэтых
+- ▁sanctuary
+- ▁premises
+- يد
+- affe
+- ҙең
+- 期
+- ▁tremulous
+- 业
+- ▁fuerza
+- sson
+- ▁рух
+- ▁attic
+- ▁сум
+- fana
+- ▁robust
+- ▁larry
+- ▁Entre
+- ▁livre
+- ▁nursing
+- lauf
+- ▁podia
+- йә
+- ▁refinement
+- ▁mfite
+- ▁overboard
+- зні
+- ĉo
+- ▁mouvement
+- ▁dispara
+- loos
+- angi
+- hong
+- اری
+- ▁bait
+- isson
+- ▁dump
+- ▁portraits
+- ▁Ар
+- ▁Igi
+- 민
+- រ
+- ▁hören
+- ▁noticing
+- ▁okazas
+- ▁appoint
+- ▁situat
+- ▁алды
+- ▁interrupt
+- fir
+- ▁discomfort
+- ▁horribly
+- axe
+- empi
+- landa
+- cuba
+- ▁disliked
+- ভা
+- ▁nero
+- ▁shrieked
+- ▁ingenuity
+- േ
+- ▁spun
+- aide
+- ▁sollen
+- ▁vegeta
+- ▁fals
+- ¿
+- ▁gertrude
+- ▁potential
+- mutwe
+- ▁binding
+- teller
+- clad
+- ▁favoured
+- għ
+- ▁scales
+- ▁onto
+- ▁laborious
+- ▁scope
+- ▁zealous
+- уе
+- ▁calf
+- ▁basti
+- ▁suppress
+- ွ
+- ▁hazel
+- ▁promo
+- ▁muti
+- ▁magistrates
+- ▁vijf
+- ▁cheered
+- нняў
+- ▁antiquity
+- ▁ungrateful
+- ▁думаю
+- ▁сваю
+- atumye
+- ити
+- ▁trobar
+- kiba
+- umunya
+- ▁operator
+- ▁harimo
+- ড
+- kama
+- 저
+- برا
+- ▁confront
+- ▁bully
+- shell
+- ▁fascinated
+- ▁rapide
+- ▁ascent
+- ▁conjunt
+- ▁Dort
+- ▁Hä
+- ordnung
+- zana
+- ▁разам
+- simila
+- bib
+- ajja
+- ▁favored
+- ▁memor
+- ▁seria
+- ▁jin
+- ▁Stu
+- ▁sow
+- ▁fühlte
+- ▁churchyard
+- تَ
+- ▁vary
+- rium
+- ▁noblest
+- 军
+- ▁처음
+- ▁contend
+- ▁weekly
+- ▁odour
+- lieb
+- schutz
+- arma
+- 化
+- ▁grandson
+- ▁Kenya
+- ▁voulu
+- ▁sortir
+- ▁Ти
+- ▁feller
+- уюць
+- ▁consternation
+- 兰
+- ▁umwanya
+- 봤
+- ێن
+- ▁prezent
+- hampton
+- nec
+- ejoj
+- miti
+- ент
+- agrada
+- ierten
+- ▁dispatch
+- ▁repay
+- тың
+- ▁marvelous
+- شه
+- ▁nasty
+- ▁auszu
+- ▁ocho
+- জন
+- ▁ports
+- uno
+- ▁damals
+- ▁disturbing
+- ▁nave
+- 罗
+- ▁earning
+- ▁naho
+- ź
+- قد
+- staff
+- ▁Mitglied
+- ▁tym
+- ▁fails
+- ▁hillside
+- ▁immensely
+- ▁antony
+- smith
+- knecht
+- Ä
+- hra
+- holders
+- stä
+- Mo
+- kirwa
+- schrieb
+- ▁مد
+- ▁Michael
+- ▁différent
+- ▁unfit
+- gesch
+- ▁sorg
+- ▁blij
+- ▁nyingi
+- ▁sag
+- ▁دەر
+- ▁فکر
+- ▁sidney
+- ▁blushing
+- ely
+- ▁sly
+- тон
+- کس
+- вага
+- ▁buzz
+- ▁Ende
+- ▁daresay
+- alta
+- Ж
+- 군
+- wch
+- ▁navigation
+- тая
+- ▁Real
+- ▁healing
+- senga
+- куп
+- ▁Би
+- ▁warlike
+- ▁implied
+- მი
+- рҙы
+- ▁pesca
+- ▁woodland
+- dà
+- нын
+- uganda
+- ▁italien
+- ▁swaying
+- زم
+- ▁brink
+- 다는
+- ▁compi
+- يو
+- ▁praises
+- rirwa
+- ▁welches
+- sloten
+- ▁série
+- բ
+- lom
+- toire
+- cret
+- ▁pau
+- ▁correctly
+- ▁pitcher
+- する
+- tuig
+- crib
+- ▁sentit
+- ▁dispersed
+- дук
+- ▁votes
+- ▁bridges
+- ▁elector
+- eiro
+- য
+- чен
+- ەکانی
+- taal
+- ▁combine
+- ▁ўсіх
+- ▁mien
+- ▁annex
+- mort
+- ▁wearily
+- telefon
+- ▁composer
+- brü
+- ▁cruelly
+- čči
+- ▁Pala
+- dire
+- mg
+- ▁dije
+- дзей
+- ▁propaga
+- ▁teritorio
+- ▁Green
+- kant
+- ▁gentleness
+- richtung
+- buga
+- zuri
+- ▁attacking
+- fera
+- ▁valued
+- ▁ebenfalls
+- ▁performing
+- ترین
+- ▁fixing
+- zono
+- pí
+- ▁Luc
+- oci
+- ▁бок
+- န
+- Re
+- ▁ell
+- ▁inland
+- ▁adjacent
+- ▁malicious
+- ▁studi
+- page
+- زە
+- ▁salv
+- ▁richest
+- اط
+- ▁abate
+- ▁certe
+- ▁barri
+- ▁fluttering
+- 開
+- nacht
+- abagore
+- ▁dough
+- räum
+- ▁beads
+- ▁powerless
+- ▁dashing
+- ின
+- reke
+- ▁сям
+- ▁electro
+- ▁hoy
+- bind
+- ▁disputa
+- ▁brin
+- ▁volcano
+- sprache
+- ▁providing
+- woord
+- ▁mounting
+- ▁nothin
+- ▁massachusetts
+- mpo
+- ▁whirling
+- ទ
+- ▁preskaŭ
+- ▁daybreak
+- ▁twinkling
+- ▁evelyn
+- nahm
+- ті
+- banza
+- bola
+- ခ
+- ▁politique
+- rano
+- ▁перад
+- ▁illustrate
+- мал
+- 큰
+- 표
+- ▁cabbage
+- ▁beatrice
+- ▁glaring
+- basha
+- няў
+- دي
+- ▁чалавека
+- ▁moviment
+- ٌ
+- کات
+- ▁gewo
+- ▁dès
+- ▁nzu
+- ▁vibra
+- ▁mentally
+- ▁Cat
+- version
+- こと
+- ▁Ker
+- ▁вядома
+- 然
+- ▁sufferer
+- ałem
+- ொ
+- ▁warmed
+- ▁zoon
+- cham
+- tate
+- ▁Rock
+- ▁cheat
+- ▁glided
+- ▁futile
+- ffle
+- bic
+- ▁ashobora
+- empresa
+- rata
+- ▁cultural
+- ▁veux
+- ▁kuma
+- habit
+- ▁swollen
+- ▁blinded
+- ▁legte
+- ▁située
+- وق
+- ▁сталі
+- ▁Person
+- mado
+- ▁zuster
+- ▁Aha
+- ▁història
+- ზ
+- zor
+- ▁unsuccessful
+- ▁mientras
+- ▁gratification
+- ▁такая
+- shye
+- эм
+- స
+- books
+- schijn
+- ▁tossing
+- ▁leisurely
+- ▁transition
+- egna
+- 했었
+- ланы
+- vro
+- sichtig
+- ▁orbit
+- ▁gern
+- ква
+- dite
+- ▁achieve
+- ▁indignantly
+- ▁pistols
+- ▁wowe
+- 後
+- کان
+- льна
+- dagi
+- ▁Vous
+- middel
+- ▁часам
+- ▁entreat
+- 지는
+- ▁இருந்த
+- ▁turk
+- ▁hernam
+- ▁diet
+- 自己
+- ▁sûr
+- ▁forts
+- ▁여자
+- ▁dorm
+- ▁langue
+- thal
+- mbye
+- ▁emerge
+- ▁Rei
+- bikora
+- ▁nurses
+- ங்கு
+- ▁promotion
+- ▁darrer
+- нап
+- ▁manufacturer
+- ▁ribs
+- ▁surpass
+- ▁deputy
+- ▁kuna
+- ▁Laŭ
+- ▁territori
+- ▁новы
+- ▁commented
+- ▁joshua
+- 口
+- ▁deliberation
+- рма
+- ▁coins
+- ukura
+- ▁morbid
+- ▁reflecting
+- ▁dochter
+- Ubu
+- 잡
+- 左
+- ▁pensée
+- ▁imagin
+- ubw
+- ▁pell
+- nki
+- ▁retro
+- besitz
+- ▁muscle
+- ▁Met
+- ▁Mittel
+- abakozi
+- ைப்
+- ▁guides
+- nę
+- ▁لێ
+- ▁Now
+- كان
+- ▁spinning
+- ▁Pli
+- '11'
+- ▁bathed
+- ▁tackle
+- ▁guise
+- ▁geste
+- ▁crusade
+- ▁faz
+- ▁audible
+- خۆ
+- ▁burrow
+- ▁qualified
+- ▁drown
+- ▁Ul
+- ▁whisky
+- ▁imply
+- ▁ligne
+- حا
+- ▁beethoven
+- ▁thompson
+- ▁sentir
+- ▁Gegen
+- bba
+- ▁grandi
+- 今
+- ▁момант
+- umbu
+- جه
+- ▁illustrated
+- ▁tails
+- ർ
+- ▁कर
+- яз
+- ▁Band
+- ހި
+- ghe
+- ало
+- ▁drooping
+- zucht
+- umwe
+- ▁crook
+- puso
+- জা
+- ▁surveyed
+- ▁grasping
+- turi
+- ▁slack
+- lista
+- ▁lachen
+- gaben
+- ▁commissioners
+- 很
+- ▁wearied
+- feri
+- ▁Mel
+- 반
+- ▁veiled
+- сад
+- ▁murmuring
+- ணி
+- 工
+- ▁liverpool
+- 等
+- ▁Icyo
+- ▁Aquests
+- iddwa
+- mous
+- ▁samo
+- brac
+- hold
+- ▁comparative
+- brach
+- ▁hallo
+- ▁бара
+- ▁بن
+- ▁similarly
+- taki
+- ▁transcend
+- ▁tangled
+- kunft
+- ▁tidet
+- twenty
+- ▁status
+- ҡан
+- ▁peck
+- ▁incessant
+- ▁davy
+- ▁galloped
+- ▁stature
+- street
+- ▁eclipse
+- ▁titre
+- ša
+- udde
+- ▁yal
+- ▁fatto
+- ▁schä
+- kultur
+- ▁honours
+- леш
+- ▁Bw
+- ▁railing
+- альны
+- ▁vieille
+- ▁Ке
+- ▁Ні
+- ▁squad
+- ▁sulphur
+- ▁retir
+- кую
+- รา
+- ▁transformation
+- rabi
+- ▁harris
+- ▁Usono
+- oa
+- خان
+- ▁цікава
+- ▁sticking
+- ▁Kigali
+- ▁fringe
+- abba
+- ▁doet
+- уға
+- шай
+- آ
+- ▁farms
+- ▁gehad
+- adora
+- ▁mariner
+- ▁zahlreiche
+- torio
+- ▁uncon
+- ▁progressive
+- ніцы
+- ▁aisle
+- ยา
+- ▁lebt
+- ▁ĉefurbo
+- ▁உள்ள
+- ucht
+- ▁franz
+- ▁dominant
+- troph
+- ▁geluk
+- алы
+- け
+- ▁eyelids
+- ▁portugal
+- скім
+- ▁grind
+- ▁distr
+- ▁malalt
+- яцца
+- ▁accompanying
+- ife
+- ▁spice
+- lira
+- ▁cama
+- entrada
+- مِ
+- chon
+- ▁empor
+- Ha
+- ▁godfrey
+- bbing
+- ▁Cho
+- ▁orienta
+- ▁conferred
+- conc
+- ▁quiver
+- ном
+- hafte
+- ▁얼마
+- teka
+- ▁unlucky
+- ▁neniam
+- ▁spies
+- xel
+- ▁такія
+- ▁homem
+- '18'
+- ▁suffrage
+- lessness
+- ▁wensch
+- lux
+- ▁sav
+- ▁crawley
+- ▁recom
+- ▁random
+- ▁concrete
+- ▁odds
+- eries
+- ▁싫어
+- agradaria
+- ▁offense
+- pond
+- vari
+- gge
+- reif
+- arà
+- ▁choosing
+- erken
+- ▁ache
+- ▁przez
+- ▁treating
+- ▁immortality
+- ▁contro
+- ▁vines
+- ▁lectures
+- ià
+- ▁barred
+- 世界
+- ▁replies
+- ▁welsh
+- gant
+- ▁glimmer
+- ▁Hay
+- ▁gaiety
+- ▁happiest
+- ▁deity
+- nyanja
+- volution
+- ▁atari
+- กา
+- ▁resident
+- नि
+- ▁maple
+- ▁heaped
+- ▁easter
+- ▁праект
+- ற்று
+- naweza
+- ▁прад
+- ني
+- ச்ச
+- ▁truths
+- prop
+- аты
+- ▁discontent
+- ▁déli
+- Ch
+- many
+- ▁있으면
+- cutter
+- ▁With
+- ixa
+- urar
+- graph
+- ▁weave
+- рос
+- ▁quiere
+- ▁nought
+- ▁było
+- друг
+- ▁wires
+- ্ব
+- ▁suppli
+- ល
+- ▁indebted
+- ▁chevalier
+- ▁Guerra
+- нг
+- ▁natura
+- bana
+- ອ
+- ▁Japan
+- ▁endeavored
+- ▁germana
+- ▁suggestive
+- ▁repar
+- ▁deutlich
+- ▁sant
+- ▁recommendation
+- ▁appealing
+- ୁ
+- ▁russell
+- ▁nightmare
+- ▁perched
+- nois
+- ▁ingingo
+- А
+- ▁Pres
+- ▁bitten
+- üm
+- ▁chaos
+- ▁문제
+- 겠어
+- gula
+- inka
+- tuk
+- ▁Mü
+- ▁dahin
+- okuwa
+- 中国
+- ▁quién
+- рады
+- як
+- ▁vibration
+- vole
+- ▁zones
+- ▁permane
+- byara
+- ▁stran
+- ▁dai
+- ▁mubi
+- vollen
+- uil
+- ▁arriva
+- ସ
+- ▁whereof
+- ▁presenting
+- ▁selfishness
+- mächtig
+- tumia
+- ▁aufzu
+- últim
+- ▁curios
+- ▁jener
+- meri
+- gez
+- ▁scores
+- ▁seguir
+- homa
+- ▁invested
+- ▁só
+- dala
+- ▁communities
+- ▁bakora
+- ത്ത
+- ▁올라
+- jad
+- ▁맞어
+- dependence
+- ▁sita
+- ▁lounge
+- ▁slippers
+- бә
+- ▁channels
+- ފ
+- ପ
+- ▁ministre
+- ▁jemand
+- ▁suf
+- izer
+- ▁evolu
+- ванне
+- ▁pred
+- ლ
+- ▁rains
+- valu
+- oye
+- ▁aki
+- ▁satisf
+- ыште
+- asta
+- ަށް
+- 팔
+- ▁delights
+- ▁dirige
+- ліч
+- ацыі
+- ▁кит
+- indre
+- moedig
+- ▁усе
+- ▁edinburgh
+- ▁associates
+- ही
+- ▁craig
+- ▁cheque
+- dozen
+- ▁ballad
+- hrung
+- kro
+- quarters
+- ▁ici
+- ▁tema
+- ▁friar
+- ▁Who
+- ▁먹을
+- ▁classical
+- 했지
+- ▁maw
+- тән
+- ▁automatic
+- ▁Nti
+- ▁laquelle
+- ▁fulfill
+- arbeiten
+- 중
+- ышты
+- ▁pierce
+- гарад
+- ▁Vin
+- ט
+- ▁glimlach
+- ▁mischievous
+- сте
+- ▁bunk
+- ipe
+- pto
+- ▁belly
+- ▁montra
+- ▁idleness
+- ▁pork
+- асці
+- 평
+- ▁agaciro
+- ▁yearning
+- uvi
+- ▁прэ
+- ▁bitllet
+- 列
+- zeichne
+- ού
+- life
+- ḥemmel
+- argo
+- ▁coo
+- ▁Ok
+- ▁darn
+- èl
+- 點
+- 맛
+- 했잖아
+- mati
+- ދި
+- ▁schlug
+- witt
+- ıyor
+- 술
+- ▁twins
+- boj
+- պ
+- ทาง
+- ▁obligations
+- ▁attentively
+- чаць
+- 들은
+- ▁agita
+- ▁abominable
+- ▁bila
+- ▁missionaries
+- ▁justification
+- ▁candy
+- ▁jerked
+- ▁població
+- ▁montre
+- చ
+- रि
+- ▁Mill
+- ג
+- embra
+- ▁clearer
+- ▁impact
+- ▁interna
+- ▁benson
+- ▁mediterranean
+- domo
+- ▁trad
+- ▁alluded
+- onian
+- Ro
+- ▁prosecution
+- ром
+- ▁grands
+- ▁Му
+- ▁despatched
+- ▁elles
+- fähr
+- ▁diplomat
+- ▁muttering
+- ombre
+- ▁hav
+- 仔
+- 式
+- ▁kui
+- ▁bijzonder
+- ▁akvo
+- ▁amelia
+- ▁madrid
+- fassen
+- ▁testify
+- ▁paz
+- ído
+- ▁seigneur
+- ▁месяц
+- ี่
+- ▁магу
+- ▁modified
+- jat
+- ▁smash
+- ▁pasa
+- spoken
+- 做
+- ▁rhine
+- ▁network
+- ▁affecting
+- ▁herbs
+- ル
+- ވި
+- ▁participa
+- ▁jelly
+- ▁exempt
+- ▁wanton
+- ▁herum
+- ▁arimo
+- ▁vexation
+- ário
+- ▁disait
+- ▁fret
+- frei
+- нула
+- ▁lippen
+- езд
+- ▁taxi
+- església
+- 백
+- 动
+- ▁galli
+- ▁hiram
+- ▁tack
+- پى
+- ▁கொ
+- ▁meteor
+- عَ
+- ▁kwamba
+- ಡ
+- 결
+- лося
+- ▁chased
+- ▁эти
+- komst
+- ▁pausing
+- ▁나한테
+- пай
+- ▁seamen
+- ▁erano
+- ▁marguerite
+- ▁benefits
+- ▁cuenta
+- ▁sacrifi
+- ▁rushes
+- тя
+- ▁parce
+- ▁precedent
+- ugar
+- ▁나중에
+- ▁French
+- ▁natal
+- 기도
+- ▁stables
+- ▁Són
+- ▁infancy
+- 한다
+- ▁swelled
+- interess
+- ▁Го
+- gim
+- ▁reli
+- ▁vaig
+- جان
+- ▁requirements
+- ▁buff
+- willig
+- ▁infer
+- grav
+- န်
+- 机
+- ▁Kla
+- ▁sherman
+- ▁suspend
+- ▁awakening
+- ▁operated
+- ▁proving
+- ▁dogma
+- ик
+- ▁honors
+- ▁flags
+- ል
+- ▁гэтую
+- 메
+- ▁yuko
+- yungu
+- ▁holidays
+- ▁ejaculated
+- 省
+- ▁patriarch
+- ▁consumption
+- ▁Kommission
+- ▁magi
+- ▁zee
+- ▁whirl
+- ▁universally
+- ▁самі
+- այ
+- ▁agomba
+- ▁әйт
+- ▁defy
+- ▁musician
+- root
+- ይ
+- bbie
+- ▁necessities
+- ▁pieds
+- ▁colonists
+- .”
+- ▁descon
+- ▁curly
+- мар
+- ستی
+- ▁chaste
+- ិ
+- ▁أَ
+- mvu
+- respon
+- erekana
+- ҙәр
+- nyuma
+- hild
+- ▁һин
+- скую
+- ▁entrusted
+- ▁näher
+- ▁ducks
+- dev
+- ▁attribute
+- ▁thrilled
+- ବ
+- ▁fogg
+- gezo
+- ▁respectfully
+- ▁filo
+- ▁Gro
+- ovi
+- buye
+- ▁außer
+- icyo
+- ▁transportation
+- ▁огыл
+- ▁confided
+- ▁detain
+- ▁pleases
+- ▁Sim
+- ▁overheard
+- empe
+- muzi
+- ▁theology
+- ▁laurel
+- gru
+- เอ
+- 的一个
+- ▁darum
+- ▁mei
+- ▁그거를
+- ffa
+- ▁broom
+- für
+- obe
+- stones
+- مش
+- ▁tienen
+- ▁магчыма
+- 平
+- omi
+- ▁opponents
+- jut
+- ழி
+- ▁patriotic
+- ▁Blu
+- ▁exceptional
+- 間
+- ▁wenyine
+- ▁Frei
+- ▁coincidence
+- ▁Amek
+- ▁bindle
+- ▁marco
+- щ
+- ▁okuva
+- wide
+- 正
+- ▁ascending
+- ▁newton
+- ▁vais
+- ▁mused
+- ulf
+- ▁gestures
+- bbs
+- ▁planets
+- ▁natuurlijk
+- ▁matilda
+- lep
+- ▁forte
+- ▁mice
+- issen
+- ▁disguised
+- сво
+- ▁juda
+- ▁Eŭropo
+- ▁hie
+- ▁plaza
+- ▁nadie
+- ▁echoes
+- ▁claus
+- eeuw
+- ▁zunächst
+- ▁expos
+- ▁printing
+- ▁scottish
+- kapa
+- ▁mord
+- ▁clap
+- huh
+- ▁tej
+- ▁erreicht
+- ▁systematic
+- ificació
+- ▁malone
+- hia
+- ▁دارید
+- են
+- ▁한국
+- ▁douce
+- ▁ruhig
+- ▁горада
+- mper
+- zetten
+- ▁хутка
+- setzung
+- ▁eliza
+- ▁waarop
+- 经
+- ▁tipo
+- baren
+- lega
+- ▁yagize
+- ▁traversed
+- ▁outcome
+- ▁richer
+- ▁cameron
+- eilig
+- ▁capitaine
+- ▁verra
+- ▁voted
+- ▁colle
+- ▁wharf
+- ▁gegenüber
+- ▁girdle
+- ▁brooding
+- ▁solved
+- ▁noises
+- buze
+- ▁commissioner
+- gerageza
+- ▁connexion
+- ▁strenuous
+- ▁clash
+- ▁cane
+- ▁nächsten
+- ▁lors
+- ட்டி
+- ▁свае
+- ধা
+- ▁które
+- ▁mpamvu
+- 电
+- ед
+- ▁television
+- ▁conde
+- cover
+- ▁kindled
+- ▁llegar
+- ▁antique
+- ▁ignore
+- ▁baptist
+- nais
+- ▁sina
+- ▁palaces
+- ▁rites
+- ▁construcció
+- hla
+- ▁whistled
+- ▁chil
+- aggi
+- ▁мас
+- ▁rabbits
+- ▁clutch
+- ▁governments
+- ols
+- ¡
+- 九
+- 员
+- 过
+- эг
+- ▁component
+- ansi
+- ▁crosses
+- ▁potter
+- ▁siguiente
+- boek
+- ▁amagambo
+- ▁Atlant
+- ▁schoolmaster
+- ▁rustling
+- ӹн
+- ▁width
+- ▁coeur
+- ဆ
+- ▁bayonet
+- лған
+- ▁publish
+- ▁clerks
+- ▁сер
+- ▁següent
+- ▁생각이
+- zimmer
+- ▁pobl
+- ▁fetched
+- ▁Пер
+- ▁begrijp
+- dust
+- ▁Kri
+- ▁kumwe
+- ugen
+- grado
+- ▁adel
+- ▁іншыя
+- سىز
+- ▁atom
+- tud
+- ▁llega
+- ▁diversion
+- ڑ
+- ▁нашай
+- τε
+- ▁besten
+- ▁Пад
+- ▁referring
+- ▁gaunt
+- ▁veranda
+- ▁obscurity
+- quant
+- schlagen
+- mik
+- ▁benevolent
+- ▁starts
+- रे
+- ▁interroga
+- лән
+- ▁fewer
+- ▁навук
+- ілі
+- ście
+- ▁burl
+- ▁finely
+- ▁fragrance
+- ებ
+- ▁parecía
+- holes
+- chine
+- อน
+- lò
+- ▁lachte
+- 目
+- ▁빨리
+- ▁barricade
+- coj
+- rlo
+- ▁consta
+- eien
+- ▁ordained
+- ▁osborne
+- kud
+- salt
+- andika
+- ▁whiskers
+- ▁뭐야
+- ▁fariĝis
+- ক্ষ
+- далі
+- ▁starved
+- cross
+- ▁scamp
+- ▁aristocracy
+- iş
+- ▁newman
+- ▁ھە
+- ▁moat
+- ట
+- plica
+- ▁intelligible
+- ▁butterfly
+- tı
+- cake
+- dział
+- pate
+- wenden
+- cioj
+- ▁demande
+- ▁emir
+- ▁impulses
+- ▁treatise
+- namo
+- ▁boyish
+- geri
+- wheel
+- ▁outstretched
+- ▁stony
+- Ц
+- ▁numbered
+- ▁voorbij
+- ▁mbalimbali
+- époque
+- ▁Christ
+- дае
+- ▁treu
+- scheid
+- vla
+- ▁thad
+- ▁surtout
+- ছিলেন
+- Institut
+- ▁periodo
+- ▁thickness
+- ▁мал
+- ▁Ndi
+- ект
+- ▁Sami
+- وش
+- 保
+- kwiriye
+- ▁reconciled
+- cento
+- ▁бү
+- ▁dialogue
+- ▁priscilla
+- ▁regulations
+- nji
+- burger
+- ▁rendering
+- ▁hopelessly
+- ▁doings
+- psy
+- 격
+- ▁standards
+- ич
+- ▁그러니까
+- saw
+- ▁Sun
+- ▁көн
+- tando
+- ▁그러고
+- ▁decía
+- entendre
+- ▁folgte
+- 취
+- ▁appalling
+- ▁destitute
+- ▁diligence
+- ▁disastrous
+- ▁mathematics
+- ▁columbus
+- koreshwa
+- рож
+- affaire
+- ▁roe
+- ▁trouva
+- ▁hanno
+- ▁retirement
+- ▁mend
+- ▁shark
+- ▁хат
+- vara
+- coop
+- ▁hostility
+- ▁oan
+- ▁амаль
+- ▁reptile
+- ▁wipe
+- ▁aloof
+- amor
+- ина
+- duce
+- operative
+- ▁tissue
+- ▁shines
+- دید
+- ▁hubert
+- ▁hablar
+- ▁wheeled
+- 间
+- ▁amakuru
+- ستان
+- lac
+- ffin
+- üz
+- พูด
+- saka
+- ▁alight
+- ▁Durante
+- ▁explanations
+- zogen
+- pada
+- imari
+- rrington
+- ▁іншых
+- ▁Nk
+- три
+- objet
+- авалі
+- ▁shocking
+- rink
+- ▁modes
+- ▁Шу
+- ▁sector
+- ލ
+- ▁radiance
+- ▁totes
+- ▁livres
+- doni
+- ленне
+- 겨
+- ळ
+- ▁algunes
+- stelling
+- ▁inquisition
+- ▁monseigneur
+- ▁tort
+- ▁یہ
+- ▁rocking
+- ▁hilda
+- ▁pairs
+- ▁displeased
+- ▁convincing
+- ▁않을까
+- ▁billow
+- саб
+- ▁ау
+- ▁improvements
+- یکی
+- چی
+- дэн
+- から
+- ▁Així
+- ▁spoedig
+- balanced
+- 流
+- rava
+- ▁taller
+- inus
+- கிற
+- huis
+- ▁Trans
+- ▁rocket
+- әм
+- ▁torna
+- ▁wenigstens
+- ▁injuries
+- ».
+- ▁scalp
+- amp
+- bord
+- ▁socialism
+- ▁vermo
+- 重
+- ▁Mala
+- arth
+- ▁delegate
+- ▁brazo
+- ▁enlarged
+- 쳐
+- leri
+- ▁kuyi
+- sanze
+- नी
+- ին
+- ▁projects
+- ▁piles
+- ▁jude
+- bish
+- ▁speedy
+- bres
+- ▁cemetery
+- ▁modification
+- ▁pouvoir
+- ieux
+- nab
+- euses
+- ▁quicker
+- ▁haze
+- ნა
+- ▁Familie
+- ▁ruling
+- ▁firmness
+- ▁spill
+- noma
+- ▁Tura
+- ▁reasonably
+- намі
+- atte
+- landes
+- ▁écrit
+- нё
+- ▁voce
+- 一个
+- ▁uneasily
+- ▁clapped
+- wszy
+- ▁achievement
+- ▁Kir
+- ▁sonne
+- ▁packing
+- ▁perpetually
+- ▁compar
+- льнай
+- ▁destructive
+- ацыя
+- уч
+- rating
+- imper
+- ▁mortals
+- ▁irritation
+- ▁Можа
+- dler
+- ▁może
+- ▁straightway
+- ▁handwriting
+- gestion
+- ىدى
+- 她
+- ▁Natur
+- ▁palabra
+- umubiri
+- ▁humid
+- ▁cuff
+- acca
+- arden
+- ▁Namen
+- ▁chuck
+- ▁quench
+- dita
+- ▁shifting
+- ▁flavor
+- ▁Roger
+- ▁expenditure
+- ▁peninsula
+- pari
+- ▁هستم
+- oroshye
+- ▁tension
+- stimmung
+- း
+- ▁besoin
+- eleg
+- ▁quebec
+- dication
+- ▁są
+- ▁Ча
+- ▁jurisdiction
+- vizi
+- ▁dismounted
+- ▁devoured
+- ▁baja
+- культур
+- ▁shaded
+- مَا
+- cote
+- huza
+- ▁anglo
+- ▁nahe
+- ▁sings
+- ▁wrinkled
+- ▁insulted
+- ▁өсөн
+- ziehung
+- 담
+- ▁crushing
+- weisen
+- apfuye
+- ▁burton
+- ▁miniature
+- 옷
+- ▁jersey
+- نج
+- imos
+- ▁passé
+- ▁hereditary
+- ▁allgemein
+- ▁һәм
+- ▁திரு
+- fine
+- tangaje
+- ▁concili
+- ▁hätten
+- ▁boulevard
+- жана
+- ▁liaj
+- ▁carter
+- ▁brent
+- ▁longtemps
+- ▁penetrating
+- ▁egli
+- artig
+- ▁Terra
+- ▁закон
+- шат
+- ▁meva
+- ▁taunt
+- penny
+- ▁나와
+- zibu
+- ▁soutenir
+- aggio
+- ṛuḥ
+- oubli
+- دۇ
+- ▁muerto
+- ое
+- ▁suggests
+- ▁Tod
+- ▁plaisir
+- 얘
+- ▁Roedd
+- ▁aucun
+- ▁interpreter
+- ▁Premi
+- ▁hinzu
+- ▁netta
+- ▁Bericht
+- ▁retiring
+- attend
+- ▁똑같
+- ▁trout
+- ▁retard
+- च्या
+- ▁skilled
+- ▁Kam
+- alia
+- ▁ніхто
+- present
+- ্র
+- ▁propre
+- 이지
+- ▁inhabit
+- ▁konnten
+- اح
+- ▁fry
+- үз
+- ▁extravagance
+- ▁boulder
+- ނަ
+- 城
+- ế
+- 손
+- ▁cristian
+- ▁omitted
+- ▁Krieg
+- دې
+- ▁parlament
+- ▁verlassen
+- ▁Belg
+- ያ
+- ▁연락
+- ▁jokes
+- ▁deprive
+- ▁vows
+- ▁debe
+- ▁variations
+- ▁dances
+- ▁நட
+- üste
+- انه
+- tuye
+- ▁triumphantly
+- école
+- lga
+- ▁kwanza
+- ▁magnificence
+- ▁sydney
+- avuye
+- ulous
+- ▁lleg
+- ▁пыта
+- ▁impartial
+- ▁Bob
+- misch
+- étais
+- ▁grapes
+- bte
+- რი
+- kamu
+- ▁fluttered
+- ▁чынам
+- ▁liep
+- ▁religi
+- İ
+- ▁baltimore
+- ▁mañana
+- ▁grievance
+- itari
+- ▁grains
+- posit
+- ▁slate
+- ▁hoor
+- ▁caf
+- ▁emblem
+- 십
+- borne
+- ▁growl
+- ▁concur
+- dist
+- ▁stooping
+- ▁gripped
+- ▁creative
+- ufer
+- män
+- ▁acabar
+- ▁largo
+- ▁dominions
+- 육
+- ▁Dal
+- ▁judith
+- ▁skipper
+- ▁шул
+- ▁facility
+- ▁projet
+- ▁sexes
+- ▁Ell
+- serve
+- ▁exertions
+- ▁lupin
+- ▁efficiency
+- ▁hypothesis
+- ▁кӱ
+- ▁Regi
+- ▁silken
+- flug
+- ▁squat
+- মে
+- tegeko
+- ▁communications
+- ▁reformation
+- ▁vapour
+- ▁grunt
+- ▁captains
+- ▁maji
+- յ
+- ▁könne
+- ▁captivity
+- ▁tuj
+- plate
+- verkehr
+- lito
+- ▁Smith
+- ▁trojan
+- ▁kwiga
+- sibil
+- ▁کسی
+- ▁waki
+- autant
+- ▁volver
+- ▁devised
+- ▁või
+- ▁longest
+- ▁stuffed
+- '0'
+- ▁sew
+- ▁ravag
+- ▁accurately
+- ▁khan
+- iera
+- ▁scrambled
+- ▁crowns
+- өт
+- ▁tbe
+- 일이
+- ▁busca
+- lait
+- ▁directing
+- chard
+- ▁handy
+- ▁sym
+- Ö
+- слуха
+- ▁mathematical
+- ▁நான்
+- ▁cancel
+- ▁bleak
+- njo
+- ▁quello
+- ▁situé
+- cencia
+- ▁arabs
+- ▁halls
+- ▁में
+- ▁odious
+- ▁solemnity
+- шэ
+- ▁traits
+- ▁پار
+- ▁psychic
+- ▁싶은
+- الت
+- ▁videbla
+- guzi
+- ŝi
+- illy
+- ▁coals
+- oul
+- ▁Après
+- ▁mesa
+- ▁Ды
+- ▁beset
+- ▁stoop
+- bide
+- کا
+- чыла
+- illi
+- ▁алып
+- ▁receiver
+- kijken
+- nker
+- ▁기억
+- éré
+- ▁fuller
+- ▁verso
+- ▁kerk
+- cloud
+- ▁instinctive
+- ▁가야
+- 망
+- ▁outburst
+- ▁überhaupt
+- ▁Rat
+- ▁dusky
+- คร
+- father
+- разуме
+- ▁Clàudia
+- 참
+- ▁частка
+- nale
+- ▁collector
+- ▁трох
+- ஞ்ச
+- 機
+- ▁boyhood
+- ▁cherish
+- ▁flattering
+- тычна
+- uyu
+- ▁graces
+- ▁Ня
+- ▁sixpence
+- ▁flute
+- oper
+- ▁prakti
+- stream
+- ▁leapt
+- ▁politische
+- ▁cultivate
+- ▁keiner
+- ▁millionaire
+- ▁spaces
+- ப்பட்ட
+- ົ
+- ▁stunned
+- ▁збіра
+- ▁peuvent
+- ▁resultat
+- handlung
+- ▁algunas
+- naw
+- indo
+- ▁flutter
+- imy
+- party
+- ▁zweite
+- ▁actua
+- ▁inzira
+- rän
+- ▁mauvais
+- ▁precision
+- muha
+- gevolg
+- itzar
+- ▁scant
+- ព
+- ▁Dolça
+- ▁valentine
+- ▁tarzan
+- édé
+- ▁antwort
+- ▁pater
+- laire
+- ਵ
+- ▁resto
+- naire
+- pse
+- jyanye
+- ▁Gold
+- rui
+- obscur
+- ▁spat
+- ▁другі
+- damente
+- ▁cruz
+- ▁ramp
+- ▁ئەوە
+- ▁comercial
+- ▁confidently
+- kaka
+- ▁silvery
+- 기가
+- ▁hitch
+- operation
+- ▁impetuous
+- ▁arranging
+- ▁juge
+- ▁advances
+- cza
+- ▁cheering
+- ĵ
+- ▁strokes
+- лем
+- ▁manche
+- atta
+- meḍ
+- 적인
+- ▁snakes
+- оу
+- શ
+- ▁certificate
+- ト
+- ▁Phil
+- ▁niño
+- ▁Llan
+- cera
+- ▁tuvo
+- lico
+- ▁venait
+- ▁enclosure
+- ▁pluraj
+- მო
+- ▁anon
+- شون
+- рі
+- ▁schönen
+- ▁offen
+- jah
+- ▁hacía
+- ▁fortified
+- fahrt
+- ▁garret
+- uan
+- ▁mercury
+- ▁hospitals
+- ▁integrity
+- نظر
+- ▁campbell
+- amahoro
+- 이가
+- ના
+- rouw
+- deg
+- ▁এক
+- ▁Усе
+- ное
+- ▁pigeon
+- ▁columbia
+- ▁pamphlet
+- 想
+- ▁defensive
+- ▁jog
+- ▁ідзе
+- ▁nader
+- ▁patrick
+- aĝo
+- ▁baş
+- elek
+- ▁fainted
+- ▁procura
+- Abanyarwanda
+- ▁campan
+- ▁prins
+- 붙
+- ▁posterity
+- ▁پیش
+- castle
+- buri
+- gesetz
+- ▁lwe
+- ▁utiliz
+- ▁okaz
+- อร์
+- ▁hangs
+- ▁acceptable
+- яд
+- añ
+- ▁pena
+- ிட
+- umupira
+- ▁compartment
+- ▁illuminated
+- ▁courageous
+- ▁investigate
+- ووە
+- ▁julien
+- hinda
+- ▁maintenance
+- ▁sinful
+- lée
+- وس
+- klä
+- uca
+- ▁leak
+- ▁taifa
+- lado
+- riba
+- ▁ordinarily
+- 셨
+- ▁спр
+- ▁cierto
+- ▁India
+- clar
+- пле
+- come
+- ▁baffled
+- ▁Have
+- ▁bom
+- غى
+- cente
+- ▁contemplate
+- ийн
+- cem
+- ▁hardships
+- ▁repel
+- 当
+- nigh
+- yumva
+- تے
+- রি
+- gehalten
+- ေ
+- нет
+- ▁nigger
+- һең
+- ▁toys
+- ▁confinement
+- ▁relatively
+- یش
+- ▁fari
+- ੱ
+- 尼
+- ▁chagrin
+- ▁falcon
+- ilian
+- шко
+- ▁geese
+- ▁wherewith
+- leḍ
+- ▁capo
+- ▁bako
+- ▁feder
+- ▁garland
+- nisse
+- ▁adrian
+- orden
+- ▁wis
+- ▁societies
+- kę
+- abira
+- ▁sehemu
+- ▁akora
+- ▁haz
+- ▁regió
+- ▁kapit
+- ▁luce
+- yenna
+- ▁aspects
+- ▁ултыр
+- ▁nina
+- ια
+- ▁billet
+- ▁Sir
+- ▁indication
+- ▁hurricane
+- ▁oxygen
+- ▁shelves
+- cino
+- ▁pian
+- ▁akked
+- ▁franca
+- ▁dipped
+- тө
+- ▁negre
+- ▁locke
+- 적이
+- ▁muscular
+- fahr
+- რა
+- ▁clover
+- posto
+- niu
+- ій
+- ват
+- ▁guinea
+- ▁comprehension
+- ใช้
+- ▁haggard
+- ▁gambling
+- ▁patty
+- ▁Franc
+- ▁marshall
+- lir
+- kuvu
+- ▁bravery
+- ▁tests
+- ▁requi
+- ▁souvenir
+- liśmy
+- ▁thump
+- ▁franco
+- drift
+- ▁advent
+- ▁Tele
+- ગ
+- ▁grâce
+- ▁katharine
+- ▁imperative
+- ▁southwest
+- ▁gladstone
+- vě
+- zamu
+- ▁бик
+- അ
+- ▁restaur
+- ▁coolness
+- कि
+- wife
+- ▁intuition
+- ▁музыка
+- тел
+- ▁irony
+- zentr
+- ▁rattled
+- றை
+- ▁mesure
+- ▁winged
+- ▁verko
+- ▁rake
+- ▁bruder
+- ▁District
+- 죽
+- ▁familiarity
+- رَ
+- ندا
+- çon
+- ▁muzzle
+- ▁enduring
+- ▁transmit
+- izen
+- ▁hohen
+- ▁joyce
+- hed
+- ▁désir
+- ▁Musik
+- xer
+- bido
+- 打
+- ▁beetle
+- größte
+- ▁해도
+- ungu
+- ▁ifite
+- ية
+- 했다
+- tala
+- ▁ruffian
+- ▁walker
+- ▁visual
+- ▁naughty
+- ▁nubwo
+- ▁peas
+- ▁bonheur
+- ▁theatrical
+- course
+- ▁exploration
+- ▁freunde
+- ▁gilded
+- fren
+- ▁کند
+- pipe
+- бла
+- ibihe
+- ▁chaise
+- ▁distingu
+- ▁monuments
+- ▁wolle
+- guma
+- ▁jedem
+- fällt
+- ▁komple
+- ▁doktor
+- ▁crawling
+- ▁wade
+- halte
+- დი
+- ktu
+- ▁Anti
+- ndung
+- musik
+- ▁militia
+- mije
+- abona
+- hort
+- 벌
+- ŝo
+- ▁mitten
+- ▁greedy
+- юся
+- ibya
+- '50'
+- tett
+- ▁aquellos
+- ▁arctic
+- mast
+- ▁investigations
+- ▁overthrow
+- 制
+- ും
+- fitting
+- দা
+- дал
+- ḍen
+- ▁Feder
+- ัด
+- ▁spurs
+- migrant
+- റ
+- ▁rectangular
+- ▁deacon
+- ▁patricia
+- ▁Ala
+- ète
+- วัน
+- ▁hilo
+- بان
+- època
+- ▁Mari
+- ▁assailant
+- ▁supplement
+- ায়
+- ▁crave
+- lough
+- ▁soda
+- ▁cadet
+- 타
+- kai
+- ▁surmise
+- ▁minut
+- ▁convicted
+- huye
+- ▁identify
+- ▁scouts
+- ▁welk
+- tuuka
+- ▁carlos
+- ގަ
+- ody
+- ▁vacation
+- ▁грошы
+- 줘
+- ▁cay
+- сно
+- ▁cushions
+- ▁conoce
+- ▁pictured
+- ▁perpendicular
+- ▁uburenganzira
+- ▁Umugore
+- enzi
+- anglais
+- ▁Black
+- ▁florida
+- ▁foreigner
+- unta
+- ▁undisturbed
+- пала
+- hic
+- ▁simplest
+- ruhu
+- ▁склада
+- mö
+- icio
+- ▁sede
+- ▁anderson
+- лена
+- ▁روی
+- ▁myriad
+- ▁pensar
+- กระ
+- ▁بال
+- ▁accidentally
+- thus
+- ▁langs
+- ▁scaffold
+- körper
+- ▁frenzy
+- spiegel
+- ▁prou
+- ▁хоча
+- ச்சி
+- ▁mutton
+- ▁oracle
+- zaam
+- ▁jeszcze
+- ▁پس
+- ▁monarchy
+- ▁сай
+- brou
+- ▁Tiuj
+- ▁poca
+- ▁چرا
+- rgan
+- étaient
+- oog
+- fest
+- ▁deriva
+- ▁homer
+- tives
+- ▁Ihr
+- 웃
+- Pierre
+- ▁exhaustion
+- gde
+- zijde
+- ▁зас
+- führung
+- ▁dishonour
+- ች
+- ▁rhythm
+- ली
+- ▁musi
+- Univers
+- ▁cens
+- рай
+- سر
+- ▁Kun
+- lata
+- ▁admirably
+- É
+- punt
+- ▁trink
+- ▁solicitor
+- ▁pessoa
+- ▁ranged
+- ذا
+- ਪ
+- ލަ
+- ▁shares
+- dh
+- ▁narra
+- 바
+- قل
+- liɣ
+- ▁atoms
+- кова
+- ugo
+- ▁libera
+- ▁vrees
+- ▁attributes
+- ▁먹는
+- нім
+- umuhanda
+- ▁sermons
+- ▁blazed
+- ▁posture
+- ▁corinth
+- ▁cyn
+- ▁polic
+- ▁frigate
+- ▁jason
+- ▁drap
+- ▁dose
+- ▁oval
+- erden
+- сё
+- дзіць
+- ▁Kol
+- ▁elastic
+- ishimiye
+- ▁imperious
+- ▁асоб
+- ▁shepherds
+- ▁stella
+- lusive
+- ▁bisher
+- ▁genre
+- ▁seaman
+- reiche
+- ▁шӱ
+- ▁findet
+- ▁sterk
+- ▁grandaj
+- ▁precautions
+- 的一
+- мын
+- ктар
+- cab
+- ▁proven
+- ▁physically
+- 독
+- つ
+- orge
+- ▁Mol
+- puesta
+- roep
+- ▁omit
+- ▁keel
+- ▁stets
+- ▁wrecked
+- ▁zulke
+- оз
+- puta
+- ▁Ton
+- ▁bern
+- influen
+- рэн
+- ▁defects
+- ▁würden
+- ▁complet
+- ▁waggon
+- ▁vient
+- servi
+- ▁smoothly
+- friend
+- ော
+- คือ
+- cali
+- که
+- ▁swiss
+- yah
+- ӡа
+- amiento
+- ▁partisan
+- ▁einde
+- ▁Bol
+- ▁saison
+- ▁estimated
+- イ
+- ▁entreated
+- ▁piazza
+- ендә
+- ▁confound
+- ▁Dun
+- asanzwe
+- ▁sane
+- ▁neun
+- ▁gems
+- ▁knit
+- ▁ushobora
+- ▁workman
+- пло
+- ▁barons
+- ▁geval
+- ▁blight
+- іла
+- ▁hysterical
+- เส
+- ▁elbows
+- ▁bulungi
+- ски
+- ▁foc
+- bá
+- ถึง
+- ▁trio
+- ▁App
+- ены
+- گیر
+- ▁photographs
+- ▁overtake
+- ▁Bill
+- arrêt
+- ີ
+- ▁karakter
+- hera
+- ▁کم
+- würdig
+- ggio
+- umutekano
+- masa
+- ▁کہ
+- ▁jede
+- ragon
+- ▁Außerdem
+- ▁ferocious
+- ▁moscow
+- ▁leib
+- க்கி
+- яй
+- ▁hasi
+- қа
+- ▁teatro
+- ҵа
+- ▁titan
+- ตร
+- sucht
+- ız
+- ▁currents
+- link
+- ▁maurice
+- ▁männer
+- ▁sensual
+- ▁obres
+- carp
+- ▁کنی
+- ▁frowning
+- ▁resent
+- ▁volley
+- ▁adversary
+- بَ
+- ▁senseless
+- fak
+- ▁whig
+- lyk
+- ▁mukuru
+- اره
+- bumenyi
+- กล
+- ▁pesa
+- ▁만약에
+- ▁homely
+- ował
+- ▁reprodu
+- ▁alc
+- ▁blend
+- hre
+- ▁amigos
+- ▁inexplicable
+- wür
+- 位于
+- ▁arouse
+- ▁travaille
+- ▁раслін
+- ▁chick
+- ▁cavall
+- ▁chieftain
+- strich
+- ▁consume
+- ▁Tiam
+- ▁besteht
+- ▁Tella
+- ndel
+- ▁defile
+- нкі
+- ▁Christian
+- ▁relics
+- બ
+- ಹ
+- ▁shaggy
+- ▁Minister
+- ▁verst
+- ности
+- note
+- ▁baada
+- ǧǧa
+- ▁shrug
+- ▁giles
+- ▁scots
+- ▁chacun
+- iken
+- ▁Joseph
+- estro
+- bazi
+- শি
+- ▁worker
+- nham
+- ▁Чы
+- ▁heftig
+- द्
+- ▁cust
+- ▁fod
+- ▁اما
+- ▁nombreux
+- ▁apenas
+- ▁concentration
+- وز
+- ▁overflow
+- ▁varias
+- ▁buon
+- ▁regained
+- ▁contribution
+- ▁cicero
+- ▁Tots
+- ggy
+- rudi
+- ɣer
+- ுடைய
+- оуп
+- ▁vivis
+- ▁fent
+- бой
+- гаспадар
+- ▁ribbons
+- ngereza
+- ▁eya
+- pec
+- ▁epic
+- veld
+- ▁stellte
+- ▁edifice
+- ▁fuß
+- waarde
+- ▁inspect
+- ▁Ки
+- ▁regardless
+- ▁wisest
+- american
+- gufasha
+- ▁convertir
+- ▁아무
+- ▁enforced
+- indig
+- ▁Més
+- ▁spencer
+- ▁arbo
+- ▁deutschen
+- ▁benches
+- ▁هەر
+- ▁pièce
+- ▁starvation
+- 色
+- ▁feebly
+- ▁아빠
+- ▁accessible
+- ▁desiring
+- improvis
+- звычай
+- ▁avis
+- ▁böse
+- ▁100
+- बा
+- ▁nikola
+- ▁accommodate
+- ▁vineyard
+- خور
+- ▁bruit
+- мян
+- ▁머리
+- ▁shutting
+- ▁ghostly
+- ▁masculine
+- 테
+- ▁savoir
+- 角
+- енно
+- ▁사진
+- Umu
+- lamp
+- knife
+- ▁maintaining
+- ▁gild
+- ▁Grup
+- ▁oaths
+- enze
+- ọ́
+- ▁inference
+- stok
+- ggia
+- ▁moo
+- ▁recoil
+- ▁denote
+- ▁shady
+- ▁bracelet
+- valent
+- ▁sympathies
+- воль
+- ▁congratulate
+- ▁choking
+- ▁willie
+- ▁silas
+- ▁inherit
+- ▁په
+- ▁indies
+- kret
+- 머
+- ▁passatge
+- ▁accumulated
+- ▁hints
+- ▁occupations
+- ▁geef
+- алар
+- ▁absorb
+- ▁langer
+- kab
+- ▁uruhare
+- よ
+- ▁ijambo
+- ▁knie
+- ▁gothic
+- grim
+- ▁già
+- ▁shutters
+- ▁woven
+- bouw
+- ▁gezelschap
+- ▁outbreak
+- کش
+- ▁gonna
+- ▁brillant
+- prom
+- isé
+- obi
+- ▁milan
+- ▁norah
+- ▁trouw
+- ▁gusto
+- ▁jefferson
+- ▁niwe
+- ▁één
+- дзіў
+- ▁czar
+- ▁betwixt
+- ▁difícil
+- руш
+- ▁frauen
+- ▁Tess
+- ▁amis
+- ▁melting
+- ▁haunt
+- deur
+- ▁protecting
+- ▁fugitives
+- ▁abzu
+- วิ
+- ▁gewoon
+- yeho
+- ▁Pin
+- 최
+- 투
+- 확
+- ▁sovint
+- umwaka
+- gic
+- ति
+- ▁кәрәк
+- 啲
+- ▁milion
+- ▁hauled
+- খা
+- ▁safer
+- ▁morsel
+- ▁commis
+- izindi
+- parti
+- ▁adored
+- ண்
+- ▁traf
+- して
+- ລ
+- ɓ
+- gesproken
+- 八
+- schluss
+- ▁reproduce
+- auber
+- ▁tumble
+- espai
+- ▁cognom
+- ▁launcelot
+- téri
+- ▁implies
+- ▁boundless
+- ▁inwardly
+- ▁freilich
+- وری
+- ▁exhort
+- tijd
+- ▁donker
+- eilen
+- brush
+- ▁impossibility
+- ▁plantes
+- ordon
+- ▁parliamentary
+- рых
+- быҙ
+- ▁switzerland
+- ▁gdy
+- ▁healed
+- ▁lyric
+- ▁Aquestes
+- гуля
+- ▁ivy
+- টা
+- 球
+- cée
+- ▁commandant
+- ▁hardest
+- grin
+- lc
+- épa
+- ▁어차피
+- ▁menor
+- ▁elephants
+- ▁traders
+- lj
+- ▁сваім
+- ▁Sara
+- ▁conta
+- ▁gracia
+- һен
+- ▁Fahr
+- streng
+- rati
+- ▁fortuna
+- ▁excused
+- ▁uchun
+- precia
+- ▁warwick
+- appa
+- ▁captives
+- ▁rivero
+- ама
+- plane
+- 브
+- rückt
+- ▁aucune
+- ▁пачаў
+- ▁Mwa
+- ▁politicians
+- ▁salmon
+- ▁Indian
+- ▁espanyol
+- வேண்டும்
+- ▁peeped
+- ▁consequent
+- ▁vowed
+- нта
+- sehn
+- hte
+- сам
+- ▁footman
+- ▁avert
+- বি
+- ▁vroeger
+- 해야
+- ვი
+- ingabo
+- ▁malgranda
+- ▁amendment
+- ▁cabo
+- ▁nests
+- twali
+- forma
+- ▁capitan
+- zko
+- ▁baskets
+- yê
+- ▁Ist
+- ▁bricks
+- ▁lowering
+- ▁okupiĝis
+- ▁motto
+- lava
+- ▁Pep
+- ▁figured
+- legged
+- ▁carrière
+- іі
+- ▁benevolence
+- ▁okul
+- ғыҙ
+- ▁vegada
+- agiye
+- länder
+- ▁prussian
+- лон
+- ိ
+- ▁grudge
+- ▁blas
+- ▁temporada
+- ▁provoca
+- ▁relax
+- ▁concealment
+- ▁rugo
+- ▁restis
+- ▁добры
+- 쉬
+- ▁nadat
+- ▁daarop
+- ologio
+- ▁shew
+- ▁negotiations
+- onic
+- ▁finale
+- ạ
+- бар
+- 줄
+- ың
+- ▁Bank
+- día
+- rote
+- ▁courtiers
+- ▁Хто
+- ▁kvin
+- ▁Дзе
+- ▁profane
+- ▁grandma
+- ▁nerv
+- ▁ntiya
+- ▁stubb
+- ▁fists
+- ▁также
+- gesehen
+- jara
+- yari
+- ekka
+- mbura
+- arlo
+- наў
+- ები
+- ▁vigil
+- ashakaga
+- take
+- ▁kwamen
+- ▁نییە
+- sobanura
+- ▁threads
+- хь
+- ▁cree
+- بح
+- ▁traverse
+- ▁reluctance
+- 曾
+- ▁tool
+- ▁volunteers
+- yum
+- lgan
+- ▁وجود
+- ▁existe
+- ▁lifeless
+- ▁bandit
+- ▁ponto
+- ▁Bild
+- ому
+- ▁kugera
+- ▁alighted
+- เพื่อ
+- ▁spruce
+- ▁Tru
+- ছ
+- ▁gör
+- ▁furtive
+- El
+- ▁quaker
+- Ө
+- থ
+- ės
+- ▁provoked
+- ▁spielte
+- ▁allegiance
+- ▁achilles
+- ▁endeavouring
+- bush
+- َيْ
+- ık
+- ▁الأ
+- தற்கு
+- ▁reddy
+- ▁straining
+- ▁conclu
+- ▁torches
+- ▁incessantly
+- 럽
+- ▁acquisition
+- ▁straightened
+- Orient
+- ніца
+- کھ
+- 个
+- rce
+- 별
+- ▁jonathan
+- ▁thigh
+- ▁clair
+- ორ
+- ▁interpreted
+- ▁такой
+- ▁Boston
+- анне
+- дары
+- zeichen
+- নের
+- ▁Yi
+- ▁miaka
+- ▁aŭto
+- 태
+- ▁muryango
+- ▁taper
+- ▁그렇지
+- ında
+- ▁resolutely
+- ▁euro
+- ▁miners
+- ř
+- ▁caprice
+- closed
+- agh
+- ▁flores
+- ▁kibi
+- other
+- 합
+- gründe
+- เข้า
+- worm
+- ▁bondage
+- ▁throws
+- ffel
+- ▁vuur
+- ▁gewalt
+- ▁vernon
+- ▁exceptions
+- ▁பொ
+- مار
+- vali
+- ▁victories
+- ▁dunkel
+- ▁brightened
+- ▁fay
+- bewegung
+- zaal
+- tabl
+- թ
+- ада
+- ▁educational
+- brechen
+- ▁gerard
+- ▁preface
+- ▁implored
+- ষ
+- ▁wardrobe
+- eusement
+- ▁relates
+- sbury
+- ▁devour
+- 니깐
+- ▁tutta
+- ▁charitable
+- 六
+- state
+- ▁Kuki
+- ▁timidly
+- ▁Sor
+- ▁булма
+- ▁drums
+- hame
+- ▁registra
+- ▁eky
+- سک
+- ▁enquired
+- ▁computer
+- ▁penitent
+- ▁knave
+- ▁ordering
+- ▁гурт
+- ▁cerv
+- ías
+- ▁아냐
+- flex
+- ▁recording
+- ▁rattling
+- 两
+- валь
+- ▁estudiar
+- ғ
+- senya
+- ▁variable
+- ▁importun
+- ▁batalla
+- ▁baku
+- ▁président
+- हरू
+- ▁approaches
+- ▁altitude
+- ▁lilies
+- ▁Miquel
+- jol
+- ▁snarl
+- ▁cheers
+- père
+- ▁tickets
+- ▁luckily
+- ▁horas
+- gaban
+- কার
+- station
+- цей
+- ▁nomon
+- ▁pilgrims
+- ▁energi
+- enjawulo
+- ▁néixer
+- vam
+- ▁prone
+- мон
+- ▁пытанне
+- ▁Av
+- ▁proces
+- ▁fis
+- ▁valjean
+- ▁aims
+- prac
+- ▁scienc
+- ▁allait
+- ща
+- ▁musicians
+- spor
+- ▁forbade
+- ▁dictate
+- mberg
+- ▁tranquillity
+- vado
+- ▁Тады
+- ▁많은
+- ғә
+- ▁disp
+- ▁molest
+- দের
+- bewe
+- ▁prevailing
+- ്ര
+- ▁wail
+- ▁kurusha
+- ם
+- рок
+- ▁sidenote
+- mama
+- ▁cope
+- ufe
+- ▁pique
+- หน
+- roepen
+- ▁Werk
+- ▁стара
+- ▁brandon
+- ▁Pod
+- ▁reicht
+- ▁Таксама
+- زه
+- ▁corrected
+- ▁Pap
+- ▁flap
+- 본
+- ▁House
+- ▁будзем
+- ▁knob
+- visa
+- ▁Cent
+- ▁unten
+- ▁invaded
+- ▁grill
+- ▁America
+- bak
+- лаў
+- 시간
+- ច
+- ▁sheridan
+- ▁거잖아
+- ▁opposing
+- amye
+- ▁pense
+- ▁queried
+- 대로
+- fly
+- ▁kuji
+- ▁sweden
+- ▁noiselessly
+- ▁cumi
+- ▁apartenas
+- ▁foreman
+- ▁weakened
+- uré
+- ▁tenor
+- ▁huddled
+- 解
+- 影
+- ▁시험
+- ▁Со
+- ▁aversion
+- ынан
+- வும்
+- ▁persia
+- kole
+- ▁ruby
+- ▁kicking
+- ▁ether
+- ▁emphatically
+- ▁адразу
+- ▁subsided
+- ▁deception
+- ▁inaugur
+- ě
+- ▁manual
+- ▁mostra
+- дь
+- geme
+- ▁abolition
+- ▁shrunk
+- ▁vuestra
+- ▁depot
+- ▁bintu
+- ▁arran
+- wash
+- ▁vermi
+- 買
+- 少
+- ▁cheerfulness
+- ён
+- ▁indicating
+- ▁increases
+- сон
+- ▁deixar
+- ▁Central
+- ▁simul
+- кол
+- نەوە
+- ▁adjusted
+- lien
+- ▁plucked
+- ▁kannst
+- али
+- އު
+- ▁leicester
+- ▁despe
+- ▁länger
+- blue
+- त्
+- ▁chaff
+- ▁hushed
+- ▁Толькі
+- ▁inherent
+- rza
+- ▁theodore
+- ▁horseman
+- ▁chivalry
+- ▁tristram
+- posing
+- sworth
+- ▁norway
+- них
+- atura
+- মান
+- ▁moods
+- ன்று
+- ▁embarked
+- ▁milton
+- ▁rowed
+- ▁vinegar
+- genera
+- ▁mots
+- rari
+- ▁bruised
+- ▁consecrated
+- лак
+- ▁hungara
+- кет
+- ▁colossal
+- bogen
+- ▁gustav
+- лож
+- ▁scratched
+- base
+- ▁creo
+- ▁palmer
+- ▁daarom
+- ließ
+- geschäft
+- ▁trod
+- ▁privat
+- ▁God
+- ▁Pia
+- ▁jahren
+- acle
+- mbwa
+- ▁vesti
+- ▁gifted
+- تې
+- рк
+- ▁дзве
+- 귀
+- ▁junction
+- ▁minha
+- ▁adjust
+- ▁certa
+- taire
+- ▁voelde
+- cero
+- ▁accomplice
+- 太
+- gou
+- ▁pledged
+- ▁Nous
+- ▁legends
+- ▁кел
+- ▁jehovah
+- ▁español
+- ▁großer
+- ▁milly
+- ▁sighs
+- ▁Om
+- bata
+- ▁customer
+- cinq
+- ▁plein
+- ▁engaging
+- ▁joel
+- stell
+- brown
+- ▁wrung
+- чер
+- ▁ruhe
+- siri
+- wehr
+- ▁alexandria
+- ขึ้น
+- ▁дзіця
+- ▁supo
+- schütz
+- ىرى
+- ▁donald
+- ▁casos
+- ▁aching
+- ▁listless
+- سل
+- ▁Sto
+- ▁patter
+- ▁exclaim
+- ▁compta
+- retse
+- мысл
+- ▁wilful
+- лган
+- ▁hank
+- ▁shack
+- ▁interfered
+- höhe
+- də
+- ▁promet
+- ਾਂ
+- ▁parmi
+- isiert
+- ▁judah
+- ▁quivered
+- ▁barbarians
+- ▁landoj
+- 릴
+- ▁città
+- ▁раней
+- تي
+- ▁stesso
+- ▁eneo
+- zeg
+- ▁eble
+- ▁تح
+- moral
+- 느
+- ▁hemel
+- abagabo
+- ▁claire
+- ▁boer
+- ▁sidewalk
+- йце
+- étend
+- тив
+- 战
+- ▁qualification
+- ▁cometh
+- hamya
+- ▁pallid
+- haba
+- gati
+- 입
+- ಬ
+- ▁llavors
+- rende
+- ▁avoiding
+- 찍
+- ikira
+- ▁chattering
+- seɣ
+- ▁wage
+- 페
+- ▁appeals
+- ▁thanksgiving
+- ▁quanto
+- ▁Street
+- ▁angelic
+- rale
+- kî
+- ▁diplomatic
+- ▁führt
+- ▁allerlei
+- ▁dividi
+- ▁Ро
+- كا
+- کەی
+- ▁gukoresha
+- ▁cria
+- ▁temporarily
+- ▁betsy
+- 이라고
+- ▁battered
+- દ
+- ▁expanse
+- ▁abilities
+- ▁junta
+- ▁Llo
+- ▁esplor
+- ▁гэтая
+- exist
+- ▁convulsion
+- ▁cerr
+- train
+- прост
+- ▁serà
+- ▁insurrection
+- ▁pans
+- ▁può
+- ▁Nil
+- யின்
+- popul
+- ▁productive
+- ▁projected
+- ▁mop
+- ី
+- ▁kad
+- Mar
+- escri
+- philosoph
+- ▁picket
+- ▁pře
+- ▁alms
+- фон
+- ▁factors
+- black
+- ▁incline
+- ▁erwiderte
+- 看
+- strung
+- ▁freshness
+- ▁katy
+- qqa
+- ▁defending
+- ▁produit
+- ▁films
+- ▁malcolm
+- ▁biting
+- ▁cosette
+- ▁kureba
+- ▁foresee
+- ҷ
+- 河
+- ▁들었
+- ▁exerted
+- ▁komencis
+- گاه
+- oce
+- ▁ĝia
+- mbuka
+- ▁pickle
+- ▁самы
+- ▁isabella
+- ▁denomina
+- ▁baptized
+- ▁château
+- 光
+- 어서
+- ргә
+- ▁delivery
+- kung
+- ▁lustre
+- ième
+- ▁assailed
+- ▁hain
+- ▁realization
+- ▁rechts
+- وب
+- ▁Minisitiri
+- ▁imyenda
+- 交
+- ҙың
+- ▁vehicles
+- mpu
+- ▁hospitable
+- нуўся
+- އެއް
+- もの
+- glie
+- ▁сіл
+- 적
+- ▁resented
+- ady
+- ▁Chan
+- ▁bard
+- ▁bientôt
+- ▁myaka
+- ▁Anwa
+- aagd
+- ▁алма
+- ▁nigra
+- virus
+- ▁tenants
+- ▁தொ
+- iš
+- ▁ferment
+- ▁kite
+- ޫ
+- anc
+- ▁сваёй
+- ǧǧ
+- ologi
+- ▁Emp
+- ▁cheery
+- ▁insolence
+- ▁Ula
+- ▁carlo
+- ▁tudo
+- oire
+- blé
+- ziert
+- Au
+- ▁conserve
+- ▁gep
+- ▁thwart
+- ▁towering
+- ▁dawned
+- ▁precipita
+- rema
+- ▁papel
+- ▁disclosed
+- ▁acknowledgment
+- ▁unserer
+- ▁testified
+- ▁Cab
+- ▁Ly
+- '40'
+- ▁wellington
+- ▁dishonest
+- жыл
+- ▁mirongo
+- ▁curb
+- ▁같아
+- rón
+- Esp
+- ▁flashes
+- äm
+- gegaan
+- 話
+- ▁höchst
+- ▁regula
+- ruz
+- kö
+- durch
+- ▁improbable
+- ▁variation
+- ▁contributions
+- ▁behar
+- ▁альбом
+- ▁பிர
+- ▁lascia
+- ▁magazines
+- ▁nickleby
+- ▁dublin
+- arren
+- ▁facilities
+- intel
+- ▁kare
+- ▁winchester
+- ▁corral
+- ▁fût
+- ▁гадоў
+- ▁Се
+- ▁hearken
+- ▁Yett
+- ▁utilisé
+- ▁보면은
+- ▁colt
+- pois
+- ▁joyfully
+- ▁nich
+- ▁International
+- ▁perspective
+- ▁skirmish
+- 웠
+- ▁verda
+- лення
+- nimmt
+- ▁naast
+- ▁voet
+- ▁வந்த
+- ▁barrels
+- ▁teddy
+- ▁perdre
+- ▁barking
+- liste
+- ▁vriendelijk
+- ▁immoral
+- рәк
+- য়া
+- шая
+- ombra
+- عا
+- Ca
+- ▁torpedo
+- ▁whispers
+- ▁playful
+- ignit
+- bavu
+- ▁termina
+- ▁blonde
+- ▁folding
+- ▁resulting
+- ޯ
+- ▁cox
+- ▁ursula
+- ▁institute
+- ▁renov
+- eater
+- ▁convictions
+- ▁fondly
+- маны
+- kano
+- ▁senor
+- zado
+- ▁आहे
+- ▁axxam
+- ▁President
+- ▁julie
+- ṭa
+- ▁securely
+- ▁burr
+- ▁são
+- ▁conjure
+- ▁scrape
+- ነ
+- ▁жанчын
+- ▁commotion
+- ▁stating
+- ▁carn
+- ▁Пера
+- 살
+- ump
+- gata
+- ▁tennessee
+- ▁hauria
+- ▁realitza
+- ▁adviser
+- ▁heading
+- ▁generated
+- witz
+- 났
+- ▁sait
+- ▁accursed
+- ▁remonstrance
+- kungu
+- tailed
+- ▁metals
+- ҙан
+- ▁discord
+- лага
+- ▁riders
+- ▁وال
+- ▁intercept
+- ▁wield
+- ▁irene
+- ▁cromwell
+- ▁iniquity
+- ▁afresh
+- भा
+- ▁vividly
+- ▁befallen
+- ▁gushyira
+- ▁donar
+- праб
+- нікаў
+- ▁consulting
+- ▁redeem
+- ▁arches
+- ▁apologize
+- ญ
+- ▁belgian
+- ▁sota
+- ▁culpa
+- lita
+- member
+- ▁paced
+- fract
+- ▁furs
+- ▁compost
+- ▁partition
+- chau
+- ▁perceptible
+- 學
+- ▁magdalen
+- ggo
+- ▁selber
+- уж
+- yiga
+- ▁Jack
+- ▁grinning
+- ▁hoewel
+- ▁eure
+- ▁Krome
+- ▁benefactor
+- 石
+- ▁petticoat
+- ▁mortification
+- ▁collabora
+- ▁retreating
+- ▁transported
+- roch
+- ▁schel
+- kuli
+- ▁kenne
+- ▁haunts
+- ▁찾아
+- ▁creating
+- лады
+- ▁slab
+- ▁institu
+- ▁historians
+- ▁barbar
+- umuyobozi
+- ঘ
+- geon
+- ▁duda
+- ▁대학
+- ▁tumbling
+- ▁sommes
+- ▁financ
+- ▁varia
+- ▁exhaust
+- ▁abstain
+- ▁addresses
+- ▁slippery
+- ▁Fil
+- string
+- ▁deixa
+- ▁справа
+- ▁fuerte
+- pande
+- kje
+- ▁illinois
+- ▁giorno
+- ▁cigars
+- ▁tyr
+- ▁Nekk
+- ▁brust
+- nol
+- ▁forcibly
+- ▁clima
+- ▁novelist
+- ▁кем
+- ▁incorpora
+- ▁эшлә
+- бәр
+- 른
+- illé
+- ▁fanatic
+- ▁blanca
+- ▁stanley
+- ▁gateway
+- ▁flare
+- цё
+- œ
+- вест
+- ▁heeren
+- ▁fruitful
+- ▁دل
+- 只
+- ▁doubly
+- ▁دیگه
+- дей
+- 이고
+- ▁demanding
+- ▁gracefully
+- tabi
+- ▁investment
+- ▁Font
+- ▁venetian
+- ▁ngombwa
+- ▁tends
+- featured
+- strom
+- ▁sorr
+- ▁이상
+- ▁Pf
+- িয়া
+- ▁tucked
+- ▁felici
+- ▁jeunes
+- ▁prescribed
+- dräng
+- ▁geene
+- istan
+- ▁ounce
+- burgh
+- écu
+- гавор
+- coli
+- prav
+- حد
+- 즈
+- ▁ingredient
+- ▁sleigh
+- chou
+- nische
+- ▁magnetic
+- ▁Englishman
+- ▁uxxam
+- τη
+- niye
+- ▁secular
+- patri
+- ▁geheele
+- سەن
+- weza
+- ▁mariya
+- ▁calendar
+- ▁incur
+- ší
+- ُو
+- zelf
+- ▁domine
+- lington
+- อา
+- ▁mim
+- бро
+- ▁mellow
+- ▁slander
+- ▁thorns
+- ▁pregunta
+- ▁farming
+- ▁especie
+- ▁mockery
+- ▁recess
+- ▁чаму
+- ката
+- ishije
+- system
+- ▁eden
+- ▁cordially
+- ▁curses
+- ▁كۆر
+- ▁pertain
+- ▁thames
+- ▁furrow
+- ▁выніку
+- ▁années
+- ▁messrs
+- اں
+- ▁whiteness
+- tzeko
+- ficient
+- ▁extends
+- ▁icya
+- സ്
+- ▁grease
+- ▁fibre
+- ▁meilleur
+- strahl
+- Lo
+- posicion
+- ▁plight
+- lingva
+- ▁apprentice
+- شم
+- ▁nightfall
+- ▁flannel
+- итель
+- ▁ئۇنىڭ
+- ▁forsaken
+- ▁vivas
+- ▁interven
+- ▁boldness
+- år
+- гна
+- bub
+- ririmba
+- ▁footnote
+- 달
+- ▁affectionately
+- ▁duidelijk
+- ▁Час
+- ант
+- ▁cyril
+- ubuyobozi
+- ▁involuntary
+- ▁жыцця
+- ▁stumbling
+- ▁boasted
+- quai
+- ▁Dom
+- ▁ceaseless
+- ▁einigen
+- ▁Ста
+- kah
+- sigaye
+- يدۇ
+- ▁merced
+- ակ
+- txe
+- ▁Sohn
+- züge
+- ▁milito
+- 'No'
+- ▁sú
+- agne
+- ▁authentic
+- ▁shrinking
+- ▁digo
+- anisha
+- ук
+- ▁improper
+- 이래
+- auto
+- ▁genie
+- ▁counties
+- ▁sickly
+- ▁ngā
+- ▁Vil
+- kawa
+- ძ
+- koff
+- ▁implica
+- ▁fondo
+- كۈ
+- чным
+- ▁plaintive
+- ▁decrease
+- nui
+- ▁delicately
+- schreiben
+- ▁tinha
+- ▁Gerard
+- ▁falk
+- ▁lore
+- ူ
+- ▁күп
+- شا
+- ▁wistful
+- ▁situació
+- ▁dado
+- ▁russians
+- arco
+- ▁administer
+- 커
+- ▁месцы
+- ▁hides
+- aving
+- ▁Nacional
+- ▁unfold
+- емся
+- ▁выда
+- ▁wabo
+- kibuga
+- ▁وقت
+- ▁pathway
+- ▁zugleich
+- agra
+- mahl
+- ▁revealing
+- ▁disciple
+- ▁kompren
+- ▁Washington
+- ▁unmistakable
+- ▁tuft
+- ▁Каб
+- matik
+- quelle
+- pont
+- ▁deepened
+- трап
+- дзіцца
+- 已
+- ▁oppressive
+- ▁claw
+- ▁dejar
+- ▁peuple
+- acion
+- ▁leadership
+- ▁rainy
+- capitul
+- ▁behoor
+- ▁Ле
+- ▁waarvan
+- ▁vite
+- ▁fruitless
+- ▁occidental
+- ▁gradu
+- ▁kion
+- ▁pasar
+- ▁principi
+- 数
+- they
+- vik
+- ▁swearing
+- holder
+- ▁constituent
+- informa
+- ▁vassal
+- thie
+- ▁groves
+- ▁insure
+- ▁Você
+- ▁Church
+- ▁дзяцей
+- 何
+- ▁ziek
+- ▁poke
+- ▁buzima
+- يت
+- ▁pensi
+- ▁다음
+- ▁لطفا
+- ▁tapped
+- ▁шлях
+- ▁gravi
+- ▁ethics
+- ▁spectator
+- ▁перш
+- ▁enlarge
+- ▁flirt
+- lipo
+- sack
+- бур
+- ▁tolerably
+- 더니
+- ▁nymph
+- ▁muff
+- ▁Zwei
+- ▁corte
+- ▁rebre
+- ރާ
+- establish
+- ▁reluctant
+- auteur
+- ▁From
+- current
+- ▁ntibi
+- ੋ
+- ▁öffentlich
+- ▁presumption
+- ▁pamoja
+- ▁supremacy
+- ▁sparks
+- ▁fuese
+- meslay
+- эд
+- ▁wailing
+- ▁impertinent
+- ▁expresses
+- ▁absurdity
+- ▁montague
+- ▁provides
+- ҡта
+- onze
+- ▁superintend
+- ▁confirmation
+- gypt
+- ▁Laura
+- طر
+- ▁cartridge
+- ▁egin
+- muni
+- ▁Karl
+- ▁Poli
+- ▁oar
+- ▁medieval
+- zmer
+- ▁caring
+- ▁leise
+- ив
+- ▁populace
+- fällig
+- ĝas
+- ▁speechless
+- ▁vergessen
+- ▁gratify
+- ▁muda
+- ▁outlines
+- 말
+- celle
+- ▁chess
+- ▁পর
+- ▁disgraceful
+- iéndose
+- 봐
+- ▁যা
+- ḥemm
+- ▁hercules
+- umuhungu
+- voting
+- ▁senhor
+- ababyeyi
+- ▁curtis
+- zajya
+- ▁universitato
+- ▁vroolijk
+- ▁ocasion
+- 林
+- ▁서울
+- ▁yao
+- pound
+- ▁temperance
+- ▁extinct
+- цәа
+- ▁kids
+- ޮ
+- ▁auge
+- ▁একটি
+- ▁bouquet
+- Mi
+- ▁pratt
+- жар
+- ▁frente
+- schouw
+- appelle
+- ▁maeneo
+- ▁Mir
+- ▁monsters
+- ▁unfair
+- ▁stata
+- ▁ніяк
+- ڕە
+- ▁rowland
+- ocupa
+- rura
+- pende
+- 언
+- rezi
+- 절
+- ambient
+- neg
+- ▁airy
+- ▁часта
+- ▁boar
+- ▁occupying
+- goye
+- ▁Yidir
+- anira
+- ataj
+- sulta
+- ▁Lle
+- guen
+- zog
+- wego
+- ▁그까
+- ▁потым
+- ▁sofort
+- ▁condemnation
+- теш
+- ▁relinquish
+- ғас
+- ۇق
+- angiye
+- вое
+- ▁누구
+- ދ
+- tegura
+- ژی
+- ▁baked
+- ▁reconcile
+- ▁rhetoric
+- ▁lucilla
+- мест
+- ▁topics
+- ▁performances
+- ▁developing
+- ▁камп
+- চা
+- estil
+- robi
+- manya
+- ▁principalment
+- ▁nominal
+- ▁abnormal
+- trodden
+- ▁aristo
+- مې
+- ▁insulo
+- ▁steve
+- ▁statesmen
+- ndy
+- ▁сапраўды
+- ▁deference
+- 共
+- ▁прызна
+- autor
+- ▁scatter
+- ▁spars
+- ▁griz
+- ▁distrikto
+- ▁intruder
+- 냐고
+- ▁producer
+- ▁connais
+- ▁خو
+- ▁엄마가
+- ▁rhe
+- ▁jij
+- komere
+- 넣
+- ▁prinz
+- ▁sallie
+- ▁illegal
+- ▁paws
+- ▁Neben
+- kron
+- ▁quixote
+- ▁nane
+- ▁حرف
+- ▁Actualmente
+- hua
+- ▁triple
+- کت
+- vira
+- ▁трэ
+- ▁equipo
+- фер
+- hah
+- jada
+- ▁smashed
+- standing
+- ▁thirtieth
+- ▁tuza
+- ▁psychological
+- ▁ounces
+- ▁velo
+- uja
+- շ
+- ▁elegance
+- ▁throbbing
+- ▁있으니까
+- ▁recreation
+- ула
+- ▁skr
+- ▁poetical
+- ▁climax
+- ▁compli
+- ▁буй
+- 真
+- ಪ
+- ▁superfluous
+- ▁decisions
+- interromp
+- ▁Sand
+- ▁Olu
+- ▁һөйлә
+- ▁wann
+- ▁Jak
+- ▁enveloped
+- ▁projecte
+- cité
+- rome
+- ▁자기가
+- ваюцца
+- ▁sliding
+- ▁Ry
+- ▁harding
+- aller
+- ▁دور
+- kampf
+- ▁sighing
+- Hal
+- ▁société
+- ▁cleaned
+- ▁stripes
+- وي
+- ▁liberties
+- ▁wagner
+- ▁voiture
+- зан
+- ▁să
+- ▁photo
+- erweise
+- ▁comprehensive
+- ▁mowbray
+- ▁susceptible
+- wohl
+- cide
+- ▁Ty
+- ას
+- inze
+- ▁fossil
+- ▁है
+- ziel
+- ▁ponies
+- ▁күҙ
+- ▁hellen
+- ▁thoughtless
+- 王
+- ▁sordid
+- ▁peacock
+- ▁시작
+- ▁herod
+- ▁გა
+- ▁blond
+- Di
+- ▁gavumenti
+- ▁furthermore
+- larni
+- ▁Brazil
+- ▁refreshing
+- baya
+- ▁Ari
+- ▁skim
+- ▁bleibt
+- kosa
+- ▁affront
+- tsu
+- ▁mateixa
+- arbre
+- ▁jedes
+- bereye
+- ▁testimoni
+- ▁baxter
+- anthrop
+- ▁enraged
+- shingiye
+- ▁unreal
+- ▁recourse
+- ▁zoû
+- ▁Aya
+- ดู
+- ▁awal
+- ▁spaniard
+- ▁pouch
+- ▁tested
+- ▁executioner
+- ▁brazen
+- ▁habitants
+- ▁quale
+- ▁hymns
+- ▁última
+- ▁unselfish
+- ▁veritable
+- cliffe
+- ▁selten
+- ▁grange
+- қә
+- ▁scheint
+- ▁Pel
+- ▁komerc
+- ▁residents
+- ▁Damit
+- ▁mettre
+- ▁unclean
+- ▁estimation
+- ▁tiresome
+- ▁오래
+- nij
+- ▁raymond
+- ▁lunatic
+- ▁fearfully
+- दा
+- ▁bezoek
+- ños
+- ழை
+- ▁pigeons
+- ▁offender
+- '20'
+- дәр
+- ▁maternal
+- падзе
+- ícia
+- ▁Ан
+- ▁ambaŭ
+- gb
+- uhoraho
+- nium
+- ▁rogers
+- ց
+- ނި
+- ▁uncovered
+- endelea
+- ாது
+- ▁quiero
+- ática
+- ▁merciless
+- ลง
+- akomeye
+- ▁barba
+- ▁Uma
+- ▁Drac
+- ▁veut
+- жь
+- ▁repress
+- berna
+- ▁Stand
+- ▁imminent
+- dicion
+- ออก
+- ▁horde
+- ▁batal
+- ▁cleaning
+- ería
+- ▁Bɣiɣ
+- ▁такога
+- ▁clearness
+- ▁busily
+- య
+- ▁ellis
+- ▁ueber
+- bron
+- urteil
+- ▁얘기를
+- ▁astray
+- ▁naturalist
+- ▁misty
+- ▁knots
+- kê
+- ▁twisting
+- mbro
+- ▁fortitude
+- ▁rally
+- gerichtet
+- ziga
+- ▁kubaho
+- ▁infamous
+- ▁camilla
+- trof
+- ▁privation
+- –
+- ▁sovereignty
+- ▁dagegen
+- меры
+- ▁coldness
+- َلَ
+- ▁barley
+- ▁군대
+- ▁kansas
+- ▁مرد
+- مال
+- ▁stroll
+- ▁congratulation
+- 셔
+- ▁sarcastic
+- ▁siège
+- ▁Dann
+- ▁seals
+- ▁hayo
+- koop
+- ▁amounted
+- ▁unspeakable
+- ▁reçu
+- ▁pitied
+- ▁factories
+- talo
+- ▁dago
+- ccu
+- йшоў
+- ▁funk
+- ▁sojourn
+- ▁survivor
+- koresheje
+- ▁psychology
+- ▁revive
+- ▁கூற
+- ▁juliet
+- ▁ndani
+- ▁intervention
+- ▁creak
+- ▁eldon
+- ▁període
+- ▁redemption
+- கிறது
+- ▁repris
+- ▁பல
+- islam
+- aĵojn
+- horo
+- ▁Carl
+- ▁suport
+- enaar
+- ▁indescribable
+- 啦
+- ▁중에
+- ▁marking
+- kisa
+- ▁dwellings
+- ▁offerings
+- ▁걔는
+- ▁yakoze
+- ▁musa
+- យ
+- ▁грамад
+- èxit
+- maha
+- ▁Está
+- ▁politic
+- wärts
+- ▁chord
+- ▁observance
+- waga
+- ▁celo
+- lieden
+- ▁capitalist
+- ▁چیزی
+- ▁hues
+- માં
+- 言
+- ▁obw
+- chio
+- ▁мой
+- ▁Anna
+- ▁gig
+- obten
+- şi
+- ထ
+- ▁توانم
+- rando
+- ▁ordinance
+- ▁catalogue
+- ถูก
+- ▁percival
+- ▁insurance
+- ▁arbitrary
+- ▁розных
+- gian
+- czę
+- ▁آب
+- umugore
+- raff
+- grat
+- ▁höher
+- ▁blindly
+- وێن
+- ▁truthful
+- taha
+- ▁häufig
+- น้ํา
+- ▁сур
+- tegereza
+- ▁친구들
+- ▁vitality
+- ალ
+- ▁einst
+- gling
+- сім
+- ▁fairest
+- ▁필요
+- ▁마음
+- ▁paternal
+- ▁vocation
+- ▁swine
+- ▁suda
+- ▁alternately
+- ▁leiden
+- ▁처음에
+- yobora
+- ▁einzu
+- вол
+- ੁ
+- ラ
+- 直
+- ற்
+- lijken
+- ▁weighing
+- ▁shelley
+- ▁jî
+- ▁zweiten
+- ▁encontrar
+- ▁continuar
+- ñas
+- ural
+- ▁sympathize
+- ▁nämlich
+- ▁broeder
+- ▁гына
+- ▁aperis
+- ▁carol
+- ▁soleil
+- ▁fij
+- ಗಳ
+- ▁joie
+- ▁пераважна
+- ▁prussia
+- ▁kalm
+- ▁reverie
+- ▁recollections
+- งาน
+- леп
+- mph
+- ▁bathing
+- ▁establishing
+- ڻ
+- 人口变化图示
+- ▁cache
+- ▁aspire
+- ▁fasten
+- ▁guineas
+- 달라
+- ▁sabre
+- ▁dispon
+- kong
+- ▁suff
+- ▁wooded
+- toned
+- حم
+- ▁voluntarily
+- ▁rails
+- cimiento
+- ▁chauffeur
+- ▁perseverance
+- ▁propositions
+- ▁ascribed
+- pflicht
+- ▁mold
+- ▁catal
+- ▁handelt
+- ▁langsam
+- ▁demonstrated
+- ▁Villa
+- wą
+- ▁Lebens
+- падобн
+- ▁zaken
+- եւ
+- 世
+- 场
+- ▁partners
+- ▁Buganda
+- ுள்ள
+- ▁manipula
+- ▁nook
+- ▁skate
+- ▁lookout
+- πα
+- ▁permanently
+- 政
+- petto
+- ▁Hö
+- poro
+- igihe
+- ▁campus
+- ▁precepts
+- اض
+- ▁calcul
+- ▁pursuits
+- ▁contrivance
+- ▁Pont
+- ▁fraction
+- ▁literatura
+- ▁spike
+- ▁цябе
+- ▁cornelia
+- lala
+- xic
+- ▁pana
+- kubwira
+- ▁ridges
+- dico
+- ▁학원
+- ▁bessie
+- вид
+- ▁verlangen
+- ▁primi
+- ▁eustace
+- ▁rebecca
+- ▁invincible
+- ▁besieged
+- 但
+- ▁straits
+- ▁ŝia
+- umuriro
+- carna
+- ▁soc
+- ▁fishermen
+- ▁believers
+- ▁apparel
+- ▁hearers
+- winkel
+- ẹ̀
+- ▁alphabet
+- ▁hodiaŭ
+- ▁través
+- ▁conegut
+- ▁elkaar
+- ▁sèrie
+- ▁ronald
+- qü
+- ▁applies
+- ▁grate
+- clothes
+- ▁gaol
+- zuba
+- uburyo
+- ▁clause
+- ▁politician
+- ▁pondered
+- ▁dunia
+- ▁excluded
+- ològic
+- liko
+- ▁Clara
+- ှ
+- 院
+- ▁chemistry
+- кий
+- incendi
+- ▁sangre
+- ▁bany
+- ▁spine
+- still
+- impr
+- essor
+- intérêt
+- ▁rampart
+- ▁refreshed
+- ▁voici
+- ▁musste
+- ▁judicious
+- nyumba
+- ▁instrumental
+- bui
+- tuba
+- ▁vaga
+- hay
+- ▁tochter
+- girira
+- ▁sparkled
+- ▁protestants
+- ▁présente
+- mén
+- latin
+- ▁insufficient
+- ключ
+- ikin
+- туры
+- ▁regional
+- ▁größten
+- ków
+- 沙
+- uomo
+- ▁Kong
+- ▁arbeitete
+- ▁arbitr
+- ىلى
+- گۈ
+- хы
+- ▁option
+- ▁deuce
+- ▁Var
+- ▁каза
+- ük
+- ▁rhoda
+- ▁secretari
+- ▁Every
+- ղ
+- 처
+- หรือ
+- ▁expansion
+- わ
+- ▁gemein
+- ▁trata
+- industrie
+- éné
+- ▁couleur
+- ▁subscribe
+- ▁exposition
+- ▁unlikely
+- ▁baroness
+- seits
+- utz
+- loup
+- irent
+- 건
+- ▁করেন
+- Do
+- ▁marcia
+- ▁Tanmateix
+- 设
+- ▁timu
+- ▁clang
+- ▁engagements
+- ▁Fle
+- usta
+- ▁promoted
+- ▁harmon
+- اوە
+- ▁나랑
+- ▁graciously
+- ohlen
+- ▁lächeln
+- 向
+- 학생
+- ▁linked
+- digen
+- ▁cultiv
+- nick
+- ▁trumpets
+- ▁dreamy
+- ▁Weg
+- ମ
+- ೇ
+- 首
+- ▁vele
+- ▁тӱ
+- gä
+- ▁travelers
+- ▁сучасн
+- erekeye
+- ▁nobler
+- ▁refers
+- ▁getan
+- ▁formally
+- ▁disseny
+- lú
+- ств
+- ▁filthy
+- дак
+- ▁pueden
+- ▁calculation
+- ▁prêt
+- tā
+- ▁madly
+- 루
+- ▁sniff
+- ▁excitedly
+- ▁Sankta
+- kleid
+- cumu
+- ▁panther
+- است
+- ▁Праўда
+- ▁ніколі
+- ▁glistening
+- ▁volcanic
+- ▁vastly
+- ▁manchester
+- ▁strut
+- issent
+- ▁oneself
+- ▁없고
+- ▁adore
+- ▁presentation
+- إِ
+- ▁Дл
+- incomprehensible
+- ▁aggressive
+- ▁völlig
+- ▁نہیں
+- ▁operating
+- aram
+- ▁کجا
+- bourne
+- ▁scrutin
+- ▁tape
+- hul
+- ▁smelt
+- ▁ouvert
+- ▁Catalunya
+- ▁freddie
+- psi
+- дара
+- 지도
+- ▁پو
+- ▁frankness
+- beḥ
+- հ
+- zga
+- 古
+- ▁abamu
+- ▁boundaries
+- ▁noah
+- ▁idee
+- ▁канцэрт
+- 日本
+- ffeɣ
+- تىن
+- saal
+- ▁manger
+- ▁attainment
+- ▁liever
+- ▁craving
+- miye
+- ▁consign
+- ▁gesto
+- blü
+- ບ
+- ହ
+- ▁tormented
+- ▁harrison
+- ନ
+- 美
+- smelling
+- ▁marx
+- ннем
+- ▁variant
+- ▁approbation
+- ▁beckoned
+- ▁animation
+- ▁zoudt
+- ▁deshalb
+- akoze
+- ▁povis
+- ▁sable
+- ▁humil
+- ▁yagiye
+- bida
+- ▁temptations
+- ▁wilhelm
+- ▁injurious
+- ▁trifles
+- ▁느낌
+- ▁julius
+- ▁blur
+- ▁sobie
+- trak
+- ▁gras
+- รู้
+- ▁dudley
+- ▁paradox
+- ▁believer
+- ▁requiring
+- dici
+- ganga
+- сар
+- ▁rasp
+- ▁expressly
+- ▁Cri
+- muri
+- ▁Трэба
+- ▁clasping
+- urg
+- ▁administrative
+- ▁ordeal
+- ▁چې
+- geschick
+- kanda
+- gold
+- ▁mliḥ
+- ▁Telefona
+- ▁breeches
+- ▁notamment
+- ▁segundo
+- ▁devoid
+- ▁Coll
+- ▁نام
+- ří
+- adh
+- ▁Londres
+- ▁evade
+- ▁whiskey
+- ▁barimo
+- ▁hyn
+- ▁papal
+- ▁fading
+- cito
+- riri
+- ▁injunction
+- ▁pharisee
+- 音
+- 程
+- expliqu
+- ▁holly
+- หน้า
+- ▁mewn
+- ▁studis
+- ▁physi
+- ▁cruise
+- ከ
+- ▁exig
+- ▁flown
+- ക്ക
+- няй
+- ijoro
+- ▁ferro
+- haupt
+- footed
+- automat
+- arsi
+- ვე
+- ▁snel
+- trained
+- ddin
+- ▁Además
+- ▁barracks
+- क्ष
+- ▁clamour
+- ліз
+- чная
+- verlo
+- жал
+- glu
+- 돌
+- ▁punctual
+- 情
+- ▁emerson
+- чаны
+- شار
+- ▁reconciliation
+- ▁scarecrow
+- stimmt
+- ▁tribune
+- ▁яра
+- ▁browning
+- ▁prohibition
+- ▁plata
+- vida
+- дзея
+- oedd
+- drà
+- ਗ
+- ▁scrupulous
+- ▁menys
+- ▁knitting
+- ▁pools
+- ▁welchem
+- ▁mlle
+- oux
+- ▁vervolgde
+- ▁versions
+- ▁선생님
+- ▁sacerdot
+- ங்கி
+- ުން
+- ▁verschiedene
+- ▁ببین
+- 창
+- ▁mosquito
+- ▁latch
+- existent
+- ▁Qual
+- ▁comuni
+- Й
+- nega
+- ▁allora
+- ància
+- ▁premature
+- 回
+- ▁yemma
+- ▁merriment
+- ▁apprehend
+- 系
+- ▁flavour
+- ကို
+- vri
+- schnitt
+- ▁Daniel
+- bà
+- ▁suitor
+- ▁راه
+- tawi
+- ▁wur
+- ▁impenetrable
+- ▁Day
+- greifen
+- கின்றன
+- ucci
+- ▁individuality
+- ▁پای
+- รับ
+- ▁experimental
+- ▁plow
+- ▁cloister
+- матэрыял
+- ▁Royal
+- ▁studierte
+- 둘
+- iji
+- ▁conrad
+- ▁Each
+- ▁splinter
+- ▁mortimer
+- ▁vell
+- ivamente
+- ▁Punkt
+- ▁indications
+- жат
+- ▁dike
+- ▁chopped
+- ▁universitat
+- ეს
+- ▁madman
+- accia
+- ▁dadurch
+- ▁projecting
+- ӹм
+- ▁warden
+- ▁mesmo
+- ደ
+- 패
+- 星
+- ▁verlaten
+- gerera
+- ▁denkt
+- people
+- கூட
+- 시에
+- ▁sentenced
+- ▁freundlich
+- ▁degradation
+- ▁항상
+- 队
+- ▁அவன்
+- ▁gottes
+- овыя
+- zani
+- ▁penance
+- illus
+- că
+- ▁paroles
+- ▁ashton
+- ▁lessen
+- ▁monkeys
+- ▁crowding
+- ▁decora
+- gukunda
+- ▁renown
+- fio
+- sohoka
+- ▁һүҙ
+- ▁imaginative
+- ▁کردم
+- ▁especialment
+- ▁blooming
+- rusha
+- ▁stretches
+- ▁basse
+- мир
+- ▁Ram
+- даецца
+- ▁funkcias
+- mmun
+- ура
+- ▁trinity
+- olla
+- ▁League
+- abaye
+- ۈپ
+- aardig
+- ▁geography
+- inspiring
+- ▁malpli
+- ▁Mul
+- ▁forsake
+- ▁verre
+- ▁شۇ
+- ▁bezig
+- ਬ
+- ▁glared
+- ▁Rück
+- ▁plunging
+- ▁revival
+- rious
+- ▁necessitat
+- holz
+- ▁гыч
+- ▁celebrate
+- ▁Monte
+- ▁diminish
+- städt
+- ▁Kagame
+- hov
+- wale
+- ▁apud
+- kva
+- ões
+- ▁verkoj
+- рем
+- ▁multajn
+- ▁дома
+- ▁omens
+- ▁banners
+- ▁recount
+- 특
+- 動
+- ▁jordan
+- kuza
+- лік
+- ▁democrat
+- соб
+- има
+- ▁expanded
+- ▁substitu
+- ▁Сам
+- ▁drawers
+- ▁ndirimbo
+- ▁krijgen
+- ▁shrubs
+- ▁entstand
+- ދެ
+- ▁sopra
+- phal
+- ▁engineers
+- experi
+- loven
+- ▁cassi
+- genze
+- ▁además
+- ▁nickname
+- ▁counsellor
+- ▁tempting
+- ▁colla
+- ▁stung
+- ouest
+- paka
+- lotte
+- зім
+- ▁devis
+- ણ
+- ▁하루
+- senge
+- यो
+- ▁augusta
+- quier
+- ▁lease
+- ▁verlegen
+- улы
+- ▁crouching
+- ▁explicit
+- ▁pretensions
+- ▁bityo
+- ிருக்கிற
+- kaga
+- idan
+- ▁inmates
+- stup
+- ന്
+- ىغا
+- ▁Angeles
+- ног
+- ▁pavilion
+- ▁싶다
+- mär
+- ▁slecht
+- ▁tangle
+- ▁lamentation
+- льных
+- ▁Sou
+- ▁reise
+- ▁llevar
+- häu
+- ▁currency
+- ڵێ
+- ▁Nga
+- ▁outlet
+- ▁selon
+- чкі
+- ▁malo
+- geschlagen
+- ▁паміж
+- ▁inshingano
+- ▁fenêtre
+- ibly
+- wys
+- ▁transmitted
+- ▁sandwich
+- ▁keith
+- лом
+- ▁savagely
+- kleide
+- ្រ
+- ▁separately
+- ▁warmer
+- ▁genannt
+- ▁arrayed
+- რე
+- ▁scrutiny
+- ▁spire
+- ▁clutching
+- ▁Deutschland
+- ▁verwendet
+- ▁abused
+- ৰা
+- ▁rupert
+- ▁completion
+- ▁medici
+- ▁trotted
+- ▁norfolk
+- acost
+- àtica
+- ▁Ost
+- ނީ
+- ▁vigilance
+- ▁paintings
+- ▁Second
+- ற்க
+- osamente
+- ▁заяв
+- treat
+- ▁astronomer
+- ▁inquisitive
+- ▁tanzania
+- ▁Weltkrieg
+- ▁mulher
+- uya
+- ▁Radio
+- ▁broth
+- tela
+- rwaho
+- ▁distorted
+- ▁impelled
+- ▁solicitude
+- дел
+- ەکەی
+- wyr
+- itation
+- ▁burgundy
+- ▁scare
+- ▁proteg
+- ▁defensa
+- аўся
+- ▁jura
+- ▁zeigte
+- ▁tanca
+- ▁palais
+- ங்கள
+- ▁Lea
+- ▁basta
+- ▁shipping
+- ▁barbe
+- ▁confide
+- ▁aún
+- ож
+- haki
+- ▁yarn
+- ▁turret
+- ▁імі
+- ongeye
+- för
+- ▁soothe
+- ▁gedachten
+- suchen
+- ▁fainting
+- gama
+- ▁twigs
+- hielt
+- nnek
+- risto
+- ▁bounty
+- acte
+- hira
+- Le
+- mung
+- ▁perils
+- 咁
+- tegereje
+- ▁enmity
+- ▁fürst
+- ▁Kabaka
+- ▁radiation
+- ▁Vic
+- ▁estable
+- ▁lendemain
+- Mu
+- ▁Kiu
+- kubita
+- ashoboraga
+- ▁Dabe
+- ▁perspiration
+- ▁lurking
+- ▁patted
+- لت
+- pica
+- เช
+- ▁dissent
+- اسی
+- іліся
+- ▁juno
+- образ
+- ▁depois
+- ▁dated
+- werp
+- ▁يَ
+- haka
+- ▁grape
+- fiel
+- ▁podem
+- ▁outro
+- shimye
+- ▁Mein
+- mibereho
+- 指
+- ▁niemals
+- ▁faci
+- simo
+- ▁vian
+- ▁indefinite
+- ▁calmness
+- schilder
+- ▁imagining
+- ▁poland
+- ▁locomotive
+- ▁ludicrous
+- 車
+- ▁corpo
+- kasi
+- ▁potato
+- ▁trotz
+- င်း
+- ções
+- ▁dolls
+- шәа
+- ▁cask
+- ▁comença
+- ▁hugo
+- ordinate
+- amaguru
+- ▁concord
+- ▁mortar
+- beau
+- ▁anchored
+- ▁possessor
+- צ
+- ▁доўга
+- ▁Texas
+- ▁gelegenheid
+- வில்லை
+- ować
+- ▁blacksmith
+- ▁wachten
+- ▁가지
+- ▁gloria
+- ▁comply
+- цтва
+- ▁politika
+- ▁ramble
+- ▁quickened
+- овай
+- ober
+- ▁lure
+- ▁garni
+- sirika
+- ▁돌아
+- ▁pug
+- ▁dingy
+- ▁гады
+- ▁contrive
+- пор
+- ▁sì
+- பு
+- ▁sogleich
+- ▁складана
+- ึ
+- ▁afin
+- ▁клуб
+- ▁fuse
+- ▁constrained
+- ζ
+- ▁Israel
+- ▁аўтар
+- ▁deutsche
+- attribu
+- ▁angla
+- ▁grassy
+- gelassen
+- ▁Zusammen
+- ▁dikwijls
+- ▁stairway
+- ▁vela
+- ñe
+- view
+- ▁concerts
+- ▁jeu
+- ਦ
+- 田
+- ▁warehouse
+- ▁cowardice
+- ▁precede
+- ▁portray
+- стой
+- ▁realizing
+- ▁mío
+- 끝
+- 关
+- ▁dinge
+- ▁Рас
+- ▁nku
+- ▁ambrose
+- jack
+- ▁нашы
+- ғын
+- لق
+- ▁và
+- ▁nibwo
+- piga
+- ▁bike
+- Ү
+- ▁indistinct
+- 空
+- 接
+- ▁prosecute
+- граф
+- ▁Mak
+- ▁strewn
+- ▁glide
+- resse
+- hilfe
+- неш
+- ▁braid
+- ▁algún
+- letter
+- gola
+- دىكى
+- 름
+- 쯤
+- 线
+- ▁disappoint
+- ▁newcomer
+- ▁White
+- ▁hamper
+- pone
+- ▁designated
+- ▁intimately
+- ▁Finalment
+- Uwiteka
+- ▁incredulous
+- ▁tendencies
+- ▁França
+- ▁ŝajn
+- ▁наша
+- bä
+- ▁opium
+- ▁Acḥal
+- ▁monotony
+- ▁vagabond
+- kkel
+- ▁numb
+- asso
+- ▁pensive
+- ▁Marie
+- 七
+- ▁tactics
+- ▁poisonous
+- كەن
+- ▁Blai
+- ▁dimple
+- itor
+- ▁posse
+- ▁montgomery
+- ▁insanity
+- ▁ruthless
+- ▁haute
+- gingen
+- ▁tradi
+- ▁medicina
+- éron
+- ▁ainda
+- ác
+- ▁ambulance
+- ▁propra
+- ▁jennie
+- ▁scor
+- ▁contemptuous
+- าว
+- terior
+- wyl
+- ▁gliding
+- ▁gushaka
+- ▁peculiarities
+- اخ
+- haltung
+- 佢
+- ▁custody
+- ▁працаваць
+- рыс
+- ▁stuk
+- ▁ntacyo
+- ▁wiping
+- ▁Wort
+- ▁semblait
+- ▁amber
+- ▁humain
+- ▁lurk
+- ੰ
+- ▁announcing
+- ▁بش
+- ▁preserving
+- ▁impending
+- ▁devenu
+- copi
+- ▁parapet
+- வன்
+- ▁Fred
+- funga
+- ▁inhuman
+- ▁fürchte
+- ▁relació
+- ▁уның
+- ыі
+- ▁realiza
+- ▁mortgage
+- ▁fonction
+- ambara
+- ▁falt
+- ять
+- mizi
+- ▁Kim
+- ▁eher
+- ▁complaining
+- ▁rasch
+- ▁distinctive
+- ▁curl
+- ları
+- ▁artikolo
+- ۇش
+- бан
+- wó
+- dî
+- ▁Bundes
+- ▁Sar
+- ▁katherine
+- 那
+- ▁இது
+- ▁visibly
+- 놓고
+- ▁commandment
+- ▁truce
+- valo
+- ▁heiress
+- ▁Про
+- hala
+- ▁drowning
+- ▁embark
+- 又
+- ▁Henry
+- ▁Antaŭ
+- De
+- ▁ыле
+- ▁michel
+- matu
+- ▁krom
+- ▁tablespoonful
+- ▁دیگر
+- ▁varios
+- crown
+- ▁Later
+- gebot
+- ▁Гэты
+- ▁தெ
+- ▁undergone
+- ំ
+- ▁છે
+- 꺼
+- 허
+- riot
+- ▁mariage
+- энні
+- vido
+- ▁freezing
+- ssian
+- ▁sundry
+- ▁profile
+- schneid
+- ▁vrienden
+- ▁units
+- acque
+- ▁heaved
+- лө
+- ます
+- ashoboka
+- ▁inadequate
+- ▁planes
+- ▁bezeichnet
+- ppa
+- ▁spre
+- вор
+- ▁Guim
+- ▁Road
+- 병
+- chain
+- ▁posible
+- ▁circul
+- 字
+- 它
+- ▁musk
+- ▁Merc
+- fina
+- ▁multiplied
+- kimu
+- łą
+- раж
+- ▁chunk
+- ▁null
+- ▁clarence
+- ้น
+- нэ
+- penda
+- estació
+- ▁kleiner
+- siya
+- ▁pedi
+- buc
+- ▁کرده
+- où
+- gemeinde
+- ▁squeak
+- ▁rhy
+- ▁tradicional
+- ▁Heute
+- ื่อ
+- ▁copied
+- ▁meddle
+- ▁stiffly
+- ▁Bürger
+- ▁yiwet
+- ▁ornamental
+- coul
+- ▁Universidad
+- ▁spreek
+- ▁gauche
+- ▁kuki
+- ▁Din
+- ▁treballar
+- ობ
+- ▁dotze
+- іль
+- bildung
+- ▁obliging
+- ▁gesprek
+- ▁czas
+- ▁адным
+- ▁werkelijk
+- ление
+- 信
+- Ç
+- ▁dadelijk
+- tele
+- ▁recollected
+- ▁performer
+- ▁moors
+- thorne
+- ەم
+- ▁Jim
+- ▁annually
+- ▁magical
+- ▁этом
+- ▁therewith
+- إ
+- 겼
+- ▁scène
+- ▁influential
+- ▁wünsche
+- ▁nko
+- ▁sloping
+- manĝ
+- ▁lambert
+- ▁enferm
+- anditse
+- ▁idly
+- ▁چىق
+- ▁maca
+- ▁Casa
+- ▁Ng
+- ▁ҡул
+- ▁glitter
+- ▁undone
+- ▁rendre
+- ▁minority
+- 英
+- ▁plausible
+- ▁bailey
+- ▁nivell
+- acció
+- 볼
+- ▁알고
+- ▁loosened
+- ▁helplessly
+- teilung
+- rion
+- ▁muslin
+- ული
+- hout
+- ▁raged
+- équi
+- ▁heresy
+- ▁territories
+- ання
+- ँ
+- ▁horrified
+- ကြ
+- dov
+- fuga
+- ▁willem
+- ▁mcc
+- ▁mab
+- ▁bolted
+- ▁début
+- ▁amahirwe
+- ارت
+- ▁aristotle
+- ▁schwester
+- ▁sparkle
+- ▁sandal
+- ▁befand
+- ▁inserted
+- ▁нейкі
+- 누
+- ▁inquest
+- ុ
+- ▁accompaniment
+- 月
+- ▁sweeter
+- ▁byron
+- ▁먹었
+- ▁moderna
+- ▁hond
+- yanga
+- ▁Dam
+- ▁bekend
+- ▁ekyo
+- ള
+- 發
+- ების
+- ▁Even
+- umuryango
+- ▁commandments
+- fremd
+- ɣan
+- ▁bemerkte
+- akozwe
+- حَ
+- ▁items
+- ▁volvió
+- imbwa
+- rath
+- ▁Ella
+- ▁prettiest
+- ▁geschehen
+- ▁grinding
+- guany
+- ▁significantly
+- fydd
+- ▁faintest
+- ▁okukola
+- ▁gedachte
+- feuille
+- ▁esperant
+- yek
+- լ
+- ▁trimmed
+- ▁frivolous
+- ▁hurrah
+- dab
+- படி
+- ▁dessert
+- ▁مق
+- ▁kennis
+- สุ
+- ▁dizzy
+- ▁rations
+- tula
+- ▁scourge
+- ▁mandat
+- dach
+- heiten
+- lto
+- ▁vorher
+- ▁كۆ
+- ▁rout
+- ▁مج
+- ▁opini
+- wendung
+- ▁сустрэ
+- ▁chamberlain
+- ▁citadel
+- fassung
+- ▁logo
+- ▁каго
+- ▁plau
+- ▁crocodile
+- 各
+- ▁entfernt
+- 受
+- ▁probablement
+- eeren
+- тэн
+- льні
+- getragen
+- ▁viejo
+- rur
+- ▁Gan
+- hibi
+- vella
+- ▁unlimited
+- ▁athletic
+- ▁ulysses
+- ▁schedule
+- ▁килеп
+- ▁flourishing
+- ▁wilkins
+- ▁fuego
+- ių
+- volvi
+- ël
+- ▁sometime
+- ▁basic
+- ▁뭐라
+- قَ
+- бры
+- ▁viatge
+- teile
+- ▁watson
+- ▁bidden
+- ▁않았
+- ▁applying
+- ▁grosse
+- alba
+- ▁personas
+- ▁bewilderment
+- ▁embassy
+- ▁jap
+- bah
+- ▁beneden
+- வீ
+- 넘
+- ahabwa
+- غل
+- ▁metallic
+- ▁tiptoe
+- ▁dealings
+- ▁peers
+- 졌
+- ▁candidates
+- аюцца
+- ▁thrash
+- гон
+- ▁jay
+- officier
+- ▁Igihe
+- ▁disputed
+- ▁invade
+- ▁Euro
+- ▁manufacturing
+- ▁taqbaylit
+- ▁elevator
+- ▁gehabt
+- ▁Rob
+- რო
+- ijimye
+- ▁saucer
+- ▁поз
+- ▁wren
+- ▁breve
+- ▁ulac
+- ząc
+- rigen
+- ▁hypocrite
+- ▁sœur
+- ▁appease
+- ▁такім
+- ▁explosive
+- ِّ
+- shon
+- πε
+- emble
+- ▁louisa
+- ▁Groß
+- filled
+- ▁mingle
+- ើ
+- ▁Apunta
+- ▁itatu
+- ▁Pic
+- ilio
+- ▁heather
+- ème
+- ▁originated
+- ▁suivi
+- ▁spoils
+- ả
+- ని
+- ▁naskiĝis
+- სი
+- сыҡ
+- ▁gruff
+- akte
+- ▁ushaka
+- ▁rostro
+- 大学
+- ▁symbols
+- ▁concentra
+- ▁difficile
+- ọ̀
+- ▁Només
+- ▁Tiel
+- ▁Marta
+- ▁타고
+- 라서
+- ▁sierra
+- ṛen
+- ▁노래
+- Bo
+- ▁occupants
+- ▁거고
+- ▁strips
+- ▁freeze
+- ▁taint
+- لان
+- permes
+- bonne
+- mauer
+- ▁overtook
+- ▁Johann
+- ▁kitundu
+- ▁perverse
+- ена
+- gewicht
+- ▁adoption
+- ▁leslie
+- ▁amiss
+- ▁бил
+- ▁pasado
+- ▁mondmilito
+- ▁derrière
+- ithi
+- ▁Francisco
+- عمل
+- ىيە
+- opa
+- ▁catalan
+- ▁dedicat
+- ivu
+- ▁juu
+- ▁moaned
+- লো
+- prü
+- ▁admira
+- ▁Museum
+- ړ
+- ▁quotation
+- ▁fervent
+- ▁façon
+- ▁iliyo
+- ▁адкры
+- ▁husk
+- гыл
+- ▁partoj
+- ▁Andre
+- funz
+- ecido
+- ▁һуң
+- ▁carleton
+- ▁eenig
+- ▁Alguns
+- 없
+- ▁bane
+- ▁muziko
+- ubundi
+- ▁verbal
+- ▁obstinacy
+- داد
+- altre
+- ▁suelo
+- ▁rooted
+- mubwira
+- ▁foreseen
+- ▁actualment
+- ▁hannibal
+- He
+- ▁consort
+- ▁reckoning
+- ▁nkuko
+- ▁raleigh
+- іст
+- ▁appre
+- ▁loka
+- regul
+- ▁Umuntu
+- fatwa
+- ▁trance
+- ▁musique
+- ▁minstrel
+- ▁accents
+- ▁булды
+- ▁temas
+- ▁Long
+- ▁llegó
+- ▁йөрө
+- anca
+- ▁Biz
+- atic
+- president
+- ▁cambr
+- ▁выгляд
+- тыра
+- ן
+- ளை
+- ▁centra
+- ▁frogs
+- گرد
+- ▁abrid
+- ▁Dank
+- про
+- ▁trin
+- ▁Ame
+- ċ
+- คิด
+- ▁Queralt
+- 였어
+- ▁bastante
+- ▁юл
+- ▁blindness
+- ләп
+- urbo
+- ▁shopping
+- tě
+- rigu
+- กร
+- ▁Adam
+- ▁ritual
+- ▁Cesc
+- spruch
+- ڕۆ
+- 清
+- 至
+- 圣
+- thron
+- ▁свеце
+- werken
+- ้าง
+- ど
+- ▁verheiratet
+- ▁tournament
+- ▁সং
+- zil
+- ▁saturn
+- koma
+- bericht
+- ▁이번
+- ▁wedge
+- ▁sequence
+- itse
+- ▁violation
+- ▁Katika
+- ▁Rose
+- ▁halve
+- seitig
+- ▁shah
+- ▁retrouve
+- ▁renowned
+- ▁варта
+- ระ
+- ▁nobly
+- ▁gentiles
+- ▁luna
+- ▁moaning
+- ▁malnova
+- ▁کنیم
+- ▁martyrs
+- hani
+- stanti
+- ▁minsi
+- ▁склад
+- ో
+- ▁sash
+- ▁extort
+- ▁committing
+- ▁weêr
+- cale
+- ▁Ngo
+- Ka
+- ▁heretofore
+- ▁utility
+- 답
+- ▁unlocked
+- geschrieben
+- ▁lonesome
+- ▁보통
+- loop
+- өө
+- ▁curves
+- ৃ
+- âge
+- lema
+- ▁promenade
+- ▁lorsque
+- ▁sketches
+- ▁crashing
+- ▁carelessness
+- yeḍ
+- meli
+- ▁purest
+- itzat
+- ▁exploded
+- 기는
+- ▁scientists
+- К
+- cú
+- ▁ambush
+- كل
+- 頭
+- ▁braucht
+- ▁hivyo
+- رگ
+- cchia
+- ▁оста
+- ▁grunted
+- ▁ayuda
+- ziya
+- ▁temperate
+- ▁менавіта
+- ▁emancipation
+- ▁cheval
+- ▁virino
+- ▁sling
+- ▁Gebiet
+- ▁denove
+- ▁Shu
+- ება
+- ▁hissing
+- chim
+- gesta
+- haza
+- 료
+- 읽
+- وێ
+- калі
+- ▁ninguna
+- ▁deinen
+- ىمىز
+- kuye
+- ▁malgré
+- ▁phineas
+- ▁stupidity
+- ▁Sea
+- zona
+- ▁frolic
+- ▁mercat
+- стры
+- наз
+- ▁contention
+- clus
+- ▁oysters
+- ▁crucifix
+- ▁Ingingo
+- ▁trabaja
+- ▁lens
+- ▁triangle
+- ▁되잖아
+- ▁enhance
+- ▁asunder
+- ▁mentioning
+- ▁Бы
+- pfel
+- ▁silencio
+- ▁ĉio
+- ையில்
+- ▁schließlich
+- ▁Kap
+- zej
+- ▁befall
+- ▁ئۆز
+- న్
+- ▁Flug
+- ▁profunda
+- ▁arabian
+- ▁Viele
+- 經
+- ▁elinor
+- compl
+- kami
+- accelera
+- yeɣ
+- ▁doble
+- ▁siajn
+- зор
+- ▁Gla
+- ▁anticipate
+- ư
+- ▁girt
+- park
+- ▁пачалі
+- Ю
+- ▁एक
+- ▁alder
+- ▁خوش
+- ▁кож
+- бег
+- 법
+- 朝
+- antiga
+- ▁fram
+- bug
+- tesse
+- ▁helpful
+- ▁года
+- ▁departing
+- ▁judas
+- ▁mujeres
+- ▁disobey
+- ▁salir
+- ▁titolo
+- ▁Frida
+- ▁purposely
+- àries
+- ▁dulce
+- ▁gemeen
+- ▁trailing
+- ▁паля
+- armée
+- ▁groaning
+- ஷ
+- deux
+- ▁savour
+- ▁trader
+- ▁পরি
+- ▁azul
+- ▁fusil
+- fair
+- ▁gunner
+- hair
+- ▁ddala
+- ▁zeus
+- ▁dividing
+- ▁uplifted
+- 이잖아
+- ▁involves
+- ▁vamos
+- Italia
+- gebilde
+- ẹ́
+- ▁arithmetic
+- ▁Kel
+- 称
+- দে
+- mikorere
+- போது
+- lof
+- ▁streamed
+- dź
+- ▁blest
+- ▁Party
+- ▁Cur
+- घ
+- 長
+- ▁lässt
+- ▁whereabouts
+- ▁змен
+- ▁hetgeen
+- при
+- ▁celebration
+- ▁elementary
+- ▁anderer
+- リ
+- ▁இரு
+- ▁doubtfully
+- ▁legislative
+- ▁webster
+- ▁duque
+- ▁estoy
+- ▁miner
+- ▁belgium
+- ▁sinned
+- ▁краіне
+- ▁otter
+- ▁bedoel
+- সি
+- ▁overtaken
+- ▁준비
+- 福
+- ▁smothered
+- ▁sicily
+- ▁집에서
+- ▁thrusting
+- bitabo
+- lícula
+- 住
+- ▁павінна
+- ち
+- ▁squaw
+- ▁universities
+- ▁هستند
+- ▁needles
+- ▁plastic
+- ண்ண
+- ości
+- ▁Burg
+- ▁Gli
+- Li
+- ▁blocked
+- ▁fac
+- 각
+- 충
+- ماس
+- essi
+- тин
+- ▁aspirations
+- ▁resembles
+- рост
+- ▁greed
+- ▁mačči
+- ▁twig
+- үк
+- kten
+- budde
+- ņ
+- 國
+- ▁hovering
+- ▁diciendo
+- чнага
+- ▁stunden
+- ▁turban
+- ამ
+- ▁considerate
+- tane
+- tegeka
+- programm
+- ▁goud
+- ▁poeta
+- guar
+- ▁chasm
+- rable
+- yambi
+- ▁stabil
+- good
+- ▁macdonald
+- 赛
+- intérieur
+- ริ
+- ση
+- ▁traduc
+- ▁slaap
+- ▁Also
+- ▁ўсім
+- ània
+- ▁enigma
+- دەك
+- 희
+- ▁Universität
+- ▁miserably
+- niem
+- pò
+- ▁keer
+- ▁scarf
+- ▁gewa
+- крыты
+- quil
+- 씨
+- ▁improving
+- ▁conclusive
+- ▁henrietta
+- ▁desperation
+- ▁buddha
+- ▁constrain
+- ▁chilly
+- rú
+- рус
+- ▁mcg
+- ▁시간이
+- ▁jako
+- ▁challenged
+- ▁grating
+- ▁ҡайт
+- ▁garçon
+- ▁받아
+- ▁spouse
+- ography
+- quia
+- ▁бес
+- ▁kuruta
+- шкы
+- ▁inconsistent
+- สามารถ
+- ▁alps
+- igno
+- banye
+- ▁ewig
+- ▁alternate
+- interior
+- ▁جم
+- ไม่ได้
+- ▁shipwreck
+- génér
+- ավ
+- ▁apren
+- ▁verschiedenen
+- ▁eph
+- ِهِ
+- ▁attest
+- ارد
+- ▁pequeño
+- ▁асабліва
+- ▁metropolis
+- ▁cowboy
+- Alba
+- gevallen
+- ▁पर
+- ▁construir
+- нцы
+- స్
+- ▁calculations
+- ▁darwin
+- ▁fais
+- ▁attractions
+- juri
+- ▁zeigt
+- ekonomi
+- ▁gouvern
+- ijn
+- ▁ugomba
+- ▁frisch
+- 更
+- 봐야
+- szedł
+- ▁уны
+- ▁zabo
+- ravi
+- ەیە
+- ূ
+- ▁тӱҥал
+- ▁Ubwo
+- buck
+- ▁duniani
+- ▁Brasil
+- ▁அது
+- чно
+- ▁christopher
+- ▁inquisitor
+- sanct
+- ▁shafts
+- ▁Чаму
+- òric
+- ▁Mai
+- ▁Lluc
+- тик
+- ▁Europäische
+- 판
+- ▁ontmoet
+- ▁blackened
+- opera
+- ▁Region
+- ▁harper
+- ▁ҡал
+- 리는
+- führ
+- ▁niko
+- leute
+- zaj
+- ònic
+- ▁rebellious
+- ▁Madame
+- 是一
+- ▁dicta
+- ▁sheath
+- ▁значна
+- ▁bildet
+- missió
+- ▁policemen
+- ▁deceit
+- ▁grumbled
+- ▁phases
+- ▁advantageous
+- лары
+- ள்ள
+- әләр
+- bond
+- ▁grec
+- brado
+- ▁апошні
+- ▁quarry
+- ▁glaubte
+- اي
+- ยัง
+- ▁qualche
+- ▁chama
+- ິ
+- މު
+- ▁intermediate
+- ▁kiun
+- ▁dubious
+- ▁oyster
+- ▁Kio
+- 먹
+- shaven
+- ▁abstraction
+- üsse
+- ▁glories
+- ▁exultation
+- ▁enkele
+- கா
+- ▁hinauf
+- ▁revolve
+- ▁боку
+- ▁buscar
+- ▁kinsman
+- ▁cristo
+- ▁hoard
+- ▁denounced
+- ▁veer
+- ▁multiply
+- 쁘
+- ▁collision
+- ▁Brown
+- ▁gloomily
+- insel
+- ɣli
+- นะ
+- ▁hohe
+- ▁astronomy
+- ▁erkennen
+- ▁ئادەم
+- ▁hindu
+- ▁banna
+- ciąg
+- ய்
+- ▁murderous
+- ość
+- ▁کل
+- ▁installed
+- ▁admitting
+- ▁downwards
+- ர்கள்
+- izuba
+- ▁swoon
+- 죠
+- ▁concentrate
+- elyk
+- ▁peix
+- ▁doris
+- ▁பய
+- пач
+- ঠ
+- beard
+- ▁electrical
+- pina
+- ▁desenvolupa
+- ▁minim
+- ▁факт
+- ▁More
+- вся
+- ▁Хо
+- ▁rivière
+- ▁degraded
+- ▁voulez
+- ▁industries
+- ▁peeping
+- ிலே
+- ურ
+- ▁neniu
+- ception
+- ▁جن
+- ніць
+- ▁bicicleta
+- ▁disinterested
+- ▁besought
+- ▁bamboo
+- ▁Chris
+- ▁sibyl
+- ▁fase
+- sooka
+- ▁китте
+- ▁mbona
+- ▁detest
+- ▁indulgent
+- ▁candidat
+- ору
+- Hug
+- ▁damaged
+- ▁inconvenient
+- 골
+- ▁measuring
+- ▁flanders
+- ▁furnishing
+- ▁év
+- ▁scruple
+- ▁працу
+- ▁frenchmen
+- ▁baseball
+- ▁rochester
+- ▁precept
+- рий
+- ▁parisian
+- xen
+- ▁delirium
+- ▁lowly
+- ۇن
+- ▁арган
+- ніц
+- үп
+- ▁disarm
+- lew
+- ▁laisser
+- ப்பி
+- ӷ
+- ▁commonwealth
+- ▁volte
+- ▁mastery
+- hän
+- ▁Estats
+- ▁presumed
+- drum
+- ▁tweede
+- េ
+- ұ
+- គ
+- 神
+- tiri
+- rolle
+- ▁valour
+- ▁venis
+- เต
+- ▁돈이
+- سَ
+- ▁lashed
+- 美国
+- 하긴
+- 조
+- ▁impudent
+- ▁chilled
+- ▁زیاد
+- ▁serio
+- ▁나이
+- hud
+- dè
+- ▁propi
+- дың
+- ▁열심히
+- 老
+- ▁civilian
+- ▁flattery
+- рав
+- ▁puffed
+- ▁маюць
+- buru
+- après
+- ▁aliye
+- ▁bleek
+- ▁кап
+- fähig
+- ▁Jenoside
+- ▁hastings
+- ▁моцна
+- frau
+- ▁casually
+- ▁ultra
+- ▁groans
+- ▁celia
+- yemo
+- ▁remu
+- Ni
+- ▁звычайна
+- ▁gravaj
+- ▁niyo
+- zeiten
+- ▁어쨌든
+- ▁Gor
+- ḥen
+- ▁shuddering
+- ▁voulait
+- zustand
+- руч
+- possession
+- who
+- ▁flooded
+- ▁informació
+- commissioned
+- ▁audacity
+- ▁dezelfde
+- ▁dürfen
+- breed
+- ▁Another
+- ▁numva
+- ▁zweifel
+- ▁trampled
+- hiri
+- ▁rudely
+- ▁steering
+- ▁우리는
+- ▁flask
+- ▁peculiarity
+- ▁مِنْ
+- ▁emphasize
+- ▁crescent
+- ▁línia
+- ▁vrouwen
+- bearing
+- ▁planting
+- ▁Lliɣ
+- Po
+- ర్
+- ▁chateau
+- ▁куды
+- gica
+- ލު
+- kho
+- ▁Several
+- ▁малады
+- ▁працэс
+- ▁overlook
+- ▁Como
+- mead
+- czył
+- ▁nursed
+- ▁sep
+- ▁wirst
+- ▁пеш
+- zabi
+- ▁relapse
+- 未
+- 维
+- komoka
+- ۇر
+- дор
+- ▁важна
+- 한다고
+- ▁brigand
+- ▁trump
+- guda
+- ▁твар
+- En
+- 싫
+- ▁soaked
+- ▁segona
+- ▁català
+- ▁شود
+- ▁susy
+- ▁эле
+- ▁juna
+- ▁flog
+- 필
+- 联
+- ▁mult
+- ▁sown
+- ▁apprehended
+- remos
+- feɣ
+- ▁signified
+- ▁옛날에
+- ގ
+- рып
+- ▁ŝin
+- ތި
+- ▁kubikora
+- ▁montagne
+- ▁аднаго
+- ▁farmhouse
+- ▁걸로
+- ▁verwacht
+- сма
+- ▁wame
+- ▁gambler
+- ▁цалкам
+- ッ
+- ▁Deg
+- boneka
+- rindwi
+- ▁Xavi
+- ▁destroyer
+- ▁celebr
+- ▁doubting
+- ҩ
+- сал
+- mbuto
+- ▁pecksniff
+- စ်
+- ▁disponible
+- ▁registered
+- ▁semblance
+- ▁felicity
+- fab
+- ▁Folge
+- ▁ernstig
+- ush
+- ▁erhob
+- ▁doña
+- صو
+- ▁digest
+- kku
+- ▁enemigo
+- 지고
+- ▁hypo
+- ▁forbidding
+- ▁выбар
+- ▁gev
+- ▁pacing
+- ▁hanyuma
+- ▁jugar
+- ▁pawn
+- ▁lashes
+- ▁gout
+- ▁situada
+- ▁Gebäude
+- ▁milady
+- ▁carving
+- ޅު
+- ▁malignant
+- ▁Hon
+- mori
+- oche
+- ▁parades
+- ▁Up
+- genzura
+- ▁bagiye
+- ▁زد
+- ▁konden
+- ▁pervade
+- ▁aforesaid
+- 快
+- ました
+- ▁Vater
+- ▁wildest
+- ▁junk
+- 對
+- 元
+- ▁forbear
+- ogne
+- ▁Cli
+- lived
+- ▁advancement
+- iloj
+- ▁арт
+- 림
+- ▁workshop
+- ▁stolz
+- ▁dichter
+- ▁거를
+- ▁achievements
+- ▁lliure
+- зер
+- ▁troupe
+- ▁wrestle
+- ▁Аб
+- ▁spectre
+- ▁gallows
+- ▁retort
+- ▁стане
+- ▁thuis
+- ▁ensure
+- ▁кажа
+- hausen
+- ▁jina
+- ▁censor
+- 点
+- ვა
+- ▁soort
+- ▁loĝas
+- ▁galloping
+- ▁echt
+- 몸
+- ▁delaware
+- ▁menacing
+- ▁Platz
+- wirtschaft
+- ▁lorsqu
+- بێت
+- ▁serivisi
+- 함
+- ▁lumière
+- ▁moderation
+- ▁harrow
+- ▁아까
+- ahu
+- ▁reservation
+- ▁paler
+- аст
+- ▁paroli
+- ▁teqbaylit
+- punct
+- ▁بى
+- ▁subsistence
+- оўка
+- ▁Danach
+- ▁apareix
+- ▁suggesting
+- ▁concealing
+- ▁combina
+- गा
+- ▁onions
+- วน
+- ▁distressing
+- ▁cloudy
+- avano
+- ▁гісторыі
+- ▁mushroom
+- 放
+- umvikana
+- ▁galleries
+- ▁voit
+- schütt
+- ▁sentry
+- ▁abject
+- ▁raced
+- ▁Germanio
+- рук
+- ಅ
+- ▁disquiet
+- 틀
+- ▁따른
+- ▁hoffnung
+- 改
+- ▁schmerz
+- ▁nuestros
+- ▁pourtant
+- ingenzi
+- ▁hoed
+- teeka
+- 하다가
+- глед
+- 形
+- ▁encamped
+- 品
+- step
+- ▁말이
+- ▁continuance
+- ▁tease
+- ▁greg
+- ▁measurement
+- ▁Asia
+- ▁bibiri
+- ▁bourgeois
+- ნი
+- ア
+- ▁introducing
+- ▁según
+- ▁bitandukanye
+- вак
+- оро
+- ▁unrest
+- ▁pleur
+- ▁stow
+- ▁lanza
+- gehouden
+- ▁хаце
+- 끼
+- ▁식으로
+- ▁brusque
+- ▁succès
+- ▁strata
+- ▁salud
+- เห
+- ▁وقتی
+- ▁outset
+- ▁volunteer
+- fighter
+- ▁accession
+- ທ
+- ▁oblivion
+- ▁somber
+- ▁dante
+- ▁비싸
+- lova
+- ▁shroud
+- ▁demonstrate
+- iau
+- —
+- ▁muzika
+- ▁softness
+- ▁artifice
+- ▁winner
+- ▁pili
+- markt
+- ▁fulfilment
+- Č
+- ▁untouched
+- ▁enchantment
+- ▁isoko
+- ▁inscribed
+- हि
+- ▁scraps
+- ▁vorüber
+- ▁bimwe
+- ▁scientist
+- ▁Units
+- ތް
+- ▁zelven
+- ▁edzino
+- ▁Music
+- wirkung
+- ▁mannen
+- ▁escrito
+- banda
+- ണ
+- person
+- ▁lengua
+- ▁faust
+- ▁patronage
+- ▁denying
+- ▁sharing
+- ▁squeezed
+- δι
+- ▁reminiscence
+- ▁pubblic
+- ▁করা
+- ▁nzira
+- ▁mattress
+- hôtel
+- ▁ornamented
+- ▁heaving
+- ▁producció
+- ▁frase
+- ▁malferm
+- ▁grasses
+- ▁escriure
+- ▁epoko
+- ▁کام
+- final
+- ▁hatchet
+- ▁loknomo
+- ▁wilaya
+- ▁software
+- です
+- hiel
+- ▁pathos
+- ▁sculptor
+- ▁gasping
+- ▁такіх
+- bishobora
+- ▁speculative
+- ▁multiple
+- ▁minimum
+- َّا
+- تب
+- ▁vex
+- ▁беларускі
+- gewand
+- 将
+- ▁пачу
+- alist
+- ẓa
+- سپ
+- ▁biff
+- ▁Mann
+- iii
+- ▁geographical
+- 卡
+- 类
+- atti
+- ▁بالا
+- 때
+- akcept
+- prise
+- ▁wollten
+- ▁pauper
+- ▁midday
+- ▁sowohl
+- ▁carthage
+- ▁Ndashaka
+- ▁envious
+- ▁mynd
+- ▁expired
+- ▁bouche
+- mika
+- ▁deiner
+- 治
+- abakobwa
+- ▁봤는데
+- ▁ایک
+- ▁seien
+- ▁outlook
+- ▁tremor
+- ▁guild
+- ▁speculations
+- mię
+- ▁disposa
+- ▁treballa
+- ▁qualsevol
+- ▁nuisance
+- ▁sammy
+- ▁digital
+- ▁schrik
+- ▁cupid
+- ▁الس
+- wagen
+- ▁wistfully
+- ▁surpassed
+- leq
+- ▁resided
+- ▁vermell
+- ▁crawford
+- ▁deserving
+- ▁Organ
+- хар
+- ▁valencia
+- seeing
+- qqar
+- torial
+- ▁tabernacle
+- 향
+- އަށް
+- bruch
+- ▁parable
+- ▁trigger
+- ▁calvin
+- ɣeɣ
+- ▁suffi
+- ▁afero
+- 겠네
+- ޒ
+- ▁metropolitan
+- ▁даволі
+- preis
+- ▁genoegen
+- ▁witte
+- ▁petti
+- தோ
+- ▁provincia
+- ▁walio
+- ▁möge
+- 果
+- ▁cowperwood
+- え
+- ▁methinks
+- ▁besondere
+- jji
+- ▁drunkard
+- laced
+- ▁brotherhood
+- ▁masterpiece
+- abadde
+- ▁crucified
+- ▁unnoticed
+- ▁свайго
+- 풀
+- 還
+- urwego
+- ▁principio
+- amma
+- ▁rechten
+- ▁velocity
+- kich
+- ▁mkoa
+- conscious
+- ▁fehlt
+- shika
+- ству
+- ▁conduit
+- ▁scruples
+- ▁unmoved
+- ▁clavering
+- gendera
+- twaye
+- rebbe
+- ▁assemble
+- ▁tercer
+- éternel
+- ▁territorial
+- ikan
+- ▁Nico
+- ▁defendant
+- ▁maximum
+- fleck
+- ▁shred
+- هَا
+- маўля
+- ▁psych
+- ▁나서
+- şa
+- ▁tacit
+- zé
+- خواه
+- ▁billiard
+- ▁cualquier
+- ▁знаходзіцца
+- ▁skip
+- они
+- igny
+- ▁بخ
+- ▁나온
+- mètre
+- ▁poner
+- ▁harmonious
+- ▁mji
+- ▁laborers
+- ▁pele
+- ▁tending
+- ▁lombard
+- ▁Kion
+- ▁gehörte
+- ské
+- ▁prefect
+- ▁сказал
+- kusi
+- ▁merkte
+- дзён
+- ঁ
+- ▁squeeze
+- ულ
+- hunga
+- ▁venen
+- ▁خدا
+- ▁petita
+- ▁auditor
+- ҡай
+- ancienne
+- umukara
+- ▁لطفاً
+- ની
+- ▁이름
+- ▁pup
+- ności
+- லாம்
+- بي
+- ▁joking
+- ▁periodical
+- ▁dram
+- ▁exalt
+- oreille
+- ▁sprinkled
+- ▁breton
+- ▁parchment
+- ▁theoretical
+- ▁huu
+- ▁seor
+- ▁reliable
+- جە
+- ▁nutri
+- ▁katie
+- ▁whisk
+- ޅ
+- ▁coon
+- ẹ
+- 语
+- ▁fernando
+- steig
+- ▁repast
+- ▁danes
+- igné
+- ▁செய்ய
+- ▁ethical
+- ▁escorted
+- ؤ
+- ▁mysteriously
+- ▁abundantly
+- ▁latent
+- وج
+- aventure
+- ▁சொல்ல
+- voir
+- ▁bituma
+- ▁judy
+- institut
+- umber
+- ▁sift
+- ▁쪼끔
+- ▁moeite
+- ▁procés
+- ▁전화
+- atori
+- éclair
+- ▁normandy
+- кын
+- іза
+- ▁Angel
+- ▁tai
+- ▁connecting
+- ຕ
+- ់
+- ▁llei
+- ▁nombreuses
+- ▁farthest
+- ▁영어
+- ▁implore
+- ▁pesar
+- ▁soothed
+- ▁Russia
+- ▁miseries
+- ▁vooral
+- ▁ошо
+- бас
+- ▁выраша
+- ▁tolerable
+- honderd
+- ▁heritage
+- album
+- ▁chercher
+- ▁Nuko
+- ▁printer
+- ▁bravo
+- ▁quail
+- zul
+- shinda
+- ▁spelling
+- weil
+- ُّ
+- ouvre
+- เจ
+- dün
+- ▁adjourn
+- hagarara
+- ▁zwaar
+- ▁context
+- ▁میتوان
+- maaso
+- umwuka
+- thia
+- رک
+- という
+- ▁blanco
+- ▁marcel
+- ▁slit
+- Tw
+- ▁arched
+- ▁principe
+- нии
+- '14'
+- ▁scroll
+- ބ
+- ▁zeigen
+- رِ
+- ▁positiv
+- ▁devise
+- ▁nameless
+- ▁imploring
+- ▁sarcasm
+- ouvrage
+- тель
+- ▁nueve
+- ▁impost
+- uju
+- moor
+- ▁siku
+- ▁koning
+- عد
+- ▁peur
+- ▁archi
+- gänge
+- льнага
+- ▁Maɣef
+- ▁bonaparte
+- ▁unceasing
+- igice
+- ▁розныя
+- ▁parlar
+- ▁كەل
+- ▁bagomba
+- ▁roadside
+- ▁bates
+- ▁vanquished
+- ▁bordered
+- ▁worrying
+- душ
+- ▁caliph
+- ▁نیاز
+- ▁illumination
+- ▁Multaj
+- ▁wrinkle
+- ▁chained
+- ▁indoors
+- төр
+- ▁Doctor
+- ▁outdoor
+- ▁mingi
+- овых
+- ▁whine
+- ▁bleu
+- ▁beginnen
+- ▁Tout
+- ▁benign
+- ▁rabbi
+- ▁unfinished
+- mero
+- drawn
+- nshuro
+- vojaĝ
+- ंग
+- ▁ilikuwa
+- ▁adelaide
+- mbling
+- eṛṛ
+- ▁hoffe
+- ▁equator
+- ▁differed
+- pfen
+- ▁Today
+- ▁exclaiming
+- ▁kamera
+- fourth
+- ▁gewann
+- ▁beaming
+- ardo
+- หม
+- хам
+- 白
+- ▁coyote
+- 产
+- піць
+- ▁collapse
+- ▁zuerst
+- laag
+- idio
+- ▁thur
+- يل
+- lil
+- kanga
+- ▁correction
+- ▁flickering
+- ▁coberta
+- ▁cairo
+- ▁plena
+- ▁First
+- ▁Дык
+- ▁piedra
+- oxide
+- ▁bwanjye
+- fugi
+- misunderstanding
+- ▁hoisted
+- ▁nourishment
+- ▁volunteered
+- ▁fleeting
+- ▁reciproc
+- 车
+- inzego
+- апа
+- knik
+- ▁gemeinsam
+- ▁så
+- ▁passport
+- мыш
+- ▁discontented
+- ▁traveler
+- ▁gutter
+- ▁bugle
+- ▁llor
+- ▁mondiale
+- 草
+- ပြ
+- ▁Budapeŝto
+- ▁drowsy
+- ▁folgenden
+- ▁imported
+- ▁kutime
+- ▁cortés
+- ▁Politik
+- ▁partnership
+- ▁sloop
+- ▁presto
+- ignon
+- ▁stevens
+- ▁livery
+- მე
+- ▁wenige
+- ▁gewöhnlich
+- め
+- ▁ibitekerezo
+- ▁sporting
+- braun
+- ▁말을
+- ▁hump
+- справ
+- ▁powdered
+- Ŝ
+- ▁organiza
+- ▁google
+- ▁sasa
+- ▁courier
+- ▁blended
+- ▁dye
+- نان
+- ▁cursing
+- ▁gehoord
+- ▁이랬
+- ատ
+- ▁gratefully
+- ▁때도
+- ▁Lang
+- ingle
+- ▁contemptuously
+- 지가
+- riel
+- ploy
+- жә
+- wahl
+- kompon
+- ▁discredit
+- ▁زندگی
+- ▁restricted
+- ▁slum
+- শা
+- ▁womanhood
+- ▁Bereich
+- amahanga
+- ▁இருக்க
+- ድ
+- ▁recompense
+- ▁harshly
+- ▁konsider
+- ▁partridge
+- ▁royalty
+- abanyarwanda
+- ándole
+- ▁ляс
+- ▁recalling
+- ▁aircraft
+- 第
+- ▁ushered
+- кыр
+- ▁alcun
+- ▁Francio
+- gawa
+- ُونَ
+- ▁lieben
+- зяць
+- ડ
+- 如
+- ▁Während
+- ▁seclusion
+- ونه
+- ванні
+- ▁anfang
+- ziba
+- ▁Cra
+- ▁lawless
+- чыў
+- ▁bwawe
+- ▁yaje
+- चे
+- zg
+- ▁persecuted
+- ▁stile
+- ▁pebbles
+- simba
+- based
+- uruzi
+- ption
+- ault
+- ▁satire
+- ▁Council
+- ठ
+- ▁بھی
+- 父
+- ▁practicable
+- garden
+- ▁ferme
+- tugomba
+- ▁benedict
+- sezera
+- ▁willoughby
+- ▁infanoj
+- toki
+- овая
+- ▁minuto
+- ▁lodger
+- fh
+- nije
+- геҙ
+- ▁herab
+- ▁Buli
+- ▁averted
+- ▁waltz
+- ▁noticeable
+- ▁strangest
+- Á
+- ▁кажу
+- fälle
+- ून
+- ▁dissatisfied
+- ▁sled
+- ▁villagers
+- illard
+- ▁Nza
+- ▁troben
+- ыуы
+- ▁heartless
+- ई
+- ▁unaccountable
+- ▁졸업
+- ▁հա
+- ▁señal
+- ▁marido
+- giri
+- ▁partido
+- алася
+- eloos
+- ▁경우
+- ▁endanger
+- әл
+- ▁stifled
+- ▁festa
+- ▁distraction
+- iol
+- ажә
+- ▁prediction
+- ▁respira
+- ▁buenos
+- ấ
+- ▁zuletzt
+- ელ
+- ▁сярод
+- stęp
+- ▁wronged
+- грэ
+- ▁ambayo
+- ▁chico
+- hypno
+- ▁شب
+- ▁Gall
+- bric
+- ▁demás
+- ▁evans
+- ▁Алар
+- ▁fasc
+- ▁paga
+- တ်
+- avaient
+- ▁hoek
+- ▁wretches
+- ▁збор
+- ▁mildly
+- ەك
+- arbaroj
+- quista
+- ▁этой
+- ▁erklärt
+- ▁배우
+- ▁guiding
+- ▁caressing
+- ▁ubushobozi
+- ிலும்
+- ▁Stein
+- ▁apprehensive
+- ▁recital
+- ▁trouvé
+- ▁barry
+- ▁cruiser
+- ▁quince
+- ▁voie
+- 리고
+- ▁adventurous
+- ▁wahrscheinlich
+- ▁ساعت
+- ავ
+- ▁clayton
+- त्र
+- bunda
+- ▁enlisted
+- circle
+- ▁indiana
+- ▁possède
+- bö
+- 했는데
+- рыш
+- لاپ
+- ▁texture
+- round
+- ▁badge
+- iamo
+- ▁Necessit
+- ▁clip
+- ▁gevaar
+- ▁essex
+- ▁хоць
+- ▁emit
+- ற்ப
+- ▁yaku
+- ▁oddly
+- ▁Mara
+- ▁Chicago
+- ▁accumulation
+- 量
+- ▁coincide
+- ▁drog
+- ▁свету
+- ▁dictator
+- staande
+- ▁preventing
+- ▁clouded
+- ▁guter
+- ʼ
+- ీ
+- ▁conversing
+- ەوێت
+- struck
+- ▁nombro
+- azioni
+- ▁இட
+- ▁terminated
+- ▁chirp
+- ▁paved
+- ▁Ĉiu
+- ключа
+- ▁palo
+- ▁ongelukkig
+- ▁jenkins
+- ▁Please
+- uṛ
+- ▁prix
+- vard
+- ▁Wali
+- ▁Abo
+- ▁itegeko
+- ▁ouden
+- nnan
+- ▁intact
+- kubo
+- ▁mida
+- ▁loĝantoj
+- ▁declining
+- ▁wichtige
+- film
+- imperi
+- ▁vertrek
+- ▁malheur
+- bina
+- ▁minuten
+- ▁meurt
+- ▁boos
+- élève
+- ▁penal
+- ▁yayo
+- movado
+- ľ
+- 持
+- ▁decency
+- ▁ngingo
+- ▁этого
+- ▁yew
+- ▁колер
+- ▁banish
+- ŝanĝ
+- жыў
+- ▁vorbei
+- வர்கள்
+- vish
+- ▁plymouth
+- ▁reina
+- ▁blinding
+- ҟ
+- ▁cooling
+- stav
+- ▁maryland
+- orerwa
+- uwo
+- त्या
+- ▁Only
+- ▁сваіх
+- ▁philo
+- 학기
+- olog
+- 이니까
+- ň
+- ▁Unternehmen
+- ▁troisième
+- ▁gouvernement
+- ▁madge
+- ுகிறது
+- ிக்
+- kammer
+- 場
+- 开
+- ▁hovered
+- ubururu
+- ▁intimation
+- wane
+- ▁vojo
+- ▁memoria
+- гө
+- ист
+- ▁carcass
+- ▁yap
+- cta
+- сек
+- ▁هەیە
+- ▁сме
+- ▁groping
+- ▁prophetic
+- ▁baas
+- ▁editorial
+- ▁mom
+- ▁defiant
+- ▁quay
+- ▁cicl
+- ▁contemporaries
+- ▁Geschichte
+- ▁생각하
+- ▁trespass
+- ▁되나
+- opo
+- ▁relied
+- ▁obert
+- ▁cortes
+- ▁heedless
+- ▁witches
+- iru
+- ▁dormir
+- ▁intricate
+- ▁요즘에
+- ▁vuestro
+- ▁pintor
+- avantage
+- ▁terminal
+- ▁participation
+- umunsi
+- ▁Ew
+- öz
+- proof
+- ▁decorat
+- ора
+- kaw
+- ▁permis
+- ▁tipp
+- ▁espècies
+- ▁internet
+- 꾸
+- ▁hypocrisy
+- ▁noneho
+- ▁회사
+- Ч
+- zunehmen
+- ▁mingling
+- һың
+- ▁табе
+- ▁midden
+- ▁compagnie
+- ▁disgusting
+- nitz
+- ▁Qq
+- ▁smoothed
+- ီ
+- هات
+- ▁overcame
+- ിയ
+- ▁distil
+- ▁entreaties
+- ▁ydy
+- ▁meekly
+- ▁bosh
+- ▁Landes
+- ▁کەس
+- ▁flourished
+- 江
+- bivuga
+- avança
+- accou
+- kwenda
+- ▁cafe
+- ▁expectant
+- ▁адной
+- ▁ике
+- ▁bushy
+- ▁heroism
+- κο
+- ▁lids
+- ▁мужчын
+- ▁puedo
+- ▁envied
+- bda
+- ▁beacon
+- ▁venster
+- iteit
+- خانه
+- ▁Kuri
+- ▁Qa
+- ▁parfois
+- 再
+- ▁bravest
+- lsa
+- ▁beneficial
+- ▁limestone
+- ▁Whi
+- ▁Rome
+- ▁flaw
+- ▁luxuries
+- ▁madonna
+- ია
+- ▁zara
+- ▁nécessaire
+- ▁государств
+- 早
+- obulamu
+- ▁isolation
+- ▁فِي
+- дно
+- ▁ponderous
+- ▁react
+- ▁على
+- ▁Fal
+- களுக்கு
+- ▁starlight
+- ▁Erst
+- ▁traum
+- ▁tromp
+- 석
+- ▁plutôt
+- ▁адбываецца
+- ▁involving
+- мәт
+- ▁econom
+- ▁walnut
+- ▁mosque
+- вары
+- ritse
+- ▁recep
+- ▁mouthful
+- ▁invoke
+- ▁поч
+- ▁seiten
+- 혀
+- ▁thoroughfare
+- ▁Ziel
+- ▁ardour
+- ▁scoff
+- ▁myśl
+- ▁districte
+- icles
+- 슬
+- ▁academic
+- guera
+- ▁regulation
+- ▁cumberland
+- ▁halting
+- ▁granny
+- ographie
+- ▁futbol
+- ▁보여
+- uvo
+- dade
+- ▁reassured
+- ▁lequel
+- zeichnung
+- ▁betrothed
+- した
+- steigen
+- suzum
+- ▁economical
+- ▁chuckle
+- ▁segment
+- ▁урын
+- ák
+- organitza
+- ходзяць
+- ξ
+- kond
+- ▁revolving
+- ▁cynthia
+- ▁afloat
+- 并
+- ▁geliebt
+- vention
+- ▁labored
+- цә
+- ▁вопрос
+- ҙән
+- shying
+- ▁fui
+- 해야지
+- ▁privacy
+- ▁repeti
+- ▁loosely
+- ▁droite
+- ▁irishman
+- سو
+- ▁extraordinarily
+- ▁tinge
+- ▁marquise
+- ▁virtually
+- ▁butterflies
+- ▁rendezvous
+- ▁crevice
+- ▁gesti
+- ▁provoke
+- ▁leḥ
+- ▁geringe
+- Bahn
+- ▁பெரு
+- amateka
+- ▁seriousness
+- ▁exaggeration
+- ▁appelé
+- ρο
+- ▁contentment
+- klin
+- ▁erzählte
+- ▁literal
+- ▁tsar
+- europa
+- hú
+- ▁должны
+- 권
+- ▁médecin
+- ▁bardzo
+- ▁limitations
+- pya
+- ▁colouring
+- ▁müsse
+- ▁graduated
+- ▁право
+- ▁kno
+- ▁tortoise
+- ▁яно
+- ▁seguida
+- ຫ
+- ▁nshobora
+- بری
+- human
+- ▁wanyama
+- احت
+- ▁cynical
+- ئي
+- ▁ryari
+- ▁wange
+- läu
+- ▁authorized
+- ▁Ré
+- ▁sitio
+- ▁wavering
+- ▁brilliantly
+- ▁edict
+- َوْ
+- ▁relentless
+- ▁Kelkaj
+- ▁Victoria
+- ▁matar
+- ▁indirectly
+- އަ
+- ▁bestand
+- ▁tutte
+- building
+- சை
+- 환
+- ▁conspirators
+- ▁straightforward
+- ▁Play
+- ▁solace
+- دەم
+- mán
+- ▁scornfully
+- zaga
+- ▁bough
+- вен
+- ▁좋아하
+- ▁dumm
+- ▁compositor
+- ▁fortifications
+- သည်
+- ▁yebɣa
+- hydr
+- ▁siberia
+- ▁yorkshire
+- ▁deseo
+- маҡ
+- ▁curling
+- ብ
+- ▁보이
+- ▁있거든
+- ▁bihe
+- ▁Hoch
+- ▁hairy
+- ▁Sh
+- صل
+- ▁privy
+- здар
+- ▁scornful
+- ▁раён
+- ▁allotted
+- ▁recesses
+- ▁Xi
+- when
+- ▁comprehended
+- ▁Biel
+- zni
+- ▁செய்த
+- ▁maakt
+- ▁christendom
+- ▁ҡуй
+- 思
+- льнікаў
+- ▁turkoj
+- ▁mobile
+- ▁reappear
+- 명이
+- ▁tenement
+- ▁dieselbe
+- investiga
+- ▁courant
+- ▁overseer
+- ▁vooruit
+- ▁Turk
+- igitaj
+- થ
+- ▁calhoun
+- ▁перыяд
+- quita
+- ▁yale
+- ▁yelling
+- ▁französisch
+- ▁Ёсць
+- 록
+- ▁surplus
+- ▁nourish
+- slaan
+- kreuz
+- ▁Boy
+- anywa
+- ▁hastening
+- ▁varies
+- ▁consummate
+- ▁austin
+- ▁көр
+- ▁issuing
+- lər
+- ▁malady
+- ▁absorbing
+- tanda
+- ▁jeweller
+- ފައި
+- ▁mormon
+- అ
+- ▁emmeline
+- ▁outrageous
+- ▁prevalent
+- masomero
+- ▁serge
+- ▁edwin
+- ▁obscured
+- ▁наступ
+- ▁brilliancy
+- ▁nkuru
+- ▁Genís
+- ▁spiral
+- ▁Akken
+- ▁cornwall
+- ▁motioned
+- ▁degenerate
+- ▁indirect
+- ▁characterized
+- ▁imparted
+- ▁abolished
+- ísimo
+- ▁된다
+- ònia
+- ▁miko
+- beeld
+- دام
+- geblieben
+- ▁lizard
+- ▁llengua
+- ▁johnnie
+- ▁Wu
+- ▁zoodra
+- ▁brightest
+- ▁پنج
+- ▁predicted
+- ▁aggr
+- adju
+- ▁rover
+- ▁weaving
+- ografia
+- ▁wenye
+- ▁broadcast
+- ▁colleagues
+- ▁rosalie
+- ▁nawet
+- ▁Kiam
+- ▁poker
+- atandukanye
+- ▁karen
+- ▁sneered
+- ▁Wood
+- women
+- 품
+- ▁проблем
+- ސަ
+- κά
+- astrolog
+- ▁тол
+- ▁Buch
+- öd
+- ▁bekommen
+- ▁knoll
+- ▁narrowly
+- ▁우리나라
+- ▁фільм
+- ombo
+- ▁coro
+- kaya
+- ▁rumble
+- ▁прыня
+- ▁usanga
+- ▁unheard
+- ▁ziemlich
+- ▁alaska
+- ▁contemplating
+- ▁hug
+- ▁listed
+- ▁Marcel
+- ▁juro
+- ▁kinini
+- ▁sombra
+- ▁Simon
+- ▁squeers
+- ▁булып
+- breite
+- ▁Зна
+- ▁bathe
+- ▁predecessors
+- bajije
+- bog
+- ▁eugen
+- ▁minste
+- ▁extracted
+- ▁swoop
+- ▁serenity
+- ▁gitabo
+- ▁kurze
+- ▁marshes
+- ▁softer
+- ▁flipp
+- ▁instructor
+- ▁sería
+- ▁بزرگ
+- ▁بیشتر
+- ინ
+- ▁ardor
+- ▁Cira
+- sezerano
+- ▁assuring
+- ▁comforting
+- sagen
+- étique
+- мәне
+- 쓸
+- ▁annihilate
+- osiyete
+- ▁manufactured
+- ▁nhw
+- ▁negra
+- ▁Fine
+- ▁Ilaq
+- ▁nanjye
+- ificar
+- lord
+- ▁northeast
+- ▁càrrec
+- ▁madeleine
+- ▁tranen
+- ▁Florida
+- pula
+- ▁falsch
+- ▁karere
+- кото
+- ▁flax
+- ктен
+- ▁shoal
+- ▁labyrinth
+- ▁rôle
+- ▁perezida
+- ▁dissolve
+- ▁Umwana
+- ▁walton
+- itsch
+- ▁Old
+- ▁salutation
+- itzen
+- usha
+- ố
+- ▁espíritu
+- xar
+- ▁daarna
+- ▁maggior
+- ▁loth
+- ▁umutwe
+- ▁cull
+- fí
+- ਂ
+- ▁vestige
+- ▁episcopal
+- 非
+- ▁dokter
+- ▁propio
+- ingiz
+- ▁fremde
+- ikipe
+- ማ
+- ▁cleopatra
+- ▁hammock
+- ▁treasurer
+- 提
+- ▁histories
+- tinted
+- ▁striped
+- ▁montan
+- ▁councillor
+- ▁беларускіх
+- ▁pompey
+- ▁dotted
+- ▁intervening
+- ▁realities
+- ▁basement
+- ▁Best
+- ோம்
+- એ
+- ކަށް
+- nnem
+- ▁sideways
+- ▁hydrogen
+- ▁participant
+- ög
+- maat
+- ▁hoarsely
+- ▁perkins
+- خورد
+- ئا
+- ▁conjunction
+- 욕
+- initia
+- ▁camping
+- ▁forfeit
+- ▁constantine
+- тыка
+- ▁haren
+- ിൽ
+- ڵا
+- ▁gulp
+- ▁panted
+- ▁testing
+- ▁حس
+- ▁내려
+- اتتى
+- ▁templo
+- ▁diligent
+- ▁змя
+- ▁termine
+- ▁icyemezo
+- ▁nostre
+- ▁Portugal
+- ▁konsistas
+- ต่อ
+- பட
+- ▁thundering
+- ▁hardship
+- gesund
+- 채
+- ▁debauch
+- ähne
+- ▁mainland
+- 에도
+- ▁yicaye
+- ▁trapper
+- сов
+- ▁gael
+- ▁cataract
+- ▁entangled
+- 主要
+- ▁bulb
+- йыш
+- ▁probe
+- ▁exclude
+- zamen
+- ▁decid
+- ların
+- fide
+- ▁உட
+- ▁olaf
+- ▁więc
+- abayobozi
+- ▁compagnon
+- ▁dishonor
+- ▁instantaneous
+- ▁succumb
+- iterambere
+- ▁seid
+- fuhr
+- ▁stronghold
+- ificat
+- аԥ
+- wege
+- ▁Estis
+- activité
+- ▁amazon
+- ▁marjorie
+- ލާ
+- ▁nelly
+- っ
+- ▁descubr
+- ▁abiding
+- Ind
+- ▁chanson
+- ամ
+- ▁Си
+- conduct
+- ▁global
+- rían
+- верх
+- ▁bigot
+- ▁quarrelled
+- ▁Ross
+- ▁Hamburg
+- ▁되니까
+- حق
+- ▁faites
+- ▁Liaj
+- ▁researches
+- ▁balcon
+- ▁explored
+- jiji
+- ზე
+- ▁begriff
+- ▁dictated
+- 으면은
+- 使
+- ▁outskirts
+- ▁siècle
+- ▁shuffle
+- 公司
+- ▁тысяч
+- épouse
+- ▁Final
+- embro
+- ▁bâtiment
+- ▁Sala
+- ▁peacefully
+- shadow
+- ▁affectation
+- 香港
+- ▁interminable
+- 镇
+- ▁dalle
+- izibwa
+- ahindu
+- ▁raiment
+- ▁adverse
+- ▁adoration
+- ▁Жа
+- ▁depending
+- ▁comparing
+- ▁lamented
+- ▁конкурс
+- ▁характар
+- ▁Company
+- ▁glittered
+- ▁এটি
+- ▁nyina
+- ▁йө
+- dydd
+- ən
+- ndosi
+- дала
+- ▁Francesc
+- ▁suave
+- ▁enacted
+- shira
+- ▁trouvait
+- kaze
+- ▁паша
+- zept
+- menyetso
+- զ
+- ▁iglesia
+- ▁barnaby
+- ▁advertising
+- ▁조금
+- ▁bloed
+- میر
+- nef
+- ции
+- ▁signora
+- ▁கட
+- 스타
+- ▁fondness
+- ▁defenders
+- ▁accorded
+- ▁remit
+- ▁archdeacon
+- ▁todavía
+- ▁daphne
+- ▁palestine
+- кыл
+- śli
+- বর্
+- endroit
+- ▁internacia
+- ordinaire
+- ▁برو
+- ▁excessively
+- ▁wanao
+- ้อง
+- เรื่อง
+- ▁betrachtet
+- ▁messiah
+- ▁cleft
+- twala
+- ▁hori
+- ▁Gusa
+- رسی
+- ▁Dra
+- ▁digitized
+- ▁reynolds
+- ▁brussels
+- ▁Vaig
+- ▁wuchs
+- ▁affirmed
+- ▁너네
+- ▁overflowing
+- ്യ
+- ▁Muḥ
+- ▁sleepless
+- ▁spät
+- vola
+- ▁precisa
+- ծ
+- 검
+- ▁busied
+- ▁talbot
+- czu
+- cilla
+- ▁enlighten
+- امی
+- ▁crumpled
+- ▁колькасць
+- شقا
+- ▁bash
+- ettaka
+- ▁rechte
+- ▁grecian
+- ▁pursuers
+- ▁چیز
+- ▁pinched
+- ▁interessant
+- ڵە
+- 꼬
+- ▁elliot
+- ▁signifies
+- ▁miał
+- ▁tunic
+- ▁examin
+- ىدىن
+- allemand
+- ▁dall
+- ▁subdue
+- ▁dello
+- gono
+- ▁fusta
+- өс
+- ▁Imbwa
+- ▁disagree
+- ▁tapping
+- brak
+- ▁yabi
+- ▁interesa
+- ▁annals
+- ނާ
+- ▁brav
+- ▁odin
+- ▁None
+- เลย
+- iḥ
+- cili
+- ረ
+- ীয়
+- ▁diligently
+- ▁brom
+- ▁ceremonial
+- ▁waarheid
+- ▁থেকে
+- ▁spricht
+- ▁armas
+- ▁études
+- ▁drained
+- ▁birdoj
+- ▁eich
+- ▁motivo
+- wah
+- тек
+- ▁labourers
+- цыйны
+- ▁vehement
+- ▁reappeared
+- дав
+- kisi
+- ▁uhr
+- ▁vehemence
+- ▁télé
+- 使用
+- ▁determining
+- itaire
+- ▁Great
+- ▁relativ
+- በ
+- лись
+- อก
+- ▁Carla
+- ▁lḥe
+- ▁kart
+- ▁modifica
+- ▁shrieking
+- ▁glorified
+- ▁pallor
+- ▁kelkajn
+- ▁flinging
+- ▁Main
+- ▁eddy
+- мө
+- ▁bony
+- ▁Berta
+- kko
+- ▁australian
+- ▁bihugu
+- ▁fallait
+- ▁бацька
+- ▁끝나고
+- igned
+- ▁erfolgt
+- جر
+- ▁брат
+- ▁Queen
+- irem
+- ▁Kann
+- ▁evolved
+- ҡҡа
+- ▁loĝata
+- ▁gitondo
+- ▁снег
+- nosti
+- ▁ownership
+- ▁cuán
+- ▁witty
+- áis
+- ▁plumage
+- 맨
+- ▁Damià
+- ▁salisbury
+- haven
+- ▁derselben
+- ▁kastelo
+- ▁stimulated
+- ▁buckingham
+- ▁Imp
+- masaha
+- ər
+- ▁respectively
+- text
+- ▁diverted
+- ▁pelle
+- ▁jock
+- ҽ
+- ▁predominant
+- ▁павінны
+- ▁verzoek
+- ▁König
+- ▁investigating
+- ▁durchaus
+- envol
+- ▁hotly
+- ▁exquisitely
+- ▁dynasty
+- ▁metaphor
+- ▁Nna
+- ▁rêve
+- java
+- ▁codi
+- ▁demà
+- eisen
+- ▁adeline
+- ▁obed
+- wezi
+- ▁zino
+- ▁manoeuvre
+- ▁ancestor
+- ▁vesper
+- ▁exceeded
+- ▁песні
+- ▁reduction
+- inconnu
+- çà
+- сіць
+- ▁Love
+- ▁قرار
+- ▁reddish
+- ▁arrogant
+- ▁despues
+- ▁Three
+- quito
+- ▁어느
+- ▁disclose
+- ▁Baada
+- ▁aurora
+- ▁livid
+- ▁এর
+- ambaye
+- ▁collapsed
+- ▁jake
+- ▁heureux
+- ▁morally
+- ▁Crist
+- ▁effectual
+- wumva
+- volta
+- ▁efecto
+- ▁wager
+- 었던
+- ▁gaping
+- ▁thud
+- ▁laboured
+- ▁плане
+- ▁аша
+- ▁redress
+- ▁wose
+- ▁cancer
+- دەست
+- terri
+- ▁gebeurd
+- ▁sensibility
+- ▁கரு
+- ▁notwendig
+- ▁Yal
+- ▁belli
+- ▁якім
+- эх
+- ▁guanyar
+- ▁System
+- ▁crear
+- цю
+- ▁refi
+- ▁advisable
+- ▁counterfeit
+- 嗰
+- ▁voldrien
+- ▁denmark
+- gonjwa
+- ▁consolida
+- ▁swede
+- ▁lucid
+- leggen
+- ▁segura
+- 받
+- रो
+- ▁quería
+- ціць
+- ▁reproaches
+- ▁vergeten
+- kristo
+- ۈم
+- ▁bueno
+- ტი
+- oby
+- ▁bangi
+- ▁micawber
+- 呀
+- ▁prodigal
+- صد
+- ▁staggering
+- math
+- ▁Ĉiuj
+- ▁fragile
+- ▁sehe
+- سته
+- ▁ikindi
+- bha
+- ▁wou
+- ▁صدا
+- ▁dairy
+- ▁너도
+- ▁malign
+- ▁Philip
+- order
+- ▁montri
+- turutse
+- dě
+- ▁smack
+- کرا
+- tenti
+- ▁என
+- ▁studoj
+- ▁wedded
+- ▁outcast
+- ழ்
+- ▁بې
+- emewe
+- ▁classification
+- двор
+- ▁vertu
+- ▁acquiesce
+- ▁eyo
+- ▁caractère
+- ▁recognizing
+- ▁bristol
+- witted
+- ▁comunitat
+- ▁capitol
+- shyamba
+- track
+- ▁perry
+- мәк
+- escence
+- ▁standpoint
+- ▁cuidado
+- ▁calor
+- befehl
+- ▁psycho
+- ▁brod
+- сот
+- ク
+- ▁będzie
+- 잠
+- ▁herhaalde
+- ▁produir
+- কি
+- fringed
+- lekt
+- trois
+- ▁сярэдн
+- ▁harvey
+- ▁obstruct
+- ▁ennyo
+- arrière
+- ▁silenci
+- ▁juga
+- ▁pinned
+- ▁Später
+- ▁سامي
+- 集
+- ▁kolekt
+- pocket
+- ▁mugabo
+- ▁schrie
+- ▁decreed
+- ▁bamaze
+- маҫ
+- ▁voudr
+- 感
+- 变
+- ▁títol
+- ▁gash
+- ▁carson
+- ҳә
+- ▁enirejo
+- ▁이미
+- ▁collective
+- ▁내일
+- ▁باشه
+- ▁Emi
+- ▁frosty
+- gird
+- す
+- ▁Cemetery
+- tekereje
+- ▁embracing
+- ▁Estudi
+- ▁welchen
+- ▁hippo
+- ▁criteri
+- ▁schatten
+- ▁withhold
+- ▁reproached
+- ▁vivre
+- ▁kapo
+- ਣ
+- ▁aussitôt
+- ▁dissolution
+- ▁stealthily
+- ▁thar
+- ▁лий
+- ▁yitwa
+- ▁repre
+- даваць
+- ▁reserva
+- ▁bray
+- äum
+- ▁bushel
+- ▁summary
+- ▁crumbs
+- ▁fleurs
+- ▁okcidenta
+- ▁auxiliar
+- ▁cambia
+- бат
+- ▁charging
+- вую
+- ▁gentry
+- ▁gutera
+- เน
+- тты
+- ři
+- ▁brauchen
+- ▁trajecte
+- ▁languish
+- 京
+- 参
+- ▁manifold
+- ▁baking
+- ▁Yer
+- ▁stacy
+- assassinat
+- ▁stati
+- ▁robot
+- чет
+- ਟ
+- ▁solle
+- สม
+- னா
+- 자고
+- ▁disso
+- viti
+- ▁Yn
+- ▁canaan
+- ▁Ihnen
+- ▁allude
+- ▁diari
+- impuls
+- ▁ولی
+- ည်
+- ▁mustache
+- វ
+- ▁erroneous
+- ▁monopoly
+- 族
+- ▁subjection
+- ▁보내
+- ▁seltsam
+- ▁hilt
+- '60'
+- ▁darting
+- ▁Gew
+- ▁bruise
+- ▁sufficed
+- gelaten
+- oggi
+- ▁viaje
+- ▁piteous
+- ▁tidy
+- teksto
+- ▁interrupting
+- ▁raoul
+- ▁habitually
+- wezen
+- мах
+- ▁bonna
+- jorn
+- ▁nulla
+- مەت
+- ской
+- ▁Yehova
+- ▁колькі
+- 넌
+- ▁যে
+- 座
+- ▁caresses
+- ▁stolid
+- ▁sniffed
+- pital
+- ▁stale
+- ндар
+- ▁priesthood
+- ثر
+- ▁diferent
+- ▁камен
+- гүй
+- ▁hopped
+- ▁expended
+- ▁thorndyke
+- ▁sunbeam
+- ▁turmoil
+- ਦੇ
+- дзь
+- выш
+- ▁gukomeza
+- kiro
+- ецер
+- ▁practiced
+- խ
+- ▁Nuntempe
+- ▁chasing
+- minster
+- tiere
+- দু
+- ▁scattering
+- ▁facilit
+- রণ
+- ▁États
+- ▁colorado
+- ▁reversed
+- دون
+- ▁disapprove
+- ▁incomplete
+- ▁ہیں۔
+- ▁congenial
+- ▁uncanny
+- ില
+- ▁Нам
+- ▁Bydd
+- ▁স্
+- ▁haast
+- ื้อ
+- 깨
+- ▁initiative
+- ▁shoved
+- ▁throb
+- ▁chris
+- ▁surge
+- 형
+- ▁faible
+- аванне
+- ▁frue
+- artisan
+- ▁mackenzie
+- ▁lancelot
+- ▁bourbon
+- ▁gatatu
+- ▁waterloo
+- ▁Ик
+- ▁usefulness
+- ግ
+- ▁remedies
+- ె
+- ▁нельга
+- ▁mtu
+- ▁sausage
+- ▁бөт
+- ▁Rolle
+- ▁jugend
+- ▁finite
+- ▁culprit
+- ▁etiquette
+- ▁представител
+- ▁pitiless
+- ▁jovial
+- ▁martyrdom
+- looze
+- ▁corrobor
+- яш
+- பா
+- Comp
+- führer
+- ▁tandis
+- һыҙ
+- ▁vivant
+- ▁πρ
+- سې
+- ▁maestro
+- ▁набы
+- ▁دوو
+- ப்பா
+- 블
+- ▁microscope
+- ▁omnibus
+- жын
+- judi
+- ေး
+- ▁staple
+- ▁communi
+- ▁subsidi
+- ▁specioj
+- ल्
+- ▁mimic
+- ▁Mana
+- ▁olvid
+- ▁système
+- ▁unfamiliar
+- ▁Больш
+- ▁reuben
+- ▁Rod
+- ▁thundered
+- ▁foregoing
+- burdened
+- ляў
+- ritz
+- த்துக்
+- ▁Kre
+- śmie
+- ံ
+- ▁alacrity
+- ▁persoon
+- ▁steamboat
+- midi
+- ონ
+- ▁middleton
+- здым
+- ▁gallantry
+- ▁ingratitude
+- ▁maturity
+- ▁dazzled
+- 百
+- ▁gerne
+- تُ
+- ▁brag
+- ▁loomed
+- tewe
+- ▁fleda
+- ▁һал
+- Ы
+- ▁stabbed
+- 洲
+- வன
+- ▁Komm
+- ▁aristocrat
+- itia
+- ▁steaming
+- intambara
+- ▁rigging
+- ▁maisha
+- ▁honra
+- ▁wegg
+- ۾
+- 完
+- ▁arguing
+- ▁chastise
+- mbaga
+- ▁boiler
+- ▁starr
+- মু
+- ▁petals
+- ▁militaire
+- ▁dunkle
+- rushanwa
+- ▁gito
+- ▁People
+- ▁prinzessin
+- pool
+- ▁назва
+- ▁breathlessly
+- ▁awarded
+- ossa
+- ático
+- ▁roasted
+- ▁gleichen
+- struktur
+- ▁taille
+- ▁volkomen
+- geschenk
+- ▁soldados
+- сацыя
+- ▁entendu
+- ▁Big
+- ▁demeanour
+- ▁rained
+- ▁follies
+- ▁bajya
+- ▁горад
+- ▁rebut
+- ▁Neida
+- written
+- даюць
+- ▁complement
+- ими
+- ɛaw
+- ▁afecta
+- ▁reconnu
+- ▁whirlwind
+- ▁Совет
+- ▁Palm
+- ▁Bern
+- ▁ŝtato
+- lix
+- cteur
+- ▁Temp
+- ▁Ла
+- ▁поўна
+- கொள்ள
+- ▁française
+- ▁sunken
+- ▁chien
+- ▁arras
+- kufu
+- ▁Gina
+- ▁governing
+- ôme
+- ▁ishirini
+- ▁tmurt
+- ▁avowed
+- ▁viceroy
+- ▁кніг
+- ಳ
+- зала
+- fläche
+- ▁shipped
+- ▁ஆக
+- ▁Mata
+- ▁closest
+- โดย
+- ▁repulsive
+- walk
+- ▁mollie
+- yisa
+- ▁thronged
+- country
+- ▁вынік
+- cella
+- ▁farce
+- ▁schlimm
+- ▁chiama
+- ▁diners
+- ހެ
+- trans
+- ▁urba
+- ▁lute
+- 現
+- ▁conceivable
+- blätter
+- ▁satellite
+- zí
+- brien
+- ▁enlist
+- ▁everard
+- ▁testa
+- யிர
+- ▁یاد
+- เห็น
+- ▁desgracia
+- ▁canterbury
+- ▁affirmative
+- ▁Avui
+- ▁cocked
+- ▁рада
+- ▁eruption
+- ▁girlish
+- ▁folgen
+- ▁gewissen
+- ▁ناو
+- ▁misunderstood
+- ▁preposterous
+- ▁шэраг
+- ▁despotism
+- ▁cuarto
+- 官
+- ாலும்
+- ▁nimble
+- ▁수업
+- ▁Мый
+- ▁balk
+- ▁annoy
+- ▁pudo
+- ▁mystical
+- ▁stanton
+- ધ
+- 普
+- 着
+- ▁blouse
+- ▁bridal
+- นัก
+- perti
+- ▁drying
+- czyn
+- авання
+- mütig
+- ▁paraît
+- ▁nicol
+- луу
+- ▁Ир
+- ▁clarissa
+- ▁concluding
+- ▁whimsical
+- ▁Award
+- responsibilities
+- ▁супраць
+- ▁هەموو
+- ▁engraved
+- cando
+- ▁defied
+- ัย
+- chreck
+- roost
+- ▁moed
+- ▁daarvan
+- ▁pasó
+- ▁Tiziri
+- ▁mezepoko
+- ষ্ট
+- shimishije
+- onym
+- Oriol
+- ▁inquiet
+- ▁detection
+- ▁Prince
+- chodzi
+- zonder
+- ▁zoodat
+- ▁nomina
+- ▁lebte
+- ▁outlined
+- mashusho
+- ▁Umukobwa
+- ▁Regierung
+- ▁informing
+- ▁expelled
+- ▁Because
+- kräfte
+- ▁Кан
+- hield
+- dził
+- thirty
+- ▁grandan
+- ▁linden
+- ▁பே
+- رن
+- kwo
+- ziko
+- ▁oasis
+- ▁diable
+- ▁উপ
+- ▁sterben
+- qay
+- ▁sobald
+- ▁booty
+- ▁dyke
+- ▁foxes
+- ▁catharine
+- ▁engrossed
+- ▁sepulchre
+- ▁sanguine
+- وک
+- ▁Gir
+- ▁Germany
+- piro
+- ▁Sitz
+- deelen
+- humoured
+- апла
+- չ
+- ▁gendarme
+- objekt
+- ▁frequented
+- wych
+- entrepren
+- ▁feud
+- ▁ноч
+- fluss
+- ▁predecessor
+- ▁josephine
+- ▁weston
+- agent
+- ▁ngeri
+- ▁என்
+- ▁cabell
+- ▁soiled
+- engine
+- ▁mutima
+- ▁называ
+- għa
+- ▁mountaineer
+- infini
+- ▁없지
+- bär
+- 험
+- ubutegetsi
+- อีก
+- ▁greasy
+- رش
+- kommt
+- ▁poised
+- grove
+- ▁sarebbe
+- ▁prelate
+- ▁Uko
+- ▁Llu
+- ▁whosoever
+- お
+- ▁gipsy
+- ▁அவன
+- acqua
+- ▁Universitato
+- ▁insulting
+- injira
+- ▁fresco
+- kubiri
+- ▁fons
+- ወ
+- ▁quattro
+- ▁будет
+- ▁azekka
+- ▁выклік
+- ▁gallon
+- ▁Juan
+- ▁Edward
+- ▁сябр
+- ▁bias
+- ▁ajouta
+- ▁pauline
+- ▁comtesse
+- ▁lilly
+- market
+- ▁modestly
+- ▁punta
+- ▁schenk
+- ▁successes
+- '80'
+- 했을
+- ▁termination
+- ▁auntie
+- теп
+- ▁Alex
+- ▁nkunda
+- স্ত
+- 기로
+- ▁evangeli
+- îne
+- ▁kial
+- เร
+- ▁dictionary
+- ▁unhappiness
+- ▁vérité
+- ▁Não
+- ▁dónde
+- ▁grabbed
+- ▁substituted
+- ята
+- пон
+- ▁comtat
+- ▁Erfolg
+- aĝa
+- няе
+- ▁Ulac
+- antaj
+- ▁ceremoni
+- ▁kalt
+- ▁methodist
+- ▁lovingly
+- ▁Xènia
+- 앞
+- ▁raffles
+- ▁Gy
+- ▁Kwi
+- ▁Chu
+- ىكى
+- kki
+- spring
+- ▁disappearing
+- havien
+- ით
+- umuco
+- hisha
+- ▁voul
+- ▁ھەر
+- brow
+- engeri
+- ▁sitzen
+- ▁phillips
+- gruppe
+- ▁rzekł
+- àl
+- ▁überall
+- ▁wholesale
+- avam
+- ▁enorme
+- ▁kidnap
+- losen
+- ▁particle
+- ▁exasperated
+- ▁چطور
+- 收
+- ▁boffin
+- М
+- ggya
+- ▁homestead
+- prin
+- ▁charcoal
+- ▁merchandise
+- ▁Center
+- amashuri
+- ▁yawning
+- ngles
+- ▁hardness
+- ▁Kino
+- ღ
+- ▁provinco
+- ▁dryly
+- ▁Kab
+- ▁وہ
+- ▁hubo
+- ▁bwose
+- ▁lining
+- ▁questi
+- ஊ
+- ▁timothy
+- sided
+- ▁slower
+- ritten
+- ▁servicio
+- häuser
+- ▁wê
+- ▁giddy
+- änglich
+- ▁pila
+- contamina
+- avion
+- ▁sumptuous
+- ▁teaspoonful
+- 송
+- ▁gairebé
+- ▁recherche
+- zame
+- spraak
+- baze
+- ▁drooped
+- iyê
+- regierung
+- ▁haunting
+- ▁yellan
+- ▁ndege
+- ▁glue
+- ▁rave
+- zongera
+- ▁지금은
+- ▁camisa
+- бед
+- 展
+- ▁aghast
+- byabaye
+- ▁choisi
+- ▁kubaka
+- embarras
+- ▁philippe
+- ▁womb
+- ▁lind
+- roque
+- ▁느낌이
+- ▁diamant
+- ▁reassuring
+- 著
+- 近
+- ▁deficient
+- ▁invaders
+- ▁untuk
+- ▁crippled
+- ▁świ
+- ▁heave
+- ▁uncommonly
+- pox
+- ▁спі
+- ರ್
+- 론
+- ▁атрымаў
+- ▁জন্য
+- ▁felon
+- ▁cyiza
+- zieht
+- ▁Zweiten
+- hava
+- ▁jammer
+- ಟ
+- ដ
+- ▁auprès
+- 김
+- 뽑
+- ▁richelieu
+- ▁cheveux
+- ▁agencies
+- چه
+- ▁liggen
+- ▁nomination
+- 째
+- ▁dienen
+- ▁schul
+- ▁insert
+- ▁huwa
+- ▁працуе
+- agombaga
+- ▁banana
+- ▁قوي
+- ▁foreboding
+- ▁bequeath
+- ▁défend
+- үй
+- кин
+- ▁complied
+- работ
+- ▁ئىدى
+- ▁obeying
+- rā
+- partei
+- ▁bragg
+- ▁quilt
+- मि
+- ▁weakly
+- ▁profusion
+- ▁siendo
+- ▁terence
+- ไม่มี
+- ▁suivante
+- تىپ
+- ഷ
+- bbel
+- ▁preso
+- saxon
+- ▁contemptible
+- ▁managing
+- ▁اگر
+- нутр
+- schwer
+- ื
+- िक
+- ▁independently
+- тычны
+- ▁Regel
+- ▁darkest
+- ▁Blo
+- cene
+- ▁kwaad
+- äus
+- ▁analogy
+- ▁дзяўчын
+- ▁epithet
+- jährige
+- ▁إلى
+- ▁meagre
+- ▁대해서
+- ▁Gib
+- ▁garanti
+- zol
+- ▁perpetua
+- ▁neuro
+- öyle
+- jugu
+- ▁peggotty
+- ▁nötig
+- ▁poultry
+- ▁teufel
+- bility
+- ▁mummy
+- ▁farthing
+- freiheit
+- களில்
+- ▁wahrheit
+- ▁simp
+- ▁stroked
+- ▁spied
+- ▁colleague
+- iriwe
+- wî
+- Alb
+- américain
+- പ്പ
+- 总
+- ▁sleeper
+- ▁township
+- ▁schwa
+- ▁splendidly
+- ▁schoolboy
+- ▁wangen
+- ▁stoic
+- ▁okwo
+- ▁begreep
+- 演
+- ▁bailiff
+- ▁동생
+- ▁ukora
+- ▁Kü
+- ▁zest
+- ĵo
+- ▁finestres
+- ▁certo
+- лыя
+- ▁Кон
+- formation
+- ▁attired
+- ▁рыб
+- ▁humphrey
+- 广
+- ▁segunda
+- ▁conquering
+- لك
+- ▁direcció
+- ского
+- ▁personnel
+- ▁esposa
+- ▁Лі
+- ▁தன்
+- ▁транс
+- ▁comunica
+- ▁Winter
+- ▁소리
+- ▁acquit
+- ▁dirk
+- ▁subterranean
+- ▁فقط
+- ▁forefinger
+- 还
+- ▁behalve
+- balo
+- ▁unfolded
+- ▁zealand
+- віта
+- vreemd
+- stuhl
+- ހ
+- ▁While
+- ▁Sança
+- ▁Brit
+- ൻ
+- ▁vindictive
+- ▁encampment
+- ▁све
+- 伊
+- ▁smitten
+- ▁approving
+- कार
+- ▁senyor
+- ▁праблема
+- ultima
+- ▁wireless
+- ▁angeles
+- ▁norton
+- 达
+- ▁누가
+- ▁evangelist
+- ствен
+- біць
+- णा
+- слав
+- ▁앞에
+- ▁Cependant
+- ▁anonymous
+- 青
+- ▁composé
+- خواست
+- ▁personnage
+- ▁supre
+- aventura
+- ític
+- ▁plait
+- ▁indifferently
+- ▁docteur
+- ▁händen
+- tokea
+- нулі
+- 탈
+- cause
+- ▁físic
+- pang
+- venta
+- ▁kreeg
+- ▁acquaint
+- ▁considerat
+- ▁compensate
+- ▁flav
+- ▁அமை
+- ▁disgraced
+- хан
+- ▁mace
+- imiryango
+- ▁imprudent
+- ▁compressed
+- ▁гадзін
+- ▁persevere
+- ▁enrich
+- ▁ishati
+- ▁cooled
+- ▁snip
+- ▁Bak
+- цоў
+- ▁frying
+- clé
+- ▁memorandum
+- ▁turbulent
+- ▁cierta
+- ▁deputies
+- ▁resorted
+- ર્
+- tuli
+- ▁bagenda
+- ▁crashed
+- ağı
+- ▁Dda
+- qualifi
+- ñu
+- shake
+- reisen
+- เป
+- ჩ
+- ▁intrusion
+- ▁بسیار
+- ▁sonnet
+- ▁Nyuma
+- ▁howled
+- ▁intrude
+- ذر
+- kā
+- ▁generously
+- ▁adaptation
+- ▁magambo
+- ▁வாழ்
+- alcalde
+- ▁geduld
+- ▁repented
+- neḍ
+- ▁loco
+- рэй
+- wahi
+- èle
+- hush
+- ▁мове
+- ▁amsterdam
+- ▁secluded
+- ▁hindrance
+- ▁speculate
+- ▁йӱ
+- ▁helfen
+- fifth
+- ▁manĝas
+- ▁libroj
+- пак
+- ▁discarded
+- 件
+- ▁apartenis
+- пен
+- ▁napkin
+- gry
+- ▁hoof
+- ▁thé
+- ▁pública
+- rillo
+- kow
+- 는지
+- ▁impudence
+- trifft
+- ▁важны
+- ▁grazing
+- ▁subito
+- ▁Lord
+- ▁elekti
+- ބު
+- ▁fece
+- กว่า
+- ▁багат
+- ото
+- ▁entwickelt
+- ▁eenvoudig
+- ▁philistine
+- ▁submissive
+- ▁embarrassing
+- ▁ئۈ
+- ▁중국
+- ▁befinden
+- ▁snapping
+- lith
+- çar
+- ▁eigenlijk
+- آور
+- пис
+- ▁bulwark
+- 광
+- ▁versailles
+- ▁árbol
+- ▁yehova
+- ▁nennen
+- ▁شهر
+- jú
+- ▁outraged
+- ▁moose
+- ▁crutch
+- ▁descendant
+- urupfu
+- katu
+- abaga
+- вон
+- ▁European
+- rzył
+- ▁diferentes
+- ▁arbres
+- ▁index
+- ▁fertil
+- ▁manchmal
+- '10'
+- ▁paard
+- ▁sinne
+- ▁йәш
+- ▁teresa
+- ገ
+- cznie
+- ▁없잖아
+- ▁grenz
+- ац
+- ▁تک
+- ▁posar
+- ▁مرا
+- анд
+- ▁Entwicklung
+- ▁comanda
+- ▁Іх
+- ▁infirm
+- ▁cited
+- ▁ھەم
+- zaza
+- ▁Jones
+- metric
+- ▁communicating
+- ▁canopy
+- ▁flit
+- ▁Civil
+- чат
+- ▁beobacht
+- кого
+- ▁کنند
+- ▁vestido
+- ▁வரு
+- ряд
+- kufa
+- ▁volstrekt
+- ▁aperture
+- ▁amasezerano
+- ▁retrace
+- ilidad
+- ▁있었는데
+- ▁nyir
+- ▁Tiene
+- ▁ogu
+- ಯ
+- ▁bikaba
+- ▁lejos
+- ▁innocently
+- ▁rapidement
+- kür
+- ▁dint
+- ▁effectively
+- гры
+- ▁Temas
+- ▁calci
+- や
+- ▁surmounted
+- ▁goblin
+- ▁پرو
+- ▁echa
+- өлө
+- tesi
+- alimenta
+- ▁daytime
+- reliance
+- ▁stonden
+- ▁byabo
+- ganza
+- ฝ
+- ▁thrift
+- ▁gyda
+- ▁요새
+- ▁Ambaŭ
+- ▁loveliest
+- ▁uphold
+- université
+- temperatur
+- lander
+- ▁constit
+- ▁steered
+- brother
+- ▁croire
+- ▁regularity
+- 더
+- ▁погляд
+- ▁comporta
+- ▁inaccessible
+- ▁unquestionably
+- κε
+- ▁leopard
+- 我們
+- ▁갔는데
+- ▁straf
+- ▁cheaper
+- ▁overlooking
+- 쪽
+- batera
+- ▁cuya
+- ▁halfway
+- ▁Iri
+- ▁worten
+- ▁schouder
+- ▁dejected
+- dito
+- ▁와서
+- lehrer
+- ibuka
+- ▁filial
+- pida
+- ▁critique
+- 었거든
+- ▁civility
+- ▁ambao
+- amini
+- isce
+- ▁wretchedness
+- issimo
+- ɛedda
+- ▁jolt
+- 았어
+- ▁publicity
+- ▁verdure
+- ▁postpone
+- ▁herrlich
+- ▁wokulski
+- ▁glossy
+- ▁dennis
+- ▁Bü
+- ▁bennett
+- ▁cyari
+- ▁mixing
+- ланд
+- ▁blink
+- ▁compla
+- ▁poplar
+- ▁frederic
+- kozesa
+- ▁Wahl
+- ▁якой
+- ىڭ
+- 원이
+- ▁fräulein
+- ▁meditative
+- àlbum
+- ▁cocoa
+- ▁Beispiel
+- жив
+- ▁geheimnis
+- ▁puppy
+- mbuzi
+- ▁Sho
+- ▁prohibited
+- entrée
+- ▁cripple
+- bwiriza
+- ▁piú
+- ▁debía
+- ຈ
+- ▁spout
+- agaragara
+- pyr
+- ▁broadway
+- ▁credited
+- ▁devilish
+- ▁mirada
+- ▁Ай
+- kyali
+- ▁ambos
+- 리가
+- ▁고등학교
+- ▁buffet
+- ▁என்ப
+- ола
+- ▁Quine
+- мам
+- shiriki
+- ▁preciso
+- ▁ôl
+- lohn
+- ▁lurch
+- ▁있었어
+- sonnen
+- pura
+- ▁titus
+- ▁obulungi
+- ▁mohawk
+- ▁atonement
+- ▁smite
+- ▁stewart
+- ▁creaking
+- ων
+- гол
+- ▁invece
+- wissenschaft
+- ▁havi
+- żył
+- ่าย
+- жаны
+- ▁가자
+- ifique
+- ▁килә
+- accés
+- ▁botiga
+- ▁fetters
+- ▁begonnen
+- ▁bakaba
+- ▁brett
+- ▁competitor
+- ague
+- нікам
+- стве
+- ▁yawned
+- imiz
+- ▁fastening
+- ▁goeden
+- ҙәре
+- ▁specialist
+- ▁staid
+- ▁burying
+- ז
+- ▁Gavumenti
+- ▁confederacy
+- ▁prescription
+- ▁unscrupulous
+- တွင်
+- ▁дакладна
+- ▁somerset
+- ▁furono
+- ҵә
+- ▁toilette
+- ▁assim
+- ▁सं
+- mə
+- ▁Luft
+- graben
+- ▁dispatched
+- ▁voltant
+- ▁graze
+- ▁rowing
+- ▁için
+- ▁zwak
+- ▁Home
+- aigües
+- ջ
+- ▁momentarily
+- авы
+- ▁steeple
+- ▁Helen
+- szę
+- ▁taɣ
+- ▁முடிய
+- isée
+- ▁만들
+- ▁Court
+- ально
+- ▁compassionate
+- ▁цікав
+- ▁efface
+- ▁officially
+- ▁suspecting
+- ▁antecedent
+- ▁peligro
+- 業
+- ▁gibson
+- mbwe
+- ▁pastime
+- ▁contrasted
+- йшло
+- ▁emphatic
+- няць
+- ாமல்
+- ennemi
+- ▁beamed
+- ▁pastoral
+- skinned
+- ራ
+- ▁harlowe
+- ▁obtenir
+- ▁rezult
+- даг
+- ڕی
+- ▁kabila
+- ▁ozma
+- ▁jonas
+- krat
+- ▁okuloj
+- سي
+- ▁artful
+- 义
+- ▁avui
+- ทําให้
+- ▁vermin
+- ▁commença
+- ▁volgde
+- loving
+- ▁slang
+- yisi
+- twintig
+- ▁geleden
+- ▁aviat
+- ▁vigilant
+- மர
+- ▁cosmo
+- ▁Uwo
+- ▁immovable
+- ▁핸드폰
+- ▁enumerate
+- ▁arque
+- ▁mince
+- ▁итеп
+- astre
+- ▁Egi
+- ▁bosque
+- ▁wären
+- boom
+- റ്റ
+- ▁astounding
+- ▁grasshopper
+- ▁sagacity
+- ▁결혼
+- ▁dimension
+- ▁residu
+- ▁pseudo
+- hängen
+- cieux
+- ▁divan
+- 변
+- ▁Conserva
+- ▁மற்றும்
+- ▁значыць
+- ▁principales
+- дры
+- ▁moglie
+- ▁Centre
+- ▁اول
+- izado
+- fc
+- ▁کیا
+- 하시
+- ▁electoral
+- ▁mocked
+- ibikoresho
+- ▁bikomeye
+- ▁sympathis
+- ▁avrebbe
+- ▁Serra
+- oxy
+- ▁forêt
+- ▁hoss
+- ▁quindi
+- ▁azure
+- ▁jade
+- gekomen
+- ▁Pakistan
+- ▁является
+- 움
+- ശ
+- كُ
+- ▁único
+- ▁estrella
+- อด
+- ▁thaw
+- ▁gratifying
+- ▁کمی
+- ▁pouco
+- ▁privileged
+- зму
+- ▁człowiek
+- ▁terrestrial
+- ▁zwölf
+- จํา
+- ▁저번에
+- ची
+- ▁saracen
+- ▁prestige
+- hamagara
+- ▁stave
+- ▁efect
+- ▁labouring
+- ▁다음에
+- ▁diverge
+- ▁beastly
+- icios
+- ▁swirl
+- ▁gestalte
+- ▁expedi
+- ▁voel
+- ປ
+- ላ
+- Igihugu
+- ▁subscription
+- ▁surviving
+- ▁automatically
+- ▁presbyterian
+- ▁imaginable
+- prez
+- ▁Гр
+- ▁아닌
+- ▁versió
+- ԥш
+- ▁dû
+- 邊
+- ▁savez
+- ▁simpson
+- ▁менен
+- لىپ
+- ▁participe
+- ▁transit
+- ▁trompe
+- fuzi
+- schrijven
+- ▁hyacinth
+- 居
+- ▁fidget
+- ▁Blau
+- bok
+- ▁incurred
+- shami
+- ▁dandy
+- ▁goût
+- ▁necessaries
+- ▁trough
+- ivugwa
+- ▁trok
+- wow
+- ▁Team
+- ▁filth
+- ித்த
+- ▁кеүек
+- фу
+- ▁traurig
+- ▁terry
+- ▁verzeker
+- 夫
+- ▁helplessness
+- ър
+- ▁controlling
+- ிக்க
+- ▁деп
+- ▁bitewe
+- οι
+- rurimi
+- ▁Schu
+- ▁partida
+- ▁writhing
+- ▁prowess
+- ▁Before
+- ▁expound
+- ▁preguntó
+- ▁creditors
+- ▁bijyanye
+- мор
+- ▁filed
+- richter
+- ignor
+- ▁mourned
+- built
+- eût
+- ▁Festival
+- ▁oats
+- gida
+- нулася
+- льную
+- ▁muzeo
+- ▁unmarried
+- flügel
+- Espanya
+- ▁هل
+- ▁Out
+- ▁голо
+- ▁scratching
+- ▁Nzi
+- ▁sahen
+- ▁monitor
+- رین
+- ▁pecuniary
+- 東
+- ▁mazarin
+- ▁mitchell
+- ▁película
+- ▁skillful
+- ▁کردی
+- ▁shiny
+- ▁thornton
+- ▁Name
+- ▁unimportant
+- ▁17
+- spieler
+- urugo
+- ▁reginald
+- ▁павінен
+- ▁pedestal
+- 命
+- ▁ehre
+- ▁чакае
+- funkci
+- ▁runaway
+- ▁creix
+- ▁уйла
+- ৱ
+- ▁unbelief
+- ▁fleming
+- yise
+- ▁surgery
+- ▁downright
+- ▁redeemed
+- ▁jeremy
+- ▁dominated
+- ▁extinguish
+- ों
+- ▁гэтае
+- ▁consoled
+- ▁кеч
+- ▁durst
+- һө
+- ▁рука
+- ▁Kuba
+- ▁televisió
+- 反
+- rutse
+- ▁Puig
+- engage
+- ▁golpe
+- miş
+- gizwe
+- ▁blanka
+- ▁genteel
+- ▁impulsive
+- ماي
+- ▁exploring
+- ގޮ
+- 党
+- ▁deduction
+- ▁gypsy
+- ோடு
+- ▁prohibit
+- ▁얼마나
+- ▁jedno
+- ▁hyper
+- ▁ruim
+- ▁waked
+- ▁Jane
+- ▁brewster
+- ▁servitude
+- ▁admonish
+- ▁rijtuig
+- tiye
+- ▁deuxième
+- ▁기억이
+- برد
+- ▁conversed
+- ▁suerte
+- يق
+- ▁seduce
+- ಸ್
+- ▁تمام
+- نگی
+- ுடன்
+- 꼭
+- hindi
+- ▁stimulus
+- گۆ
+- ▁Hong
+- voeg
+- годна
+- achtet
+- middag
+- ▁starry
+- öö
+- ▁podria
+- ▁imitated
+- ▁véritable
+- 땐
+- 波
+- ▁supplication
+- ރީ
+- йҙар
+- ▁troch
+- ▁grimace
+- entreprise
+- ▁좋아하는
+- otti
+- kritik
+- ▁высокі
+- ▁produc
+- ▁adjustment
+- ▁deluge
+- ▁tarry
+- lomb
+- ติ
+- church
+- تار
+- знаём
+- ▁montreal
+- ▁gesamte
+- ▁supérieur
+- ▁ehren
+- ▁germain
+- ▁Jackson
+- ▁Text
+- ▁پول
+- platte
+- nyira
+- ulla
+- ▁lobby
+- ▁choke
+- ▁impede
+- ▁hindered
+- ӓт
+- ▁хотел
+- ▁пункт
+- ▁Cela
+- treiben
+- ▁palpita
+- ▁syr
+- ▁nibyo
+- âtre
+- ▁distinguishing
+- ▁rotha
+- һында
+- ɣel
+- tendant
+- म्
+- ▁якіх
+- ▁ஆனால்
+- ▁easiest
+- ▁vollständig
+- ▁아는
+- ▁лӱ
+- ▁uppermost
+- выр
+- ▁Twi
+- ▁Fund
+- ▁Town
+- ▁industria
+- ภาพ
+- круг
+- 드라
+- ▁annoying
+- gema
+- issaient
+- ັ
+- ubugingo
+- ïda
+- ▁ironical
+- ▁politiki
+- ▁scharf
+- crum
+- ▁halo
+- ▁inspira
+- guruka
+- ▁сваё
+- ▁batatu
+- әһе
+- ▁stupendous
+- ▁serikali
+- kuk
+- ▁setmana
+- ▁حالا
+- girwa
+- гавары
+- ▁abstracted
+- drome
+- ▁mohammed
+- ▁blockade
+- ▁gehören
+- چھ
+- ▁interchange
+- ааи
+- ▁patrino
+- ктор
+- bäume
+- ▁soldaten
+- ▁sedert
+- rimmed
+- ▁درست
+- ▁Sergi
+- ох
+- 府
+- ▁найбольш
+- ▁natasha
+- 型
+- ▁assisting
+- ٹی
+- ▁mistrust
+- ▁يەر
+- ▁friction
+- ▁barrister
+- charge
+- ▁benefici
+- ▁булыр
+- 们
+- ▁iroquois
+- ▁troubling
+- hinde
+- ▁pali
+- ▁squint
+- ία
+- ▁traves
+- يات
+- плат
+- ▁barton
+- qual
+- ▁planter
+- そ
+- ▁frequency
+- ▁forefathers
+- ▁unhappily
+- ▁hubiese
+- ▁convoy
+- ন্দ
+- ▁seguito
+- ▁Rosa
+- ▁starboard
+- ▁durham
+- floor
+- ▁calculate
+- ▁dauphin
+- ▁jednak
+- ▁offenbar
+- 이면
+- inici
+- ▁vehemently
+- casse
+- нән
+- ્ય
+- Umwami
+- 死
+- ▁пачатку
+- ▁crumble
+- ▁inflamed
+- ▁manifestly
+- ▁marianne
+- کەم
+- ▁displaying
+- ▁deciding
+- vance
+- ▁Spring
+- ▁Iran
+- ▁estudis
+- ▁johann
+- ▁кок
+- ▁Beau
+- ggu
+- ▁proximity
+- ▁circling
+- ▁screech
+- ▁vouchsafe
+- ▁먼저
+- 的一个市镇
+- ▁filosof
+- ▁dennoch
+- ▁schwach
+- ▁quedó
+- ▁endeavoring
+- ▁şey
+- garra
+- bewußt
+- ▁florentine
+- ▁reinforcements
+- 밤
+- ▁уже
+- ▁strayed
+- ▁proprio
+- ء
+- dringen
+- ▁publishing
+- ▁ferocity
+- 母
+- ▁avarice
+- ▁winked
+- ▁pompous
+- ▁nyamara
+- strijd
+- ▁뭐가
+- ▁крок
+- ▁наступны
+- ▁repulsed
+- ▁돈을
+- นิ
+- сія
+- ▁gusoma
+- ▁appro
+- ҙар
+- ޝ
+- ể
+- 取
+- ▁fanciful
+- 地区
+- mbira
+- ್ಯ
+- бле
+- ▁spielt
+- ▁voltaire
+- ▁gable
+- ▁dros
+- รี
+- 野
+- ▁inoltre
+- ▁pagar
+- chino
+- kolebwa
+- ▁봤을
+- morning
+- ▁promen
+- ▁diferencia
+- eña
+- minister
+- টে
+- ▁développe
+- ▁Costa
+- ുന്ന
+- ▁frantically
+- schläge
+- ▁adelante
+- tinya
+- ▁pounding
+- ▁desist
+- ifika
+- чыя
+- ▁undertone
+- ▁hostilities
+- เมื่อ
+- ታ
+- ▁sceptre
+- ক্ত
+- ▁Hund
+- ▁음식
+- ▁adroit
+- ▁verstehen
+- ▁patsy
+- рҭ
+- ▁tilted
+- ▁Bwe
+- ޔަ
+- ▁instituted
+- ական
+- ɣal
+- ▁engineering
+- ▁albany
+- ▁flamme
+- ▁getroffen
+- ▁materially
+- bani
+- ја
+- ▁manuel
+- さ
+- ▁stamping
+- ▁conocido
+- mbaraga
+- ▁Patr
+- ▁chisel
+- 스트
+- лоп
+- ▁southeast
+- ▁swarming
+- omwana
+- grup
+- ▁Ramon
+- somesa
+- ▁Gü
+- ▁sluggish
+- ▁sampson
+- ▁overpowered
+- ▁невялікі
+- ▁Spanish
+- ▁symptom
+- ▁indolent
+- ▁булһа
+- ▁deane
+- ując
+- ستا
+- ภา
+- vět
+- ▁από
+- ▁occupies
+- ению
+- ▁terrier
+- ħa
+- ▁grated
+- ▁pussy
+- ▁watoto
+- ▁insistent
+- ▁armchair
+- reihe
+- ▁medden
+- ตา
+- ▁freak
+- ▁eternally
+- வர்கள
+- ▁spécial
+- ▁дакумент
+- ഗ
+- ▁précis
+- ▁mesos
+- لغان
+- लि
+- ▁веле
+- ▁вочы
+- ▁확실히
+- 께
+- ▁mozart
+- 岛
+- ительно
+- ▁tão
+- ▁다르
+- ▁lakini
+- ▁Nella
+- ▁oído
+- льная
+- ▁iawn
+- ▁stepan
+- ▁chew
+- ▁cleverness
+- ▁maladie
+- ▁Йо
+- ▁wailed
+- ør
+- ▁recited
+- ▁propietat
+- ދާ
+- ▁walikuwa
+- ▁heureuse
+- ▁intimated
+- ▁depicted
+- ▁plug
+- ▁suppression
+- risch
+- ӓн
+- ▁cuales
+- ▁éclat
+- ▁Tial
+- ▁ekzist
+- рана
+- ▁reverent
+- ▁metaphysic
+- ajuda
+- ▁detested
+- អ
+- ▁découvert
+- ▁schicksal
+- ▁لەگەڵ
+- 샀
+- ▁tampoc
+- ▁yelli
+- ▁lazily
+- ▁breit
+- ▁relic
+- ▁usher
+- ▁अस
+- dür
+- ▁Motor
+- ▁downcast
+- ▁estão
+- ▁inseparable
+- ▁loĝantaro
+- ▁oregon
+- ▁adela
+- جد
+- ▁Mach
+- ийг
+- saki
+- ▁dufite
+- ▁beschäftigt
+- 局
+- ▁сабой
+- ▁schwierig
+- ▁cray
+- ления
+- reiten
+- ibihugu
+- чин
+- ▁pamp
+- วา
+- ▁appris
+- ▁awkwardly
+- 社
+- événement
+- ▁இல்லை
+- ▁aggravate
+- ▁zwarte
+- ▁surveying
+- Qua
+- ▁relat
+- major
+- ▁répét
+- latina
+- ▁discerned
+- betrieb
+- лик
+- 높
+- ▁boisterous
+- ▁unchanged
+- ▁meditated
+- лых
+- ▁tush
+- ފު
+- ▁Wohn
+- ▁disordered
+- hib
+- ▁слово
+- ▁impotent
+- لای
+- ▁pratique
+- ▁plural
+- ▁portar
+- ▁fervor
+- ▁Julià
+- kuuma
+- ▁둘이
+- ▁betraying
+- ▁Präsident
+- ▁terstond
+- மைய
+- ▁diferenc
+- ▁kurwanya
+- ▁snort
+- ▁나오는
+- ▁geological
+- stemming
+- ahise
+- ▁upande
+- ▁clerical
+- ▁따로
+- ▁taxation
+- ença
+- ▁arabella
+- ىنىڭ
+- ▁Kenn
+- ▁solcher
+- umukinnyi
+- ▁вышэй
+- સ્
+- ▁rumbling
+- ▁beweging
+- ▁obligat
+- levé
+- ▁langzaam
+- ▁antioch
+- ▁nacia
+- ▁acquiring
+- ▁erfolgreich
+- ހާ
+- ▁niveau
+- episodi
+- spitze
+- ▁stepmother
+- broch
+- mentioned
+- ستم
+- ▁Бына
+- ▁ratio
+- mpton
+- ுக்க
+- ▁Kunst
+- uwd
+- fior
+- åg
+- ▁Wirtschaft
+- mugoroba
+- ▁меня
+- ▁nshya
+- ▁evitar
+- écoute
+- ▁που
+- ▁цар
+- ▁supporters
+- mitima
+- ▁höchsten
+- ▁congratulated
+- ▁nourri
+- ิง
+- schriften
+- ▁conceded
+- ている
+- หนึ่ง
+- ▁sydd
+- ▁Farners
+- ▁wrapping
+- ▁gym
+- ▁bram
+- ▁grandpa
+- ▁sophist
+- ▁hydro
+- 증
+- ▁fotografi
+- ▁giro
+- ▁rencontr
+- قدر
+- ▁sozial
+- ▁niña
+- ▁saisi
+- ▁miembro
+- ▁kunywa
+- glasses
+- lisi
+- ור
+- сыл
+- ▁tumbler
+- คํา
+- ॉ
+- ▁prerogative
+- рес
+- 끼리
+- ▁persistently
+- ▁umutungo
+- ▁ostr
+- awake
+- ▁chlor
+- ▁pferde
+- iéndo
+- ▁mended
+- ▁Беҙ
+- futa
+- ને
+- ▁clapping
+- ▁statisti
+- ▁stanza
+- бере
+- ▁objeto
+- ▁watchman
+- ▁sparing
+- ▁ecclesiastic
+- ▁foaming
+- exdem
+- ▁marchand
+- ▁gevonden
+- ▁strategy
+- 片
+- ▁douleur
+- ▁اتاق
+- gendere
+- ▁Тым
+- ▁charleston
+- chow
+- ▁huck
+- ashyizwe
+- ▁asiatic
+- ▁cael
+- ▁கூ
+- ▁studious
+- ▁licked
+- ▁minerva
+- ▁deplorable
+- wärtig
+- ▁bedeutet
+- ▁persönlich
+- ▁müßte
+- މުން
+- ▁fenwick
+- ▁sickening
+- werfen
+- ▁baldaŭ
+- ▁joueur
+- ▁unwise
+- ▁tariff
+- ▁abhor
+- َعْ
+- ▁süd
+- ▁feina
+- ನ್ನು
+- ▁unaccustomed
+- ▁fireside
+- ▁legacy
+- ▁assassination
+- ▁austere
+- ▁tentative
+- ▁wote
+- ▁выкарыстоўва
+- ▁presta
+- ▁bruno
+- ▁schudde
+- ▁гана
+- vable
+- 끊
+- ▁alvenis
+- ▁pricked
+- ▁tearful
+- ▁huron
+- neil
+- ihariye
+- ▁그런데
+- ▁königin
+- നി
+- gereza
+- ▁precarious
+- 於
+- ▁plotseling
+- ▁adjutant
+- ▁Reich
+- ▁masked
+- ▁ignoring
+- ▁professeur
+- idir
+- ▁scud
+- ូ
+- ▁aeroplane
+- ▁pedagog
+- 造
+- ▁département
+- ▁Reserva
+- ▁hawkins
+- 휴
+- ▁здаецца
+- ▁provocation
+- ▁بده
+- stube
+- grenze
+- Ş
+- 곳
+- 村
+- ▁längst
+- ▁Waliwo
+- ▁pourrait
+- ▁Länder
+- ▁사고
+- ▁corro
+- ▁coche
+- ▁impostor
+- ▁carroll
+- urubyiruko
+- ▁waxed
+- ▁reproduction
+- ▁vollkommen
+- occupa
+- ્ર
+- ▁postponed
+- umweru
+- ▁contraire
+- ▁dritte
+- ▁bloodshed
+- worthy
+- judg
+- shinja
+- ▁humiliating
+- capacit
+- ▁rallied
+- ▁крыху
+- lungi
+- ▁moslem
+- ▁галін
+- ▁Virginia
+- ▁versetzte
+- ▁teori
+- ▁wandte
+- ▁fiscal
+- ▁employing
+- 습
+- ▁luxuriant
+- ▁unsatisfactory
+- ▁ياخشى
+- ېرى
+- ▁constancy
+- ▁glor
+- ność
+- ▁penetr
+- iteguye
+- ▁forgetfulness
+- คล
+- ▁kugura
+- ▁availed
+- zida
+- 第一
+- ▁jose
+- தாக
+- ▁presumably
+- ▁grumbling
+- ▁싶은데
+- ▁kep
+- ▁beguile
+- ▁handsomely
+- ▁unglücklich
+- ▁colin
+- ▁specie
+- ▁Filip
+- ▁задач
+- ▁cuthbert
+- ▁biashara
+- ▁sleek
+- ▁cristal
+- gebung
+- ▁acquitted
+- ▁Northern
+- leɛ
+- ▁oppress
+- duizend
+- ▁distract
+- ▁От
+- сістэм
+- ника
+- َى
+- ▁kangaroo
+- ▁synagogue
+- ▁прынцып
+- ▁emerging
+- ▁hübsch
+- ▁Mireia
+- vloed
+- byeyi
+- ▁맞는
+- ▁dades
+- ▁괜찮은
+- command
+- ▁puss
+- ▁gma
+- ▁கொண்டு
+- ▁foolishly
+- ▁juego
+- bitekerezo
+- ▁quack
+- วง
+- ▁Doch
+- urugendo
+- viennent
+- ▁이러고
+- ▁christina
+- ▁Tür
+- முடி
+- னும்
+- மும்
+- 叶
+- ▁sullenly
+- ▁이러면서
+- ▁cossack
+- ▁penalties
+- ▁élevé
+- ▁Anne
+- ▁tightened
+- ▁campagne
+- ▁bwabo
+- ▁vulture
+- shouldered
+- ▁virgil
+- رخ
+- வள
+- сэн
+- เบ
+- ▁Turn
+- ӓш
+- ishijwe
+- ▁antics
+- ▁boudoir
+- กําลัง
+- ▁abreast
+- ▁모르겠는데
+- ▁sababu
+- tandukany
+- elijkheid
+- ▁monatoj
+- lustra
+- yemeza
+- assum
+- ▁горадзе
+- ▁dorrit
+- ▁nettle
+- iker
+- чную
+- ▁Fan
+- ▁claro
+- чем
+- ▁participar
+- ▁Netta
+- ކީ
+- ığı
+- tambula
+- ▁همین
+- ▁gewisse
+- ▁foolishness
+- ▁aparece
+- ฮ
+- 폰
+- ▁barked
+- ▁tumultuous
+- жда
+- ▁johnston
+- 国家
+- balala
+- ▁dawson
+- ▁ئاي
+- віць
+- ▁вуліцы
+- ▁bothered
+- ԥс
+- ▁llum
+- cheza
+- праграм
+- ▁reprend
+- ề
+- ▁inconceivable
+- ▁vosotros
+- ▁얼굴
+- ▁mislead
+- ▁thrive
+- ▁cheated
+- ▁mightily
+- ▁meses
+- qat
+- ▁fatigued
+- ▁caspar
+- جى
+- ▁verandah
+- ▁algemeen
+- ёў
+- ▁sceptic
+- گشت
+- ▁Congo
+- ▁principalmente
+- هد
+- ▁venera
+- ▁cuyo
+- ▁criado
+- стру
+- ▁wszystko
+- 멀
+- ▁théâtre
+- ▁waterfall
+- رُ
+- kı
+- ▁recede
+- qqel
+- color
+- ▁squall
+- bieten
+- ▁Mehr
+- ▁eugenia
+- ▁erfahren
+- ▁embroidery
+- ▁cleverly
+- ▁Dopo
+- pter
+- ▁sinclair
+- ▁footprints
+- ▁trägt
+- ওয়া
+- ▁outright
+- ▁squeal
+- lector
+- ▁Rang
+- ▁самае
+- ▁тых
+- ▁пур
+- ▁exaltation
+- ▁ҡалды
+- ição
+- ähnlich
+- ▁serena
+- ▁Sprache
+- мір
+- hri
+- ▁pequeña
+- ീ
+- ▁humbug
+- 之一
+- ▁праблемы
+- ▁Zahl
+- ▁нашых
+- sammelt
+- ▁Kara
+- ▁каманд
+- ▁vestit
+- ▁valle
+- ▁Route
+- ▁quickness
+- ▁couvert
+- ▁Kial
+- ▁आणि
+- ▁observant
+- ▁chère
+- interno
+- ▁axis
+- ▁confessor
+- ▁gayly
+- ▁diagram
+- ▁vornehm
+- ستە
+- ▁continuously
+- vich
+- ▁arrivé
+- ▁eaves
+- ▁hulp
+- 능
+- alda
+- ▁disconcerted
+- ▁juncture
+- ▁incarnation
+- глу
+- فل
+- sthetic
+- யம்
+- ▁laughingly
+- article
+- ▁nailed
+- ▁Buri
+- ▁Мі
+- ▁assemblage
+- ▁ceasing
+- ▁ręk
+- ▁Eigen
+- ▁Social
+- ناس
+- ▁그쪽
+- ▁прыклад
+- ▁differing
+- ▁corrupted
+- ▁hassan
+- ▁blickte
+- ▁resolv
+- ▁Мар
+- ▁incoherent
+- ▁straggling
+- ▁mosaic
+- ▁policia
+- ▁고민
+- ▁Aql
+- ▁навуча
+- 너
+- cji
+- ▁assyria
+- ▁auction
+- ▁laurence
+- ▁yiḍ
+- ▁dangling
+- ▁Plej
+- ▁plumb
+- वर
+- ▁Aŭ
+- ▁Guillem
+- ▁pinnacle
+- ▁sanctity
+- ▁withheld
+- ▁finestra
+- ▁갔어
+- ▁classifica
+- ▁Four
+- ہوں
+- ▁comune
+- ▁resisting
+- ▁نق
+- ▁flicker
+- mér
+- ▁беларус
+- ▁badger
+- ▁nkaho
+- ▁glint
+- ▁potomac
+- 퍼
+- ▁disabled
+- ▁présence
+- ▁quitting
+- この
+- ▁wrest
+- ▁legisla
+- US
+- ▁بزن
+- ▁guile
+- ▁semaine
+- ▁tugend
+- ҿ
+- semblant
+- ▁deinem
+- ▁sterling
+- ▁reformed
+- ▁Жы
+- ▁criticise
+- ▁dinero
+- дов
+- ▁característic
+- ▁carretera
+- ▁flattened
+- 드라고
+- ▁boswell
+- ▁exclusion
+- ▁devoir
+- цо
+- vî
+- mugenzi
+- ▁compelling
+- ▁incidentally
+- teganya
+- ▁agatha
+- ▁Gel
+- ▁schwarzen
+- ▁новыя
+- ▁Zum
+- ▁facilitate
+- ▁tenido
+- ▁fitness
+- hivern
+- ▁droll
+- ▁niche
+- unze
+- Ü
+- letto
+- ▁encroach
+- ஐ
+- ▁쪼금
+- ▁flexible
+- ▁foretold
+- ▁частцы
+- ▁kwizera
+- ▁voeten
+- ▁fitful
+- gurira
+- ▁уке
+- ▁simile
+- ▁хәл
+- ▁kennel
+- ▁applauded
+- möglich
+- ▁graphic
+- fleisch
+- نَا
+- ▁Пу
+- ▁semana
+- вали
+- arzt
+- ▁anxieties
+- ▁fluctua
+- ▁olhos
+- ▁steak
+- ▁wintry
+- த்தின்
+- ▁pecho
+- ▁snail
+- umbre
+- ▁vreemde
+- ▁venga
+- ▁사는
+- ത്
+- ▁rigorous
+- ▁chatting
+- ▁боло
+- 쉽
+- ▁немагчыма
+- 便
+- ▁wiederholt
+- kadde
+- ▁gereed
+- பின்
+- èria
+- স্ট
+- ▁strolling
+- ▁comical
+- 밥
+- ▁Test
+- ▁okw
+- 라면
+- ▁Brand
+- ▁rodolphe
+- ▁бяс
+- ▁Tin
+- ცი
+- élection
+- zika
+- icamente
+- mette
+- łę
+- ▁ekzemple
+- ubucuruzi
+- ▁intoxicated
+- ▁مثل
+- ▁foresaw
+- ▁Produkt
+- ▁malgrandaj
+- ▁safeguard
+- ▁канца
+- głos
+- isait
+- ▁llevaba
+- ▁bewildering
+- ▁huguenot
+- ▁unwonted
+- akati
+- 배
+- ▁shaved
+- ▁razor
+- ▁befell
+- ▁philosophic
+- ▁ntashobora
+- ▁omission
+- ▁retaining
+- khu
+- ▁Бул
+- ▁lionel
+- ▁sufiĉe
+- ▁compañero
+- ▁leopold
+- cello
+- abaturage
+- ▁Нават
+- ▁binne
+- 생활
+- atac
+- ▁дух
+- ої
+- ▁Nieder
+- grä
+- ▁rivalry
+- ▁hyde
+- метр
+- weight
+- èche
+- organise
+- ພ
+- dokument
+- แบบ
+- 源
+- ▁contingent
+- ▁galilee
+- 可能
+- ▁прыма
+- ▁cyumweru
+- ▁Jetzt
+- uchte
+- ▁Hano
+- ▁пайшоў
+- ▁заўва
+- ▁acerca
+- arrel
+- ▁ignora
+- სა
+- ▁housemaid
+- ▁contending
+- gift
+- ▁meditating
+- ▁provoking
+- 션
+- ▁아닌데
+- ▁Вер
+- '500'
+- ▁savait
+- ▁specifically
+- hindura
+- zamuka
+- ▁verließ
+- யோ
+- iɛ
+- кӧ
+- hook
+- ▁martian
+- ▁Акрамя
+- ▁architectural
+- begotten
+- ▁mantelpiece
+- ஞ
+- ▁diente
+- ▁frightfully
+- ▁마시
+- ött
+- ▁Wil
+- ้ง
+- ▁amply
+- sprung
+- ▁weiteren
+- важа
+- ▁кары
+- ▁puzzling
+- ▁kerak
+- ▁ministro
+- ▁pensamiento
+- unterricht
+- ▁shaving
+- นี่
+- ▁mohammedan
+- ▁옆에
+- ี้
+- discrimin
+- ▁derecho
+- თა
+- ▁redoubled
+- ▁Déu
+- ssant
+- ▁sabia
+- بط
+- ใ
+- яўляюцца
+- ▁Yessefk
+- ▁mahogany
+- ▁schweigen
+- ▁swedish
+- ▁proporciona
+- няя
+- ών
+- ▁retira
+- rigg
+- ▁separating
+- ▁monopoli
+- ەکەم
+- ▁சில
+- ▁사이
+- ▁деч
+- 氏
+- çant
+- ası
+- ▁mão
+- ▁alban
+- effa
+- ▁Leben
+- cele
+- ▁pluto
+- ▁astounded
+- ▁unlawful
+- édition
+- ▁kečč
+- kenewe
+- зла
+- ▁crafty
+- ▁correspon
+- powiedz
+- ▁cherub
+- ▁выраб
+- Š
+- ▁хачу
+- ▁bullock
+- ▁leider
+- ▁então
+- ▁protesting
+- eẓra
+- ▁elias
+- ▁Valley
+- ▁tendre
+- น่า
+- larning
+- ▁நாடக
+- 극
+- ▁Card
+- 游
+- kämpfe
+- ▁suivant
+- giriye
+- ▁stoutly
+- насцю
+- ▁fremden
+- ▁demeure
+- ▁boasting
+- ababa
+- ▁operates
+- ▁بازی
+- ▁schle
+- ▁그랬어
+- วั
+- ▁mkuu
+- өм
+- ▁трыма
+- ़
+- ▁audacious
+- ▁covetous
+- ▁statistics
+- ▁dismayed
+- lynn
+- ▁suspiciously
+- ▁conveyance
+- brä
+- amant
+- nię
+- өрө
+- ▁асноўным
+- ▁lilac
+- ▁quantitat
+- ▁oczy
+- ▁franchise
+- selmad
+- ▁있어서
+- ▁matern
+- ▁unbearable
+- ▁jeffrey
+- ▁Wakati
+- ες
+- нуцца
+- ▁Рэ
+- wyth
+- ▁kehrte
+- ohereza
+- ▁harassed
+- genzi
+- ▁fervently
+- ▁compon
+- ▁끝나
+- ▁acest
+- ▁commodore
+- ▁implacable
+- ▁percentage
+- ▁پیدا
+- ию
+- ▁olma
+- ikiriza
+- errant
+- ▁lluita
+- ▁좋아해
+- ▁دەپ
+- ҿы
+- варот
+- ▁يوق
+- ▁terrifying
+- ▁premio
+- ▁panama
+- ▁obstruction
+- ଦ
+- ▁charakter
+- forgetful
+- wurf
+- 했거든
+- ▁imperfection
+- เด
+- ▁Italian
+- ▁swarmed
+- ▁uneven
+- ▁муж
+- ▁barney
+- wold
+- ىشى
+- chiq
+- ށ
+- ▁econòmic
+- ▁Беларус
+- ▁gaarne
+- ▁ehemalige
+- educated
+- ▁mesura
+- ▁Kö
+- ▁Stelle
+- ▁jesse
+- ภ
+- ▁indefinitely
+- ▁organize
+- ▁Schwa
+- чку
+- ibanze
+- іна
+- kulembez
+- objectiu
+- ▁fácil
+- ▁музыкі
+- kanwa
+- ▁drone
+- ▁atomic
+- वे
+- ▁obuta
+- sòl
+- huka
+- ▁defender
+- verwandt
+- 잔
+- ▁blurred
+- Opera
+- ▁Look
+- ▁тудо
+- ęż
+- beshya
+- ▁leeren
+- ▁bombi
+- ович
+- ▁puffing
+- ▁кня
+- ▁pestilence
+- ▁unintelligible
+- wasser
+- ▁toestand
+- ▁analogous
+- ▁carnal
+- ▁herzog
+- ▁під
+- ▁tying
+- ▁mogen
+- ▁becky
+- ▁llista
+- ▁réuni
+- ▁Ин
+- ▁windsor
+- ▁eenigszins
+- ▁intrinsic
+- ▁knuckle
+- ▁satisfactorily
+- 落
+- ▁elijah
+- ▁değil
+- ▁scold
+- ▁presided
+- шаш
+- ▁داره
+- ▁cinque
+- ▁Isi
+- dług
+- ▁انجام
+- nyubako
+- พวกเขา
+- 客
+- warte
+- ▁кажуць
+- ىتى
+- dollar
+- ▁chink
+- ▁reprove
+- ▁nähe
+- ▁borst
+- থা
+- loof
+- され
+- ▁sard
+- ▁vivacity
+- ▁сусед
+- ▁respondis
+- ▁brutus
+- ▁moody
+- oiseau
+- ▁ferma
+- ▁fleeing
+- ▁frans
+- ▁outwardly
+- keza
+- ▁científic
+- ▁unwelcome
+- ▁trustworthy
+- ▁París
+- ▁улар
+- ▁шулай
+- sacrificing
+- ніку
+- ▁Good
+- ▁goldsmith
+- eetaaga
+- ▁будучы
+- ▁midway
+- ▁gravy
+- ▁harrington
+- ▁tercera
+- ▁Imir
+- დე
+- ▁moh
+- ▁draußen
+- ҕ
+- ▁kathleen
+- 育
+- ▁georg
+- ▁rajah
+- ▁voyez
+- ▁Titel
+- ▁einiger
+- 보다는
+- first
+- century
+- ሰ
+- 숙
+- ▁capricious
+- 呢
+- ▁aandacht
+- ӓл
+- ▁timidity
+- гле
+- သော
+- ▁caval
+- logique
+- ▁messieurs
+- ▁Johnson
+- ▁povo
+- ▁rosalind
+- ▁doorstep
+- ▁hieß
+- 돼
+- λο
+- ▁terrain
+- riez
+- ĉambro
+- ▁wesentlich
+- fisch
+- ▁bradley
+- ▁الان
+- ▁calico
+- ▁aerial
+- ▁nathan
+- connor
+- ▁parasite
+- ตาม
+- ▁swiftness
+- lje
+- ▁periodista
+- ▁Avenue
+- ▁disitulikiti
+- 走
+- ▁diplomacy
+- ρε
+- juring
+- ström
+- ▁restoring
+- ▁Parizo
+- ▁еще
+- ▁faccia
+- ρί
+- ▁scolded
+- ▁scolding
+- ▁Over
+- ▁willingness
+- günstig
+- ▁шинча
+- 승
+- ▁stammt
+- фра
+- ▁roller
+- ▁profuse
+- ▁Ferran
+- ▁debtor
+- ▁foresight
+- ▁كې
+- ▁McC
+- َنْ
+- ▁kumw
+- 들을
+- ▁barefoot
+- 봉
+- ▁zufrieden
+- ▁stratagem
+- ▁synonym
+- ▁ballet
+- ▁compliance
+- ▁bezonas
+- گەر
+- ▁fasado
+- ▁fala
+- ▁propia
+- ▁forthcoming
+- ຍ
+- ▁rudder
+- يۇ
+- estiu
+- patikana
+- approche
+- ▁pista
+- ουν
+- ▁caer
+- yec
+- schafft
+- zimu
+- ▁Яшчэ
+- 每
+- ▁reflex
+- ▁preparatory
+- ▁якасці
+- лот
+- ▁embodied
+- exécut
+- ▁быть
+- ▁چەند
+- ▁mkubwa
+- 몰
+- vith
+- huru
+- oppo
+- ệ
+- ▁overhanging
+- 政府
+- ▁aesthetic
+- ▁pageant
+- ètica
+- ▁lagoon
+- ▁comple
+- satisfied
+- த்திர
+- ▁hurd
+- ▁funció
+- ▁battlefield
+- ▁grati
+- ▁koloro
+- ▁Komp
+- جَ
+- babwira
+- ▁Academy
+- ▁aproxima
+- ▁inexorable
+- ▁невядома
+- 짐
+- schluß
+- 华
+- ▁pooh
+- ▁Melmi
+- wiɣ
+- ▁miedo
+- rizo
+- ▁стары
+- ngwe
+- ▁sebastian
+- ▁wrench
+- 류
+- ▁friendliness
+- verbindung
+- ▁저기
+- ashize
+- ibiti
+- ▁doleful
+- ḍan
+- ▁kutokana
+- digd
+- ▁праўда
+- ▁venas
+- ▁applaud
+- ▁краін
+- ▁gewone
+- ▁Forma
+- 인이
+- 떠
+- zerstör
+- ▁bobbsey
+- garagaje
+- ▁може
+- ▁없다
+- ▁liberality
+- ▁overthrown
+- ▁казаць
+- ▁israelites
+- ábamos
+- ▁arbust
+- ▁barrow
+- ▁brewer
+- ▁personatge
+- ▁çi
+- дчас
+- ▁raving
+- elimin
+- ▁posi
+- ▁pisa
+- 촌
+- ▁дапамог
+- indege
+- ▁courteously
+- ▁llança
+- ▁Latin
+- fika
+- ▁heiligen
+- ▁Йә
+- ▁unanimous
+- 請
+- ▁formació
+- முத
+- ▁дазвол
+- ▁schrecklich
+- ▁commended
+- ▁inexperienced
+- ▁caza
+- ações
+- ▁Grau
+- ▁pregunt
+- ▁ствара
+- дү
+- ▁애가
+- demonstr
+- шып
+- ▁geneva
+- ▁charlemagne
+- ▁shrivel
+- ▁tränen
+- ▁cascade
+- ▁tyrann
+- ▁pianist
+- ▁cleek
+- ▁없는데
+- ▁ўдзел
+- رض
+- ▁scorned
+- ▁unmöglich
+- 當
+- ▁wharton
+- ▁lobster
+- ▁verhaal
+- ▁réussi
+- ▁tidak
+- ▁presencia
+- ▁drawback
+- Т
+- ▁bacu
+- ▁Кар
+- ▁elaine
+- ▁pleine
+- flow
+- ▁broil
+- ▁hullo
+- plaça
+- ਇ
+- ▁connecticut
+- ▁impediment
+- ▁overturned
+- ▁ancestral
+- ▁oblique
+- ▁dived
+- мәгән
+- ▁lessened
+- ▁godmother
+- ▁inspected
+- ▁typically
+- ▁construit
+- ▁Senat
+- ▁moje
+- ▁niagara
+- ल्या
+- wią
+- ▁beneficent
+- ▁cipher
+- leiten
+- ▁maritime
+- shyiraho
+- ரும்
+- ທີ່
+- ▁adherents
+- ▁obertures
+- ▁flounder
+- ▁civilised
+- Sch
+- ני
+- ▁idées
+- ▁growling
+- лась
+- 끄
+- 識
+- ▁sinjoro
+- ḥi
+- вядзе
+- steuer
+- ▁tipped
+- ▁prag
+- ▁surname
+- ▁diversity
+- ▁aperas
+- ɣeḍ
+- ▁financier
+- ▁Insel
+- ື
+- ▁Bedeutung
+- ▁dissipated
+- ▁gikorwa
+- ▁antiguo
+- accueil
+- верс
+- ▁selves
+- ▁Totes
+- ▁unglück
+- ▁Zug
+- ▁conformity
+- গা
+- ▁publique
+- ▁blev
+- ▁einzelnen
+- idge
+- ىگە
+- ▁рэгіён
+- 착
+- ▁irritable
+- તા
+- ▁priceless
+- ▁ڕۆژ
+- ▁biddy
+- ங்களை
+- ▁gaston
+- ▁glimmering
+- шок
+- ▁urboj
+- ▁އޭނާ
+- ▁antelope
+- ▁grooten
+- ▁occult
+- ▁loathsome
+- habitude
+- ▁randolph
+- ▁mournfully
+- ▁calais
+- ḍeɣ
+- ▁restlessness
+- ▁fleece
+- ਅ
+- ทุก
+- ▁물론
+- ▁nuovo
+- bottle
+- ▁razón
+- ▁lachend
+- ▁socket
+- ंत
+- ▁건가
+- ▁abolish
+- ▁refl
+- yanyu
+- ▁caleb
+- ▁kunlabor
+- અ
+- ▁توانید
+- ▁froh
+- ▁navarre
+- ▁detall
+- ▁konsiderata
+- ڵی
+- ▁чтобы
+- ț
+- anunci
+- ▁crank
+- போல
+- foil
+- ▁nyingine
+- ▁danish
+- ▁imatge
+- ೊ
+- ▁particulier
+- ▁suspension
+- ▁encompass
+- ▁хвілін
+- ▁zoeken
+- ▁carrington
+- تاب
+- ▁лічы
+- خۆی
+- ▁cobert
+- ▁Encara
+- ssait
+- ▁directeur
+- zahlen
+- advocat
+- ▁voraus
+- ▁بىز
+- ▁일을
+- ▁livelihood
+- ▁dispar
+- ▁뭐냐
+- ▁zorn
+- guka
+- État
+- саў
+- ▁beauté
+- தம்
+- ▁perturb
+- ▁sabía
+- thier
+- ишет
+- ▁tutlayt
+- ▁저녁
+- Hospital
+- ื่น
+- ▁erzählen
+- bavubuka
+- аваныя
+- genossen
+- existence
+- ▁newgate
+- integra
+- ▁composing
+- stimul
+- ▁traje
+- efecte
+- ▁jerome
+- ▁whaka
+- өҙ
+- kład
+- ▁பூ
+- ▁Public
+- gregor
+- ▁mutually
+- لُ
+- extrême
+- ▁Zanzibar
+- ▁Mundial
+- ▁재미
+- 風
+- ▁того
+- ▁brenn
+- favor
+- ▁scheu
+- ▁بودم
+- ▁caràcter
+- ▁vacancy
+- ▁savi
+- шьа
+- ▁demostra
+- ▁volgenden
+- ▁아니면은
+- ▁blithe
+- вымі
+- ▁copious
+- ▁konstrui
+- 服
+- யான
+- ▁iedereen
+- entrer
+- ▁glob
+- bilder
+- ▁momentous
+- blooded
+- defined
+- ▁Ли
+- ▁mutiny
+- ▁devouring
+- ggett
+- ▁blinking
+- ▁selecting
+- ▁الع
+- ▁digestion
+- cini
+- rów
+- ▁өз
+- meaning
+- ▁yambaye
+- ▁kigo
+- ▁При
+- ямі
+- ▁안에
+- fabrik
+- ▁unjustly
+- िया
+- _
+- ▁مادر
+- ▁قبل
+- ที่จะ
+- ▁royalist
+- ▁amplia
+- ▁rigor
+- bidde
+- varen
+- ▁беларуска
+- ▁drunkenness
+- 뜨
+- 베
+- 计
+- ▁gresham
+- ▁вызнача
+- strecke
+- ▁destinies
+- ▁маё
+- ▁가면은
+- ▁tiro
+- ▁친구가
+- дадзены
+- ▁Maybe
+- ensimbi
+- remy
+- ▁ryose
+- morf
+- ▁prostitute
+- ▁zwanzig
+- hàbit
+- đ
+- 具
+- ▁چقدر
+- ▁shameless
+- ▁defective
+- ▁budget
+- zijn
+- ▁melhor
+- ▁larmes
+- ▁Icya
+- ▁Dé
+- ▁markham
+- คา
+- ڏ
+- ፣
+- ▁unbekannt
+- ▁혹시
+- ▁dilemma
+- ▁spitz
+- ologische
+- ▁орган
+- ▁pounce
+- ▁legislator
+- ▁Праз
+- لۇق
+- ▁Francis
+- ▁lingua
+- ología
+- ▁translate
+- freund
+- ▁드라마
+- ▁coughed
+- ▁mitjança
+- ▁infinity
+- ▁στο
+- ▁instructive
+- ▁robbing
+- ▁jardí
+- ▁cleanliness
+- saidia
+- ▁älter
+- ísima
+- ▁occup
+- ▁Emil
+- ▁nächste
+- Oce
+- ド
+- ▁applicable
+- ▁rheumatism
+- ▁territoire
+- 萨
+- 从
+- ▁größere
+- ▁தலை
+- رة
+- ร์
+- ▁Ehe
+- هایی
+- abbé
+- völker
+- ृ
+- дзяржаў
+- ▁thermometer
+- ლებ
+- ▁Kind
+- ▁gemüt
+- ▁widened
+- ▁discours
+- اصل
+- 存
+- ▁چۈش
+- adura
+- ▁muller
+- reaching
+- ▁bashobora
+- ▁parched
+- ▁giggle
+- ɣar
+- ▁exposing
+- بوون
+- geordnet
+- ▁eblas
+- ผล
+- abanyeshuri
+- ▁lebhaft
+- ▁stiffened
+- ский
+- adresse
+- ▁okufuna
+- تور
+- czego
+- ▁subjective
+- pasi
+- ▁Since
+- ▁Our
+- ▁gefahr
+- 빨
+- ▁dificult
+- ▁buddhist
+- 電
+- ▁analyze
+- ▁licence
+- ▁Geld
+- ▁furchtbar
+- ▁appropriated
+- schade
+- ▁부분
+- ӣ
+- 북
+- ▁тэатр
+- بات
+- ▁ensuing
+- ▁collo
+- ▁Thi
+- ▁groped
+- ▁spurred
+- ▁அவர
+- ▁Klein
+- accord
+- ▁Hast
+- прос
+- ▁bryan
+- ▁splashed
+- тэры
+- umwana
+- ▁gemakkelijk
+- ▁proksime
+- કે
+- ▁crackling
+- ▁veneration
+- ▁metaphysical
+- ▁Passa
+- ▁얘기하
+- amore
+- ▁yahise
+- sätze
+- ▁refrained
+- weḍ
+- ▁Wolf
+- ғо
+- ▁Бар
+- ▁Fest
+- stoßen
+- ▁arrogance
+- ▁odysseus
+- ▁crumbling
+- ▁caius
+- хвал
+- avocat
+- λα
+- ۇرۇ
+- ▁ponder
+- лығы
+- Wei
+- ▁알았
+- ઓ
+- ▁intrusted
+- ▁tangible
+- 应
+- ▁bombard
+- આ
+- ▁mallonga
+- ▁asteroid
+- ▁renewal
+- kontakt
+- ▁woollen
+- ுவது
+- ▁qualitat
+- ▁froth
+- ḥeɣ
+- ilgan
+- gelegd
+- ൾ
+- ▁Anschließend
+- ▁hetzelfde
+- ▁infallible
+- ろ
+- 奥
+- ▁erinnert
+- ▁umher
+- geraho
+- driver
+- ▁seguro
+- streich
+- ▁při
+- ▁teraz
+- ▁같고
+- ▁revista
+- ۆر
+- ▁brutality
+- ▁supra
+- ▁terminate
+- ▁sonata
+- ގު
+- ▁сэнс
+- унда
+- spoon
+- み
+- 講
+- сол
+- ▁frugal
+- هِ
+- żon
+- ▁Как
+- ▁gevoelde
+- ▁Tab
+- ▁crema
+- رات
+- ▁supple
+- ▁sembra
+- ▁ursprünglich
+- ▁verborgen
+- 才
+- ▁cambio
+- lishi
+- ▁clinch
+- ▁Hotel
+- ▁spon
+- ær
+- ▁violated
+- ▁forecastle
+- లో
+- ▁approximately
+- ▁legally
+- ▁tableau
+- ifier
+- ରେ
+- erekeza
+- ▁матч
+- ▁espada
+- ▁Nicht
+- ଲ
+- ▁República
+- ▁pregnant
+- ▁shrubbery
+- ▁unarmed
+- ▁journée
+- ুল
+- partij
+- ▁yenye
+- tegeeza
+- evolució
+- ▁ڕوو
+- ▁voorkomen
+- вши
+- ▁Jana
+- isola
+- ▁Brook
+- ribb
+- ▁übrig
+- consciousness
+- ▁meditate
+- alliber
+- ▁Kolo
+- 說
+- ▁Tek
+- ▁grootste
+- ▁الب
+- ▁povus
+- eggwanga
+- ▁wrestling
+- λι
+- buja
+- ▁Creek
+- ▁coveted
+- ▁sourire
+- ▁oblong
+- ▁알지
+- ▁starb
+- ্ৰ
+- chev
+- 樣
+- 土
+- ▁따라
+- ▁сезон
+- ▁Yiwen
+- ▁работы
+- ▁kruroj
+- ▁Vallès
+- ▁entreaty
+- ▁Manuel
+- ▁danach
+- ▁chinaman
+- ▁recog
+- ▁என்பது
+- жӹ
+- icyubahiro
+- ▁anarchy
+- ▁trend
+- peta
+- ▁maul
+- ▁mwezi
+- ▁popolo
+- avance
+- ▁gallantly
+- ސ
+- ▁resounded
+- ▁acemma
+- ▁schnee
+- ▁repaid
+- ▁klug
+- ▁завод
+- ▁sequel
+- ▁feliz
+- ▁bunting
+- иться
+- ▁فرا
+- ▁scissors
+- hartig
+- ▁harvard
+- ிருக்க
+- ▁dyed
+- ▁nominated
+- ▁advert
+- دانی
+- ▁icumi
+- ▁servile
+- ▁occupant
+- ▁reducing
+- ▁cron
+- länge
+- вр
+- ▁wavered
+- épi
+- ▁accede
+- ▁Kata
+- ▁semejante
+- ▁mujyi
+- ▁паста
+- ไหม
+- ▁candlestick
+- ▁voyant
+- ▁heightened
+- ▁juliol
+- ▁twitter
+- ungura
+- tuufu
+- ▁evermore
+- rasse
+- ▁hört
+- ▁scramble
+- ▁zili
+- ▁fastidious
+- ▁cannibal
+- ▁infection
+- ▁anarchist
+- ներ
+- ▁neef
+- জি
+- ▁sueño
+- finita
+- ▁admiralty
+- ಕ್
+- ▁respectability
+- maan
+- ţi
+- ▁disparu
+- படுத்த
+- 级
+- ▁laissé
+- ▁краіна
+- ически
+- educació
+- ▁embryo
+- 하자
+- ▁아니지
+- ▁Bou
+- ▁wierp
+- ▁Mutter
+- ▁sanders
+- ▁pisos
+- 活
+- Τ
+- ▁peregrine
+- ▁усіх
+- ▁thrush
+- ▁scipio
+- ▁droeg
+- ▁clustered
+- ▁unused
+- 번에
+- ▁fianc
+- ▁viento
+- spezi
+- panda
+- ▁Ekzistas
+- ▁gloucester
+- ▁rehearsal
+- ▁зямлі
+- ▁considéré
+- ▁hermosa
+- ▁drudge
+- ▁bête
+- ▁withdrawing
+- ▁paltry
+- ▁supervision
+- ▁widespread
+- ▁stowed
+- ▁ascetic
+- മാ
+- ▁кас
+- Plat
+- ▁casket
+- โร
+- trokken
+- lako
+- ▁brat
+- ▁brûl
+- 與
+- ކާ
+- ▁betook
+- ▁roemische
+- ▁старо
+- ▁Board
+- ▁numberless
+- ▁buggy
+- gène
+- ▁jakie
+- аводле
+- leuchtet
+- why
+- ▁ensign
+- ▁havoc
+- ▁mwene
+- ▁stimulate
+- ▁Fol
+- ▁Kuna
+- qbayli
+- ▁무조건
+- ով
+- முறை
+- ▁яҙ
+- ▁간다
+- ▁Scott
+- ▁blinked
+- ▁joyously
+- ▁airship
+- लो
+- füllt
+- ▁greu
+- ▁authoritative
+- ▁bungalow
+- ▁deficiency
+- ▁indictment
+- 置
+- ▁누나
+- ▁witwa
+- petu
+- voller
+- ▁쓰는
+- ▁Pac
+- ▁bramble
+- ▁germà
+- exprim
+- альных
+- ▁treurig
+- ▁cabal
+- ▁Krom
+- ▁blumen
+- ▁Off
+- ▁preside
+- ▁Магчыма
+- ▁unpopular
+- ▁atrocious
+- ▁маленьк
+- czne
+- havia
+- られ
+- ▁flapping
+- 하는데
+- ▁Sí
+- হি
+- gifite
+- ▁comprar
+- дән
+- ▁clem
+- ▁survival
+- ▁насельніцтва
+- ▁jennings
+- 始
+- ▁vorhanden
+- ▁movimiento
+- ▁mynheer
+- ▁segueix
+- ijnen
+- ▁scraped
+- ▁Reka
+- ელი
+- ▁dóna
+- ▁nimmer
+- ▁somethin
+- ▁enriched
+- ▁firmament
+- 딩
+- ▁website
+- ње
+- ▁philosophie
+- rozum
+- ▁magnus
+- ▁barbarian
+- umutima
+- ▁sharpened
+- oportun
+- ކ
+- ەرە
+- ▁loaves
+- ▁zuvor
+- லே
+- broken
+- ▁dismount
+- ։
+- ڊ
+- ▁països
+- އާ
+- 堂
+- ither
+- ▁olivia
+- ▁könnten
+- ▁landmark
+- owski
+- solut
+- ▁ferguson
+- livyo
+- ▁pondering
+- ▁دڵ
+- ▁தீ
+- егеҙ
+- ▁jarojn
+- ropa
+- ▁dissipation
+- ▁Junior
+- ләй
+- ▁emigrants
+- ▁esperanza
+- ▁coordina
+- acak
+- ▁мене
+- ök
+- ▁kultura
+- rwana
+- ibita
+- ▁Пакуль
+- 곤
+- 牛
+- ▁infamy
+- woon
+- ▁flushing
+- ▁gestern
+- ▁mortality
+- ▁consell
+- ▁nellie
+- ҙө
+- ▁geweld
+- ▁portico
+- ▁obstinately
+- ▁kubanga
+- ▁аўта
+- ▁snub
+- ▁besuchte
+- original
+- небудзь
+- ▁tapestry
+- ▁хутчэй
+- ▁innovation
+- ▁restlessly
+- ▁runaka
+- ▁رسید
+- crust
+- ▁quarrelling
+- ▁kufanya
+- ▁blotted
+- ára
+- ▁lamentable
+- ▁zeven
+- ▁Omuntu
+- Eulàlia
+- ▁jufvrouw
+- ▁drenched
+- ▁límit
+- ▁embers
+- kenntnis
+- ▁miteinander
+- ▁bowels
+- ahandi
+- ▁originality
+- ▁теле
+- ▁oscil
+- ▁reunion
+- ▁spartan
+- legio
+- ▁exhibiting
+- ែ
+- ▁disconsolate
+- ▁жывёл
+- ▁insurgent
+- ▁unbounded
+- မှ
+- ▁equilibrium
+- ▁Programm
+- ▁jamás
+- ▁Тэ
+- ▁vertrouwen
+- ▁okubeera
+- ▁verdient
+- qel
+- Liber
+- brique
+- mbuga
+- ▁exhortation
+- ▁roh
+- ▁politie
+- ▁tribal
+- ▁deceiving
+- ▁gunpowder
+- 研究
+- ▁detestable
+- hê
+- ▁besuch
+- ▁fazer
+- ▁airport
+- ▁vinger
+- σα
+- ▁leafy
+- ecaj
+- ▁vertoon
+- energia
+- ▁rusa
+- ▁Rais
+- ĵoj
+- ण्या
+- ▁calamities
+- ▁caterpillar
+- ▁swagger
+- ▁renkont
+- ▁isaiah
+- ▁딱히
+- ▁Volks
+- 자리
+- ▁verlangt
+- slavery
+- ығыҙ
+- ▁ruffled
+- һына
+- ▁fluent
+- ▁idolatry
+- ▁practising
+- 根
+- 만큼
+- ▁tito
+- weka
+- ▁clase
+- ▁aferoj
+- ▁deplore
+- ▁faisant
+- window
+- ▁interesse
+- біт
+- ▁midsummer
+- ▁remporte
+- ▁Manchester
+- ▁sybil
+- ▁framework
+- ▁legde
+- ▁thun
+- ▁shrub
+- ั่ง
+- ▁ulls
+- ▁Henderson
+- த்திற்கு
+- ▁strak
+- ▁vicomte
+- ది
+- ▁feia
+- ▁recipe
+- ▁Eble
+- ▁inferred
+- ▁wengine
+- دَ
+- ರು
+- samkeit
+- ▁рэчы
+- зіць
+- ▁urwo
+- весці
+- rending
+- ચ
+- ▁compañía
+- ▁몰랐
+- ▁smoky
+- ▁Frauen
+- afashwe
+- ▁plurajn
+- ▁underlying
+- ▁ijwi
+- pełni
+- ▁Йы
+- ▁parut
+- ▁advertised
+- ▁locust
+- ▁llena
+- trzyma
+- ▁necessari
+- hti
+- ▁giraffe
+- ▁locate
+- örd
+- туу
+- ▁Mike
+- ▁localities
+- ▁clod
+- ና
+- ክ
+- ▁protagonist
+- ▁Guerre
+- ▁convulsive
+- ▁ведаў
+- '90'
+- erkannt
+- ▁poche
+- ▁copperfield
+- ಿಸ
+- ▁heilige
+- ▁cleave
+- نِ
+- ▁enthusiast
+- ணை
+- лось
+- Pro
+- ▁chile
+- ▁blair
+- ными
+- ▁tranquille
+- ggins
+- ▁ہم
+- ▁senzill
+- лівы
+- ▁Dorf
+- mushinga
+- selig
+- ባ
+- ▁anatomy
+- 편
+- ▁menjar
+- мом
+- ▁verbeeld
+- ▁músic
+- ▁trank
+- ▁puisque
+- ộ
+- ▁disapproval
+- ▁disclosure
+- ▁persuasive
+- ▁unobserved
+- 武
+- މެ
+- ▁Questo
+- дең
+- ▁jingle
+- ▁yabonye
+- ▁motiu
+- anstalt
+- ▁calves
+- ▁Tier
+- ▁hemisphere
+- ▁сюды
+- ▁гатовы
+- ▁frisk
+- ▁زمین
+- ▁benannt
+- ▁Зараз
+- ▁Jud
+- ▁spotless
+- ▁якую
+- ▁corse
+- ▁purport
+- ▁drip
+- ▁cinder
+- 県
+- ▁panorama
+- ▁wriggle
+- ▁employees
+- ▁моцны
+- ▁forenoon
+- bonera
+- ▁kleur
+- رۆ
+- ▁negli
+- ▁mpan
+- ▁byonna
+- tzea
+- ▁llac
+- ▁gloss
+- தில்
+- drang
+- ▁embellish
+- ▁arduous
+- julikana
+- talk
+- ▁repeal
+- ▁jagged
+- ▁만들어
+- geschw
+- ▁mêl
+- ▁ravish
+- ▁Baldaŭe
+- ▁основ
+- гәы
+- ▁impracticable
+- ▁invaluable
+- ਸ਼
+- цам
+- ▁transmission
+- бей
+- ▁querido
+- ▁combien
+- ▁ukuntu
+- ▁rosto
+- аграфі
+- өк
+- wett
+- distinguishable
+- crates
+- ▁pintura
+- étranger
+- ျ
+- ▁unearthly
+- ▁vaixell
+- ▁louisiana
+- setzte
+- ▁reproof
+- দেশ
+- ▁identific
+- ▁Fiona
+- ▁ritter
+- vuka
+- ▁ئوي
+- kundig
+- مانی
+- ▁넘어
+- ável
+- ▁arroj
+- ▁тый
+- ▁Schiff
+- బ
+- ▁abomination
+- ▁stalwart
+- fitanye
+- ▁decayed
+- ▁deteriora
+- 었잖아
+- ▁murugo
+- ▁jules
+- ▁serai
+- ▁wär
+- abayeho
+- ▁implant
+- ▁krui
+- ▁clambered
+- ▁matrimony
+- 습니다
+- 像
+- ▁Hung
+- ▁stirrup
+- ▁dispensation
+- träger
+- ▁corpus
+- جو
+- ▁vanishing
+- ▁recol
+- ▁okulaba
+- ęci
+- mvugo
+- ಾಗ
+- ▁physique
+- ▁Wor
+- ▁Сёння
+- 示
+- ▁stealthy
+- ▁reeled
+- ▁iceberg
+- ▁pantry
+- ▁Czy
+- ▁evergreen
+- ▁impunity
+- ŵ
+- ▁가끔
+- ▁inanimate
+- ▁erwartet
+- ▁paarden
+- aŭri
+- లు
+- тыкі
+- ▁موقع
+- ▁intrepid
+- тлумач
+- ▁exercising
+- ▁akamaro
+- ▁plupart
+- ▁reverently
+- birwa
+- ▁pulp
+- ▁successively
+- ▁crypt
+- ҳо
+- ไทย
+- ▁kensington
+- ▁constitue
+- rindi
+- ▁dentist
+- vallig
+- ▁incorporated
+- ▁retail
+- ▁каманды
+- yemera
+- ▁cumpli
+- ▁utiliza
+- ▁Мә
+- ▁stability
+- ▁novaj
+- ุม
+- kiriye
+- கம்
+- ▁Catholic
+- ▁прафесі
+- ▁accusing
+- ▁telkens
+- ▁Vari
+- ▁sussex
+- ság
+- ▁hauteur
+- ▁identification
+- க்கா
+- ▁Mazal
+- ▁zauber
+- ▁relaxation
+- ▁unuaj
+- Europa
+- gefährlich
+- ▁batch
+- 얼
+- Association
+- ▁devonshire
+- ▁voluntad
+- 无
+- ▁ваҡыт
+- ▁knapp
+- geschlecht
+- ▁strangled
+- ▁vieillard
+- ▁diving
+- kolera
+- uğu
+- ▁grotto
+- wiesen
+- kwano
+- iquement
+- ▁salió
+- ▁Kama
+- ▁assemblies
+- ▁côte
+- gord
+- ▁souhait
+- ▁خواهد
+- ▁쓰고
+- ▁stationary
+- ้อ
+- emerewe
+- ▁regelmäßig
+- hughuli
+- ▁albeit
+- ▁angenehm
+- ▁edad
+- lə
+- ▁كۈن
+- ▁мая
+- ▁yelp
+- ▁juny
+- ▁yiman
+- 놓
+- ▁tulip
+- ▁vun
+- ▁uninterrupted
+- ▁بچه
+- ▁contar
+- ▁mühe
+- ▁aimé
+- ▁bevond
+- ▁départ
+- ▁imprint
+- ▁zonderling
+- รถ
+- ▁Antonio
+- ▁categories
+- ▁circumference
+- ▁Service
+- ദ
+- ▁dombey
+- ▁Major
+- ▁handicap
+- ▁saucepan
+- ▁fickle
+- ▁sprout
+- ▁literaturo
+- ▁Young
+- ▁mammy
+- ▁Middle
+- ▁gukina
+- ▁wisse
+- ▁tracing
+- 에서는
+- ▁craven
+- agricultura
+- களும்
+- gebäude
+- ▁crítica
+- ▁preoccupied
+- ▁travaux
+- ▁варыянт
+- ▁ممکن
+- ৌ
+- ▁назад
+- ▁enliven
+- ▁дек
+- ▁һыу
+- ▁flakes
+- ▁novice
+- ▁zweck
+- ▁pumpkin
+- ▁wrenched
+- ▁bohemian
+- ▁Song
+- treppe
+- ▁жаль
+- ຄ
+- ယ
+- ▁dexterity
+- ▁irritating
+- ▁wünscht
+- ▁бацькі
+- シ
+- структур
+- 纳
+- ▁witchcraft
+- abati
+- ▁slammed
+- ündig
+- ▁permitting
+- ▁submerged
+- ▁roadway
+- ▁разумею
+- robed
+- ▁veía
+- ▁discut
+- ೀ
+- ▁prominence
+- 商
+- ▁اگه
+- ▁lucius
+- ridi
+- ▁strawberries
+- ▁மன
+- continent
+- ▁raporo
+- ▁тул
+- ahoze
+- ▁kiedy
+- gevangen
+- träge
+- Ғ
+- ▁susie
+- ▁repas
+- ▁estuvo
+- ▁investigated
+- ▁йәшә
+- ▁woodpecker
+- ▁vestibule
+- ▁stifling
+- ▁bumble
+- ▁contradictory
+- ▁consistency
+- drücken
+- ماق
+- ▁шога
+- ▁louvre
+- พิ
+- ▁gènere
+- ▁reparti
+- lju
+- ▁extensively
+- ▁populous
+- 部分
+- ▁confidant
+- ▁overgrown
+- ▁Mateu
+- tafel
+- ঙ্গ
+- ▁transgress
+- ▁impertinence
+- umujyi
+- phar
+- ▁여러
+- ของฉัน
+- ▁Gö
+- ▁shoemaker
+- ában
+- бри
+- ▁Aufgabe
+- ▁میکرد
+- 택
+- 맥
+- ▁την
+- ▁Japanese
+- ކަން
+- kord
+- ▁парк
+- aktion
+- บ้าน
+- ▁Кор
+- ရှိ
+- 놨
+- 店
+- writing
+- ▁pedant
+- ifié
+- ிடம்
+- ▁feigned
+- ▁паверх
+- ▁nenion
+- ▁Western
+- ▁ensue
+- شتی
+- கொள்
+- ▁echoing
+- ແ
+- ▁jaune
+- ▁vivian
+- ▁griechische
+- هُمْ
+- ▁churchill
+- ▁palmerston
+- ▁Blanca
+- ▁вайны
+- 장이
+- ▁erzählt
+- ُوا
+- ▁café
+- ▁bloc
+- ▁школы
+- ▁કર
+- ▁ningún
+- ▁perdido
+- geber
+- ƙ
+- ▁emptiness
+- ▁Kultur
+- ▁دنیا
+- ▁Mannschaft
+- ▁tattered
+- ▁extinction
+- ▁repressed
+- ▁situacio
+- ▁신기
+- ▁stork
+- ▁grog
+- ந்தது
+- ுக்
+- ▁Water
+- ▁dismissal
+- ▁disobedience
+- ▁responsable
+- 群
+- 器
+- ▁ignoble
+- ▁bazaar
+- ▁hoist
+- larini
+- ވ
+- ▁nationality
+- isierung
+- ▁Beide
+- ▁enjoined
+- ▁poteva
+- ▁Umuhungu
+- ▁chwili
+- ▁whipping
+- ▁folle
+- ▁peso
+- arrivée
+- ▁guardar
+- lender
+- ▁marvelled
+- ▁refresh
+- ▁Group
+- ▁vivir
+- ▁ballot
+- twaro
+- ▁کمک
+- ▁керт
+- ▁sandwiches
+- ایت
+- ▁libraries
+- ▁pernicious
+- ▁галоўны
+- 雷
+- 门
+- ▁depressing
+- ▁rispose
+- ▁españa
+- στη
+- gäste
+- ▁sioux
+- मे
+- ▁heretics
+- ▁remonstrated
+- ▁jazz
+- ▁eingesetzt
+- ଏ
+- ▁recipient
+- ▁conceited
+- ▁uncouth
+- ▁encircled
+- ▁Alexander
+- ▁өҫ
+- ▁propped
+- ▁jessie
+- ▁fouquet
+- ▁goblet
+- ▁Lola
+- ▁biography
+- ▁Earl
+- ▁ruido
+- style
+- ▁utilitzar
+- ೂ
+- ▁exchanging
+- ▁ineffectual
+- ▁mpuzamahanga
+- ▁quijote
+- ▁sanctified
+- ეთ
+- ▁troll
+- 校
+- ىدىغان
+- ▁Гу
+- گان
+- ▁көт
+- ▁знаходзі
+- ▁stumble
+- ▁nsonga
+- ▁untrue
+- ndiye
+- ▁clime
+- offizier
+- ▁patria
+- artiste
+- kirch
+- ▁römischen
+- ▁adversaries
+- ▁сітуацыя
+- 附近
+- ▁initiated
+- ▁Ĉinio
+- ▁Society
+- ▁mourir
+- ▁거기에
+- ▁چون
+- aîné
+- fungwa
+- логі
+- رج
+- ▁перамог
+- ▁congrega
+- 족
+- ▁irrational
+- ▁penniless
+- ▁imbabazi
+- ▁Cloè
+- ▁lattice
+- ▁denomination
+- ▁diminishing
+- ▁agonies
+- ▁graag
+- ▁distribute
+- ▁occurring
+- ▁Clo
+- ▁organised
+- épaule
+- төн
+- ▁harass
+- ییە
+- analyse
+- 락
+- ▁incomparable
+- 川
+- ▁gedeelte
+- ▁Һин
+- nniḍen
+- ▁glut
+- смотр
+- ▁chargé
+- тҡан
+- ▁cramped
+- ▁incite
+- ▁prêtre
+- imyaka
+- ▁Ша
+- ात
+- قابل
+- ▁Kreis
+- ▁capella
+- ▁breakers
+- வில்
+- ▁parasol
+- ▁frey
+- ▁ҡыҙы
+- nyamuryango
+- Ҡ
+- ▁engraving
+- ▁domingo
+- ▁ludo
+- ▁аңла
+- ▁racket
+- ▁shingle
+- ▁داریم
+- ӹнь
+- значэнн
+- ▁surged
+- Imperi
+- ▁augur
+- ▁infrequent
+- ▁jüngling
+- teekeddwa
+- дзейніча
+- 务
+- ▁של
+- ▁beverage
+- ▁plundered
+- ▁investigator
+- тую
+- réis
+- ▁bezit
+- ▁Inama
+- ▁좋았
+- meeting
+- ▁viure
+- ▁flea
+- mouthed
+- 항
+- geworfen
+- 논
+- ▁лёгка
+- ▁ӱдыр
+- ▁alpine
+- olymp
+- ▁bambaye
+- َهُ
+- weebwa
+- 들도
+- ▁bataille
+- ▁distasteful
+- ▁advertise
+- ▁matrimonio
+- liad
+- ▁According
+- ▁indulging
+- ▁melbourne
+- ▁duplicate
+- ▁Gau
+- ▁alguma
+- ▁greenland
+- ▁desertion
+- ▁confiding
+- ▁trasport
+- avond
+- ▁һора
+- ▁Де
+- ▁buster
+- irland
+- fiɣ
+- ơ
+- ▁derision
+- ▁crass
+- ▁ruff
+- ▁hustle
+- troffen
+- ▁Baba
+- öv
+- amezi
+- ▁Stre
+- ▁nourished
+- ▁absicht
+- ▁итә
+- С
+- ىڭىز
+- ▁ئەمە
+- akorwa
+- 一個
+- ۽
+- ▁Toronto
+- ▁سۆز
+- 初
+- ▁onmogelijk
+- ▁malkovr
+- ▁들고
+- ▁ҡыҙ
+- ▁அடி
+- ▁scandalous
+- ▁waarlijk
+- ▁культуры
+- ▁kirby
+- ▁Web
+- ▁chasse
+- Geschäft
+- ▁sauntered
+- ▁بولۇپ
+- sacrifice
+- ▁snarled
+- ▁Pop
+- пуска
+- ▁Nyamuneka
+- 썼
+- ▁Energie
+- 境
+- ▁registaro
+- ▁sobretot
+- ▁infected
+- haranira
+- ▁persistence
+- ▁durf
+- ▁fumes
+- ▁باشی
+- 돈
+- spagnol
+- ย์
+- ▁contractor
+- praat
+- ▁име
+- ▁faig
+- ฉ
+- 管
+- lowooza
+- ▁condescension
+- ▁exultant
+- ▁fairfax
+- ▁inexhaustible
+- ▁كىشى
+- disposed
+- ▁praising
+- جل
+- ▁nervousness
+- জে
+- ▁elicit
+- brug
+- ▁loĝis
+- ▁இத
+- zmir
+- ▁Yari
+- 짝
+- যোগ
+- ▁shirley
+- ▁адрозніва
+- ▁circulated
+- ▁Ble
+- ங்
+- ▁adept
+- сына
+- ▁predict
+- gumu
+- 铁
+- ▁denunciation
+- ▁그것
+- ▁Tiutempe
+- ▁público
+- ▁propaganda
+- ▁sewed
+- ▁résolu
+- ecek
+- ▁clink
+- ▁может
+- حل
+- wanyi
+- 팀
+- ▁Fernseh
+- ▁лічу
+- ▁daraus
+- шьҭ
+- ▁mı
+- ▁wakened
+- modul
+- ▁cornelius
+- ▁judici
+- ▁matrimonial
+- ▁홍대
+- 뒤
+- pewa
+- ▁hugged
+- ▁enorm
+- ▁mostrar
+- ▁힘들어
+- oscope
+- motiv
+- ritu
+- ▁lungo
+- ▁serviceable
+- ▁religio
+- ▁Leur
+- ▁piedi
+- ▁paroxysm
+- ▁filino
+- ▁bleed
+- ▁гандл
+- ▁알았어
+- வரும்
+- ▁داشته
+- ouvrit
+- ▁kwirinda
+- ▁similarity
+- ▁rayon
+- ▁scholarship
+- cumbi
+- ▁droop
+- ▁Guarda
+- ▁padron
+- ▁Ви
+- குறி
+- ▁Kati
+- ística
+- ▁laddie
+- ▁Deshalb
+- чага
+- ення
+- ▁flitting
+- ▁housekeeping
+- ▁shule
+- ถ้า
+- ▁vomit
+- ▁negativ
+- ▁ferret
+- flanke
+- ▁yawn
+- ▁lourd
+- ച
+- ▁apologies
+- ▁fervour
+- 希
+- ürzt
+- ▁fergus
+- ▁Township
+- გი
+- ▁wunderbar
+- pressió
+- اڵ
+- ▁discount
+- ▁estilo
+- رق
+- ▁greenish
+- ▁Riu
+- ។
+- ▁reiterated
+- 工作
+- ▁sophy
+- пісаў
+- ▁wordsworth
+- ▁recite
+- களின்
+- lford
+- ▁blister
+- ۇپ
+- зван
+- ▁chewing
+- ▁starch
+- ▁admonition
+- ▁cleveland
+- ▁fletcher
+- ▁inscrutable
+- ▁impersonal
+- anglès
+- 说
+- ▁comenzó
+- 传
+- ▁exclamó
+- ▁hopkins
+- üş
+- 개월
+- ▁placé
+- ▁않았어
+- ▁सम
+- ▁الح
+- ▁category
+- ▁uitdrukking
+- ▁gravitation
+- ▁будынка
+- ող
+- ▁рашэнне
+- ▁extremities
+- ▁vertellen
+- ются
+- ▁mountainous
+- shoes
+- ▁роўна
+- ▁marrow
+- blind
+- เค
+- ▁thinner
+- цкай
+- ▁uplift
+- ▁hinab
+- präsident
+- ګ
+- ▁enterprising
+- ▁lancaster
+- ▁sapphire
+- ▁proffered
+- 木
+- ▁της
+- rugamba
+- üll
+- বার
+- мәҫ
+- ▁collins
+- ▁callous
+- stätte
+- ▁mercies
+- ▁revision
+- ▁schat
+- ▁presentiment
+- ▁rempli
+- geht
+- lunaku
+- 冇
+- эканомі
+- 클
+- ▁inderdaad
+- ▁légère
+- ▁차라리
+- ▁honeymoon
+- arribar
+- ▁drought
+- ▁sportsman
+- ▁mongol
+- ▁jeugd
+- ▁impious
+- ▁sallow
+- ▁jamii
+- ▁diffused
+- ▁agreeably
+- بینی
+- ▁Colom
+- flucht
+- тәре
+- шон
+- ▁распа
+- ации
+- ຊ
+- ஓ
+- ဖ
+- ▁commodities
+- ▁impassable
+- ▁неабходна
+- ▁dissimul
+- ылда
+- ▁mantenir
+- ambwiye
+- bücher
+- ▁җи
+- ▁shyness
+- 분이
+- zigera
+- ▁articulate
+- koreye
+- 兩
+- ▁intoxication
+- ▁futbolista
+- ▁Хоць
+- ▁texan
+- ▁flitted
+- ▁slapped
+- ▁جوان
+- ▁weasel
+- particip
+- acqu
+- geslagen
+- gabanya
+- ▁embankment
+- ▁가족
+- ▁سن
+- ▁모르겠
+- крэт
+- ▁briggs
+- পে
+- ▁Harry
+- ▁bohemia
+- கார
+- ▁kulturo
+- ▁ҡар
+- ▁praten
+- ▁iceland
+- ▁удзел
+- ▁Silva
+- ▁scherz
+- ട്ട
+- ୋ
+- ▁reservoir
+- ▁cançó
+- 限
+- 团
+- ▁respite
+- ▁Vieno
+- ▁schoolroom
+- scape
+- ▁schijnt
+- safiri
+- iéndole
+- ▁digne
+- beug
+- ▁Verwaltung
+- ▁caricature
+- ▁diabolical
+- ▁vocabulary
+- ▁امروز
+- ▁일어나
+- ▁verbunden
+- ▁projection
+- ▁entscheiden
+- ▁bashful
+- لاش
+- ▁kupata
+- ▁disdainful
+- ▁bagenzi
+- ▁amends
+- umvira
+- zoni
+- 활
+- ▁interne
+- ▁paralyzed
+- ▁preocupa
+- ▁그랬는데
+- ▁Jaume
+- ▁nijoro
+- ▁ئۆت
+- rundi
+- ▁Petro
+- ▁shedding
+- ▁거니까
+- ▁chronic
+- ologic
+- ৰ্
+- รุ
+- сіў
+- ível
+- 瓦
+- 实
+- fluß
+- ▁йӧр
+- ▁coronation
+- pół
+- zzu
+- ▁guteza
+- ▁Хә
+- ▁Unit
+- ▁gdzie
+- ▁scop
+- ▁jambo
+- ▁Ру
+- 빼
+- 빌
+- ▁marlborough
+- ▁interfering
+- ▁chevaux
+- ▁vince
+- ▁дроб
+- ▁indessen
+- ▁cuadr
+- ▁rolf
+- экс
+- ▁größer
+- ▁bewegt
+- ▁koud
+- ଣ
+- complacency
+- ▁poignant
+- ▁wolsey
+- ▁کشور
+- gaciro
+- ▁sentido
+- ▁pâle
+- 景
+- ▁procuring
+- روس
+- ▁poop
+- ▁Jugend
+- кола
+- гь
+- ▁enquire
+- ▁아빠가
+- ▁desenvolupament
+- ▁postmaster
+- ▁아침에
+- künstler
+- ▁overpowering
+- ▁вялікая
+- κό
+- ▁Einsatz
+- ையை
+- kandi
+- ▁nikolaas
+- ▁withdrawal
+- enthusiastically
+- ▁Carlos
+- ▁immigration
+- 即
+- 米
+- ▁chand
+- ▁ecco
+- ▁வழி
+- ▁mobili
+- ▁palacio
+- ▁llegado
+- ▁winkle
+- ▁ingenuous
+- ば
+- 技
+- ▁groove
+- ▁bavaria
+- lî
+- ▁goethe
+- ▁paix
+- خاطر
+- ▁compromis
+- bulu
+- ▁merlin
+- ▁firelight
+- ▁geometry
+- ▁އަހަރެން
+- ▁ontvangen
+- ▁энерг
+- ލް
+- ▁Такі
+- ▁пеле
+- ▁denounce
+- ▁curtly
+- ▁esperar
+- ▁Тик
+- ۋى
+- ▁بچ
+- ▁ўсход
+- ▁eenmaal
+- ▁garage
+- huizen
+- ▁continuació
+- lome
+- ▁buzzing
+- ▁pillage
+- 乐
+- ▁번도
+- ▁myinshi
+- ▁syrup
+- ▁houdt
+- ▁Theater
+- ▁маці
+- lân
+- ▁creat
+- τά
+- ▁menschliche
+- ▁pràctica
+- ▁armstrong
+- ▁immigrants
+- ▁kifo
+- ▁lounging
+- галас
+- สอง
+- ▁rotation
+- ▁inducement
+- ▁هەڵ
+- ▁preferable
+- ғанда
+- ▁kidd
+- gá
+- ▁festivities
+- ▁secession
+- ޔާ
+- iridde
+- ▁discrimination
+- ▁stalked
+- ▁monastic
+- ▁deepening
+- ▁reconstruction
+- ▁bizarre
+- ▁ludzi
+- ▁같다
+- нӑ
+- ▁prospective
+- ▁Wij
+- blanc
+- fruit
+- މީހ
+- genoot
+- yini
+- ▁marchmont
+- ▁unaffected
+- ▁nuclear
+- ▁upturned
+- 支
+- ▁alegre
+- ▁هفت
+- ▁Little
+- ілася
+- tsinze
+- ▁malade
+- zusehen
+- ▁ئار
+- печ
+- yasi
+- جار
+- ▁vlak
+- ▁Licht
+- ლე
+- umwihariko
+- ▁Carolina
+- ▁오히려
+- ▁adversity
+- ▁lovelace
+- ▁pecado
+- ▁Она
+- ▁merton
+- ensonga
+- ▁bosch
+- ▁filmoj
+- ▁tratta
+- ▁deceitful
+- ▁forecast
+- ▁Beat
+- оит
+- ▁rousing
+- 書
+- 故
+- ▁conspire
+- ▁갔을
+- ▁utawala
+- ▁eskimo
+- ▁ukomeye
+- wirt
+- struction
+- ▁Drei
+- ▁marié
+- ▁Gott
+- ▁Monat
+- escut
+- ▁modify
+- ▁inicial
+- ▁ўвесь
+- 環
+- ▁macedonia
+- ▁Bamwe
+- ▁ugliness
+- ▁boatswain
+- second
+- ▁фасад
+- ▁utilise
+- ▁maxwell
+- ित
+- ▁trotting
+- 的人
+- ▁fàcil
+- ▁Ḥemmleɣ
+- ▁તે
+- ▁Vilanova
+- ▁forbearance
+- ▁trimming
+- ▁indruk
+- ▁boog
+- О
+- ▁Це
+- ▁cypress
+- block
+- зон
+- 늘
+- ▁bedding
+- Jacques
+- ▁reindeer
+- reißen
+- ▁имеет
+- ▁adorable
+- ▁있다고
+- bright
+- ▁rivier
+- ▁savor
+- ▁krä
+- eastern
+- ▁silhouette
+- ▁venía
+- ▁послед
+- zanye
+- тҡа
+- ▁arriv
+- fight
+- খন
+- ayagala
+- ▁Metro
+- ▁ئۇلار
+- ுகிற
+- ▁senat
+- Κ
+- ဖြစ်
+- ▁mumbled
+- ▁veronica
+- 시켜
+- laki
+- ▁enviar
+- paß
+- ▁tinkle
+- աց
+- ▁popola
+- ▁devint
+- ▁shave
+- ▁krijg
+- ौ
+- ▁clennam
+- 料
+- èvre
+- ▁даўно
+- occupe
+- รัก
+- ▁지나
+- ▁Пе
+- ▁கால
+- ਦਾ
+- ▁álbum
+- 낫
+- ▁constellation
+- тэх
+- ▁walpole
+- ▁grapple
+- ▁Burundi
+- ড়া
+- ▁piteously
+- ▁pijn
+- ▁marina
+- gelöst
+- чнымі
+- واقع
+- ▁dritten
+- imprim
+- 쿠
+- exèrcit
+- љ
+- 며
+- ▁errichtet
+- เม
+- ▁Потым
+- ▁chaud
+- ▁خونه
+- ▁lurid
+- ▁Staats
+- oeil
+- ildi
+- ▁Пі
+- ▁Original
+- ètic
+- ▁любо
+- ▁alcuni
+- ▁husky
+- ▁umusaruro
+- ▁студэнт
+- ▁scanned
+- நிலை
+- leiter
+- ▁bahari
+- ▁equipage
+- ▁nocturn
+- ▁purified
+- mesi
+- що
+- coated
+- tsje
+- ▁powie
+- eignet
+- ▁жыцці
+- ခဲ့
+- ▁schlafen
+- ▁эксп
+- ▁muhimu
+- ngula
+- ▁zinc
+- ▁cessation
+- ▁Italio
+- ▁trow
+- ▁wohin
+- ылды
+- ▁dutiful
+- ▁antigua
+- ▁Bernat
+- ▁Flo
+- ▁shutter
+- ▁draped
+- offre
+- ▁tripped
+- ▁naskiĝ
+- گری
+- ૂ
+- ▁título
+- ▁Projekt
+- ▁mahomet
+- ▁eŭro
+- ▁magnify
+- kute
+- ▁exceptionally
+- نجا
+- ▁justicia
+- ubahiriza
+- ▁laboratori
+- ருக்கு
+- ▁marcella
+- ▁teasing
+- িয়ে
+- ▁hampshire
+- ▁incompatible
+- ▁prettily
+- ފާ
+- ▁podría
+- ocell
+- ▁malay
+- ▁cramp
+- muth
+- глядзе
+- utilizza
+- ▁хү
+- ▁actuellement
+- ▁prettier
+- џь
+- ▁ntekereza
+- ▁succour
+- ▁gerust
+- ▁außerdem
+- ▁Japó
+- ▁нібы
+- ▁fortikaĵo
+- ▁knotted
+- ▁Deux
+- ▁тине
+- 었지
+- білі
+- indiqu
+- ▁durfte
+- ▁Dienst
+- ການ
+- ▁bethlehem
+- ▁precipitous
+- ▁igitabo
+- ▁gauge
+- ▁celebrity
+- றிய
+- ▁서로
+- ▁quoique
+- ukwezi
+- blatt
+- ▁longitude
+- vrouw
+- ρά
+- ▁südlich
+- ▁provisional
+- ▁خط
+- ▁arcade
+- 하거든
+- ▁anointed
+- ▁патрэбны
+- 毛
+- ▁күңел
+- ทั้ง
+- 否
+- ▁فیلم
+- ▁visitation
+- ▁navega
+- ▁gamble
+- گذار
+- X
+- ▁rhai
+- ▁reduc
+- ▁lavora
+- ▁projekto
+- ▁estudia
+- ▁vortoj
+- возмож
+- ޕ
+- ▁öffnete
+- kräftig
+- ▁카페
+- 半
+- 越
+- ▁zwijg
+- ▁swerve
+- ▁actividad
+- ▁mammals
+- ouvrir
+- зал
+- ▁rind
+- ▁gaire
+- ҩы
+- comers
+- ▁komuni
+- ▁contended
+- ▁футбол
+- ▁пачынае
+- ▁살아
+- ▁blauw
+- ▁chopin
+- तो
+- ▁qualité
+- ണ്ട
+- ορ
+- ▁contour
+- ▁palate
+- ▁tinker
+- rijden
+- ▁completing
+- тең
+- ▁poursuiv
+- ▁Rest
+- ▁годзе
+- ▁пачала
+- ▁erfüllt
+- ▁konservi
+- ▁surging
+- ▁apache
+- ▁پیر
+- ហ
+- ▁jeremiah
+- ▁avalanche
+- ەتی
+- ކޮ
+- ▁bunker
+- ▁enchanting
+- ▁fête
+- ▁slanting
+- ▁meaningless
+- чэнне
+- ▁mornin
+- শ্
+- ▁самым
+- œil
+- cruci
+- ▁scherp
+- ymmetri
+- ▁Show
+- ▁olga
+- ▁rudeness
+- ▁libertad
+- ▁mediaeval
+- ▁eligible
+- ▁텐데
+- 史
+- ▁varnish
+- writer
+- ▁Quirze
+- ▁sprinkle
+- ▁armament
+- ▁rubanda
+- ▁sloth
+- ▁playmate
+- ▁nachbar
+- ▁articul
+- ▁lokoj
+- seminar
+- ▁Chile
+- bodied
+- ▁stilte
+- タ
+- ▁тэрыторыі
+- 進
+- ▁delirious
+- ▁pixie
+- ▁мусіць
+- ▁drukte
+- leitung
+- ▁해봤
+- ▁வந்து
+- או
+- important
+- arrêta
+- 찾
+- ▁famiglia
+- рэшце
+- ▁revolucio
+- schränkt
+- ▁lugha
+- ▁scull
+- ▁wolken
+- ўрад
+- ▁گوش
+- 악
+- ▁principalement
+- ▁complète
+- ▁imperceptible
+- ▁dowager
+- ▁Үҙе
+- ▁Professor
+- ▁melville
+- ური
+- ▁pří
+- rzę
+- võ
+- ▁Polisi
+- யில
+- ▁peevish
+- 必
+- injiye
+- ▁malheureux
+- ▁seraph
+- gewiesen
+- ▁пакры
+- ▁Ple
+- ▁transient
+- ▁alderman
+- ▁먹으면
+- الله
+- ▁analog
+- esdeveniment
+- ▁belangrijk
+- ▁некаторыя
+- ▁교수님
+- ▁당연히
+- ▁Ontario
+- ▁wohnung
+- ▁ineffable
+- ▁schwieg
+- ▁hefyd
+- ▁workmanship
+- ▁froze
+- ▁forgave
+- ▁civic
+- ▁twese
+- stehenden
+- acteur
+- ชา
+- ▁selecció
+- 거리
+- ▁sigue
+- ▁merci
+- дон
+- ▁prosecutor
+- ▁마지막
+- ▁porcelain
+- ▁근까
+- ある
+- ▁cravat
+- ▁அவர்கள்
+- ▁пяць
+- ▁பேச
+- ▁atheist
+- prés
+- äger
+- ▁swindle
+- آمد
+- elaar
+- ▁eunuch
+- ▁пытанні
+- gebrochen
+- ▁genesis
+- contra
+- ▁comprendre
+- ▁skilfully
+- risti
+- ಿಯ
+- righteous
+- ಲ್ಲಿ
+- ▁tablets
+- ▁kemp
+- ▁njia
+- ▁dinar
+- ގެން
+- cucu
+- ▁capitán
+- ▁irrelevant
+- ▁tècnic
+- ▁southampton
+- ▁mulberry
+- เวลา
+- ▁skunk
+- ▁выконва
+- ▁выраз
+- ▁mugore
+- ▁bluntly
+- ლის
+- ▁Game
+- ▁электр
+- เว
+- ▁gutwara
+- kamba
+- бярэ
+- ▁fend
+- ▁feign
+- ▁Hauri
+- ▁кашт
+- ৰি
+- 쁜
+- jord
+- ▁brooklyn
+- ▁gukura
+- ▁Freund
+- ▁میشه
+- ▁drugi
+- ▁panie
+- handle
+- ▁reag
+- ▁пераклад
+- ▁وەک
+- ▁membership
+- ۈش
+- ▁پشت
+- ▁tolerated
+- buuza
+- ▁partoprenis
+- ▁uwiteka
+- އްޖެ
+- ▁huwelijk
+- teekateeka
+- ▁атрымалася
+- ▁amabwiriza
+- ▁بگیر
+- ▁развіва
+- ▁karibu
+- ▁astonish
+- ▁abandonment
+- ▁ferris
+- ▁veinte
+- عِ
+- ▁sensibili
+- ▁Azio
+- ▁sphinx
+- ddwaliro
+- ▁corporeal
+- ▁элемент
+- ӹш
+- ▁bepaald
+- akundaga
+- attaque
+- ▁cordiality
+- 형이
+- ▁kitu
+- ▁ebony
+- ▁llamado
+- ▁Tren
+- ▁பார்த்த
+- Mfite
+- ▁champlain
+- ▁dorothea
+- ▁phoenix
+- ▁connaissance
+- ▁période
+- 议
+- ▁prelude
+- ▁rambling
+- ▁espacio
+- getreten
+- ▁Bona
+- ▁شام
+- ▁cię
+- fated
+- ▁madison
+- ▁ventana
+- ▁پسر
+- ҩа
+- ▁цікавы
+- ▁Bruna
+- оўцы
+- ޔ
+- ▁Thaís
+- ▁blasphemy
+- ▁intercession
+- ▁puppet
+- ▁exerci
+- ▁diminution
+- ▁tegenover
+- ▁revenis
+- ▁uretse
+- спорт
+- ቀ
+- nyigisho
+- ▁freckles
+- ност
+- ▁هاو
+- ▁beispiel
+- ▁사람은
+- thür
+- content
+- uvieron
+- கல
+- ▁Scotland
+- ଗ
+- ▁siegfried
+- 味
+- ▁degrading
+- ▁lavender
+- arenze
+- ▁calculating
+- ▁eddie
+- ▁sanft
+- ubutaka
+- ▁ibyiza
+- ▁sulky
+- ▁Fisch
+- லு
+- ▁Guard
+- ▁transgression
+- ▁obliterated
+- ▁dunno
+- ▁minnie
+- ▁eventful
+- ▁äußerst
+- 語
+- ▁appellation
+- цця
+- ▁finalmente
+- ▁vosto
+- ▁hendrik
+- ▁paese
+- ▁boor
+- baruwa
+- ▁militant
+- ▁spasm
+- sichtlich
+- ▁chiesa
+- ▁redeemer
+- के
+- ▁bedstead
+- ▁shrouded
+- mäh
+- ဘ
+- ▁gikomeye
+- ▁impassioned
+- ▁monasteries
+- 냈
+- quatorze
+- ▁algebra
+- komera
+- प्
+- ▁nuntempe
+- ▁Meister
+- ▁enlightenment
+- ▁faithless
+- amato
+- ▁resultado
+- ▁Altres
+- ▁ideo
+- ލި
+- usuari
+- ▁teoria
+- ▁Down
+- ▁gitaramo
+- ▁Algunes
+- ▁unfriendly
+- ▁invariable
+- ▁тэкст
+- ▁myrtle
+- ▁cheerily
+- ▁totalment
+- ціў
+- giza
+- ▁keerde
+- guza
+- diplom
+- arangije
+- ▁parental
+- where
+- ▁hampstead
+- ▁incumbent
+- ▁replenish
+- ▁لەسەر
+- 选
+- ▁lorry
+- ▁இர
+- ▁tolerate
+- ▁spelen
+- ▁schulter
+- ▁பழ
+- слуг
+- ▁emplea
+- ിക്ക
+- ▁retirar
+- ▁іншым
+- ▁compulsion
+- ▁மனித
+- ▁thunderbolt
+- ▁illumined
+- เกิด
+- ▁erwarten
+- bereich
+- ▁hinunter
+- ▁depict
+- ▁axiom
+- ▁праблем
+- ▁ogw
+- basomesa
+- ының
+- ▁réveil
+- stitució
+- ▁counteract
+- သည်။
+- ജ
+- ▁oblivious
+- ▁societat
+- ▁свои
+- 视
+- ▁underwent
+- ▁Mountain
+- пэўн
+- ตก
+- ▁objectionable
+- ▁connaître
+- ▁corrent
+- 茶
+- 歌
+- 则
+- ▁dîner
+- uriye
+- ▁мастак
+- ▁sahib
+- ▁nettie
+- として
+- ▁lokal
+- herrschaft
+- ▁verslag
+- dream
+- алагічн
+- ▁gelten
+- ▁impassive
+- ▁palpable
+- ▁அவள்
+- ▁wearisome
+- レ
+- 亦
+- ▁vizaĝo
+- нення
+- ▁Under
+- ▁Адзін
+- ▁کي
+- ▁bocca
+- ▁funnel
+- шел
+- macher
+- ▁patting
+- ▁davie
+- ந்
+- ▁rhythmic
+- ▁unutterable
+- ▁ĝenerale
+- ▁અને
+- ▁대학교
+- ▁heinrich
+- ▁normalment
+- शा
+- ▁جهان
+- ▁shatter
+- ▁самых
+- ibigo
+- ▁Pobla
+- ▁stink
+- ▁ilustr
+- ▁좋지
+- ▁ҡур
+- ▁représente
+- ▁retinue
+- ▁symphony
+- ▁mukobwa
+- ▁сустрака
+- 零
+- 望
+- ▁voldria
+- 塞
+- 图
+- ▁высока
+- ▁étudie
+- ▁götter
+- ▁fissure
+- ▁аҙ
+- ▁هذا
+- ▁دەبێت
+- ▁yasabye
+- miterere
+- ▁mermaid
+- ▁daarbij
+- ▁сфер
+- ▁kleider
+- mique
+- ▁thir
+- ▁nazareth
+- ▁norwegian
+- ▁superfície
+- ▁extricate
+- އީ
+- gebleven
+- ပ်
+- ▁musketeer
+- ▁sexton
+- ▁Einige
+- ▁gunboat
+- ▁ruĝa
+- ▁Kubera
+- ▁Mariona
+- ▁Augen
+- 其他
+- ▁voorstel
+- àrea
+- ццё
+- ▁Itsinda
+- ▁Democrat
+- ▁evinced
+- ▁irresolute
+- ▁unsettled
+- ▁antagonism
+- ▁agility
+- ▁ہی
+- ▁acquiescence
+- ▁généralement
+- ▁востра
+- ▁Wild
+- ▁intervened
+- ▁справе
+- ▁boscos
+- feḍ
+- ▁argi
+- ▁taula
+- hiver
+- влӓ
+- ▁mcclellan
+- ▁navigator
+- ▁зразумела
+- 莱
+- 밀
+- 专
+- ▁endowment
+- ▁Rosó
+- 세요
+- ▁Stock
+- izeye
+- ▁navire
+- ▁запіс
+- ▁merged
+- ▁ausdruck
+- ▁realism
+- ▁rufus
+- ▁گیا
+- ▁sucked
+- ▁달에
+- ਆ
+- ណ
+- sufficiency
+- ▁نزدیک
+- ധ
+- 彼
+- ▁mishap
+- ▁reputed
+- ▁grumble
+- verband
+- ▁Consta
+- riff
+- topped
+- ▁йом
+- ▁sorrowfully
+- ▁gregg
+- ▁тавар
+- ▁besluit
+- umani
+- vaardig
+- ▁pheasant
+- జ
+- ▁ṣṣ
+- ▁urubanza
+- ▁대한
+- ▁darkening
+- schot
+- ▁hitting
+- رود
+- ▁volum
+- phil
+- ▁تێ
+- ▁abashed
+- ▁regionoj
+- ▁bedenk
+- leger
+- லம்
+- bisha
+- 익
+- ਉ
+- ▁waarschijnlijk
+- 쩌
+- އ
+- babaza
+- ▁Такім
+- ▁shirk
+- ларға
+- ▁rimark
+- vité
+- litten
+- ▁carlton
+- ▁arabic
+- ▁tronc
+- ▁mathematician
+- ▁ecstatic
+- ▁seymour
+- それ
+- ▁moindre
+- ▁ekonomia
+- ▁matri
+- ▁булған
+- ▁Alice
+- ▁மக்கள
+- ▁없었
+- ▁Ужо
+- ▁informal
+- ▁tourne
+- ▁sociolog
+- letten
+- кою
+- ▁décide
+- ੈ
+- ▁absorption
+- ▁dissatisfaction
+- ▁moccasin
+- ▁fabulous
+- ▁آهي
+- ▁unanimously
+- ▁рабі
+- ▁superficie
+- ▁drummond
+- 정이
+- ▁produeix
+- ▁ҡала
+- onyine
+- fizik
+- ৃত
+- ▁houding
+- ▁giovanni
+- ▁pretense
+- ▁idrimen
+- ▁deafening
+- ▁tofauti
+- ను
+- ▁inquiringly
+- aşı
+- babaje
+- bụ
+- ▁masonry
+- ▁korto
+- ▁vizier
+- ▁fakte
+- taub
+- ▁உலக
+- ▁slipper
+- вілі
+- ▁cwiṭ
+- ▁яҡшы
+- ▁furcht
+- ledig
+- ாவது
+- àcia
+- ▁dehors
+- ▁educate
+- 응
+- ▁substanti
+- runde
+- цамі
+- ություն
+- ▁botigues
+- ▁glasgow
+- ▁réponse
+- ▁genutzt
+- ▁repudiate
+- ▁languor
+- ▁інтэр
+- ▁primrose
+- ▁wrth
+- ▁Patrick
+- ▁Zoè
+- ▁цэлы
+- ▁benediction
+- ▁Partei
+- yiɣ
+- kül
+- ▁warrington
+- ▁unconsciousness
+- ▁شدن
+- ▁Lwak
+- ▁contempla
+- això
+- ▁príncipe
+- ▁Наций
+- ▁gobierno
+- ▁klimato
+- ▁combatants
+- ▁kabi
+- caire
+- ▁viaj
+- ▁가기
+- ▁finit
+- ▁вечар
+- ▁malaria
+- ▁forfeited
+- ▁harri
+- ።
+- ټ
+- ▁chastity
+- ▁باشد
+- ▁thresh
+- ▁literari
+- ▁mistaking
+- ▁nafasi
+- λε
+- ▁fuori
+- ▁erreichen
+- ▁regime
+- ▁tottering
+- ▁optimist
+- ▁Umuyobozi
+- ▁brigadier
+- ▁tawny
+- ▁ماشین
+- ▁sauvage
+- ▁billion
+- ▁clog
+- ▁erscheinen
+- ▁Album
+- ▁Handel
+- غیر
+- ▁Cuba
+- gurisha
+- ▁stupor
+- ▁сара
+- ▁puddle
+- ▁monmouth
+- 버리
+- exclam
+- ibyishimo
+- ▁Antônio
+- ▁Объединенных
+- ▁campionat
+- ▁tudert
+- yambere
+- ▁étrange
+- basajja
+- ▁anatomi
+- simbi
+- ▁стал
+- ▁مخ
+- ▁gutekereza
+- ▁quartz
+- ▁noiseless
+- ▁kvazaŭ
+- ▁겨울
+- ▁deputation
+- ▁revuo
+- ▁مهم
+- ▁ububasha
+- ▁месцамі
+- zustellen
+- ▁presidential
+- ▁departament
+- ▁hallway
+- श्
+- ikoranabuhanga
+- ▁grandchildren
+- ▁mountjoy
+- ▁diumenge
+- angled
+- ▁griffith
+- ▁wawal
+- ▁klaar
+- فهم
+- ▁hashize
+- ▁renom
+- ▁hanover
+- arranged
+- ▁نشان
+- уеит
+- ▁demselben
+- заходні
+- ▁yabwiye
+- ▁cetera
+- ▁واقعا
+- ▁گل
+- qal
+- ▁proche
+- amazzi
+- ▁падыход
+- ▁بولسا
+- ▁birimo
+- ▁comporte
+- ▁condensed
+- ▁renaissance
+- ▁мастацтва
+- 龙
+- 火
+- 结
+- ▁спо
+- ▁décor
+- ▁banter
+- ните
+- ▁arizona
+- ▁hobson
+- verhältnisse
+- ுதல்
+- maschine
+- ▁glean
+- ▁reĝo
+- ļ
+- ီး
+- ▁Ukrain
+- ▁primavera
+- ▁cairn
+- ▁yabaye
+- ▁Быў
+- ▁mortified
+- રી
+- ▁Ву
+- ่อน
+- ambwira
+- ▁gully
+- ▁derrota
+- 进
+- ▁cochrane
+- 书
+- ▁candour
+- ▁apparten
+- ▁Segons
+- ▁Баш
+- ▁thankfulness
+- ▁coisa
+- bica
+- Vir
+- ▁blankaj
+- ▁Davis
+- ▁swimmer
+- ▁адбылося
+- առ
+- ▁immobile
+- ▁کتاب
+- rhein
+- ▁pilate
+- ٿ
+- ▁recor
+- ▁slouch
+- ▁аднак
+- гәндә
+- ▁treaties
+- ▁hawthorn
+- ▁gibi
+- разу
+- ▁йӱк
+- Amèrica
+- ตั้ง
+- ▁consecutive
+- ▁rippling
+- ▁précipit
+- 塔
+- ▁слабы
+- ▁reizen
+- ▁tradesmen
+- ▁Egypt
+- wissenschaftlich
+- ▁verified
+- ▁marqué
+- Ministr
+- ▁ouvi
+- ▁вада
+- szą
+- ▁babona
+- ▁ntebe
+- ൂ
+- ▁besitzt
+- ầ
+- ▁efficacy
+- ▁Chinese
+- ▁poliisi
+- ന്ന
+- ▁خبر
+- ▁않은
+- economia
+- ▁loot
+- ▁respekt
+- ▁maniac
+- ▁lebendig
+- ▁blackbird
+- ▁lucha
+- ▁misled
+- раст
+- ▁squander
+- 「
+- 찌
+- ▁affinity
+- ▁phoebe
+- ▁ŝatas
+- ▁baadhi
+- klasse
+- ▁بېرى
+- ijver
+- ▁här
+- fatanye
+- ▁маю
+- ▁aggie
+- ▁umubare
+- ຂ
+- ଟ
+- ▁moeilijk
+- ▁залежыць
+- ಜ
+- ▁coaxing
+- 失
+- ্ম
+- ▁apostolic
+- тычныя
+- ▁stau
+- ▁Front
+- ▁ilaq
+- bireba
+- ▁linton
+- ▁tebɣ
+- ▁discontinu
+- ▁meinung
+- ▁glorify
+- ائی
+- ▁suburban
+- ▁lágrimas
+- ▁nyandiko
+- 幾
+- ▁sensibly
+- ▁einzelne
+- ▁transact
+- ▁whist
+- ▁bustling
+- ▁barter
+- 也是
+- abandon
+- ξε
+- ▁colère
+- lecció
+- ▁sandstone
+- მის
+- obéi
+- ▁quelqu
+- ▁Jersey
+- ▁haughtily
+- ▁Вельмі
+- ▁collier
+- ▁osiris
+- ▁arbor
+- νο
+- ▁activa
+- ▁roderick
+- ▁remis
+- ▁świat
+- François
+- ▁conflagration
+- ▁dissension
+- ▁취업
+- ▁impulsively
+- acya
+- ▁gaudy
+- akarere
+- авых
+- ▁discouragement
+- 보니까
+- ▁ліку
+- бө
+- ▁giorni
+- ങ
+- ▁kwishyura
+- ▁faltering
+- өгө
+- escalier
+- ▁Betrieb
+- ▁selben
+- ▁Work
+- ▁canviar
+- ▁драўлян
+- ส่วน
+- ▁bristling
+- ▁cavity
+- ▁netherlands
+- 빵
+- 쌤
+- 画
+- ▁걱정
+- ▁staunch
+- ▁مورد
+- ▁zitandukanye
+- ▁verkeerd
+- ▁solitari
+- နေ
+- ▁banishment
+- ▁protective
+- ruption
+- ▁хацеў
+- ▁drummer
+- bivuze
+- ▁생각보다
+- ที
+- ▁raphael
+- 求
+- ▁dasselbe
+- 包
+- ▁دختر
+- ކި
+- ▁specified
+- टा
+- ▁princi
+- ▁Tochter
+- ▁ҡына
+- 마다
+- ▁heerlijk
+- добр
+- ▁griffin
+- ▁vertrouw
+- ▁тағы
+- توانی
+- ▁galloway
+- ▁melodious
+- ▁kirekire
+- ▁worries
+- 功
+- ▁bifite
+- ▁appropriation
+- ▁mwenye
+- ▁uttermost
+- ▁Abagabo
+- ْتَ
+- attendre
+- ▁flickered
+- ▁snorted
+- ggling
+- oxid
+- ▁chivalrous
+- ▁rapturous
+- ▁unprepared
+- 段
+- ▁membrane
+- রী
+- ▁газ
+- ächtig
+- ▁можам
+- ▁фіз
+- ậ
+- ▁Canadian
+- ▁assiduous
+- ▁hoeveel
+- ▁beinahe
+- müde
+- developed
+- chukua
+- pě
+- mbyi
+- ▁reconstruct
+- ▁Saison
+- umuyaga
+- ▁trapped
+- claim
+- ultimo
+- ▁ajudar
+- springen
+- ڤ
+- ▁résultat
+- 」
+- bě
+- ▁protracted
+- ▁물어보
+- ▁шырока
+- ތީ
+- ością
+- ▁Mount
+- велі
+- disciplin
+- ії
+- ▁shapeless
+- ຮ
+- ▁сказаў
+- ▁indomitable
+- 別
+- ▁randall
+- ▁unerring
+- ▁antwerp
+- ▁annul
+- ▁noemen
+- ▁municipality
+- ▁hazy
+- सि
+- ▁Stone
+- ▁nocturnal
+- ▁자주
+- স্থা
+- ▁Greek
+- ▁alrededor
+- ▁triangular
+- ▁Santiago
+- ਿਆ
+- ▁onderwerp
+- ▁crec
+- ▁füßen
+- ▁Llop
+- amateeka
+- ▁genoa
+- க்கப்பட்ட
+- ▁себе
+- ▁новых
+- achtung
+- ▁graveyard
+- ĥo
+- ▁lightened
+- ▁sediment
+- digger
+- ▁fez
+- ▁ebiri
+- ▁Wissen
+- Umuryango
+- ▁абавязк
+- ▁Gabriel
+- ▁läuft
+- ▁болгон
+- ▁retrospect
+- ▁Walter
+- ▁discordant
+- ▁portland
+- ▁بهتر
+- ▁Nick
+- illumina
+- ▁nausea
+- ▁clergymen
+- umuhondo
+- ▁fraternity
+- フ
+- 線
+- vermögen
+- لارنى
+- ৰে
+- ▁지원
+- ▁symbolic
+- roux
+- ▁briar
+- ▁bosc
+- ▁hamburg
+- staw
+- épreuve
+- ▁intuitive
+- ▁omstandigheden
+- ▁peremptory
+- ▁знайсці
+- ▁ਹੈ
+- ▁Rauric
+- ▁lèvres
+- ▁lorenzo
+- ▁venison
+- ▁aptitude
+- ▁stroking
+- ▁resposta
+- ▁undoubted
+- patia
+- ▁nzego
+- ▁lunar
+- өйө
+- ▁mwy
+- 往
+- ▁cobweb
+- ▁관심
+- ▁consegui
+- ▁entdeckt
+- ▁jeunesse
+- ▁veille
+- ▁diventa
+- ▁rakonto
+- льнік
+- ▁sommige
+- ▁Telm
+- ▁كۆز
+- ▁morrison
+- scrive
+- ▁croix
+- ▁меў
+- ▁башла
+- ▁ruinous
+- ▁Sicherheit
+- ▁Afegeix
+- 把
+- Olymp
+- ▁prowling
+- ▁galerie
+- ிலிருந்து
+- ехаў
+- ▁화장
+- żał
+- рэз
+- ▁defiantly
+- ▁repute
+- ▁refute
+- schicht
+- ▁dalton
+- ▁bwite
+- ▁sundown
+- ometr
+- flüster
+- ▁equation
+- ▁lazarus
+- ▁puny
+- ▁kapitein
+- ється
+- ▁athlete
+- ṭṭef
+- ների
+- ▁бөтә
+- ▁reassure
+- ▁جای
+- ▁hazardous
+- క్
+- ▁tagoj
+- квет
+- ▁нимә
+- ▁dringend
+- ▁enamel
+- ▁bedeuten
+- ▁gervaise
+- ▁knapsack
+- 司
+- ▁예전에
+- ▁Хәҙер
+- ▁victoire
+- ▁싶었
+- ތައް
+- ▁animate
+- ▁لل
+- ▁moshi
+- ▁religió
+- آخ
+- בי
+- ్య
+- ▁tracy
+- ▁surrey
+- ▁складзе
+- গে
+- ▁Straßen
+- ক্র
+- مەي
+- ▁Harris
+- ▁brock
+- ▁nucleus
+- 딸
+- kurikije
+- ▁자꾸
+- ▁dervish
+- ▁michigan
+- ▁загад
+- ▁хво
+- ▁parlement
+- interès
+- ланды
+- ехаць
+- ▁España
+- ▁ئىككى
+- 園
+- ▁외국
+- 份
+- ▁sponsor
+- ▁julio
+- ▁nectar
+- ▁내용
+- ▁lynch
+- ▁chaperon
+- ▁summed
+- ▁süße
+- ▁아닌가
+- ▁gigante
+- стэ
+- ▁fracture
+- ▁lengthened
+- valuation
+- ೋ
+- ▁damascus
+- ▁muganga
+- rzuc
+- ▁تص
+- ▁posteriorment
+- ▁ehrlich
+- абара
+- ▁эше
+- وەر
+- ▁пасад
+- gewaltig
+- ▁igitekerezo
+- ▁shimmering
+- ▁utensils
+- ▁ўвогуле
+- 飯
+- ▁therefrom
+- หลัง
+- ▁incarnate
+- gå
+- ▁kubw
+- ▁Гэтыя
+- спява
+- দ্ধ
+- க்கம்
+- ▁نہ
+- ▁boeken
+- yetta
+- ▁Prze
+- ▁endorse
+- ਚ
+- ▁reconnaissance
+- ▁vídeo
+- abahungu
+- ▁Joel
+- ▁rupture
+- apercev
+- ▁reproachfully
+- ▁eliot
+- ▁yiruka
+- humour
+- ▁ningú
+- ▁يە
+- ▁sterven
+- ▁gioca
+- italia
+- ▁Korea
+- siedel
+- ▁plotting
+- ▁unavoidable
+- 딴
+- ▁괜찮아
+- ▁свята
+- ▁seminary
+- ▁shopkeeper
+- ▁austral
+- ▁nomiĝas
+- ▁castillo
+- ▁Mario
+- approcha
+- ▁hovel
+- fitted
+- ▁boulogne
+- ▁ignominious
+- ▁dibuix
+- 查
+- ▁bandera
+- ▁clinton
+- ▁музычн
+- ▁repelled
+- ▁hanaud
+- ▁citizenship
+- ▁Bord
+- ▁waffen
+- wijl
+- ▁Ohio
+- ▁classified
+- ▁voldoen
+- 쪽에
+- ئەت
+- لەم
+- 關
+- 头
+- 目前
+- ▁urugero
+- ▁roving
+- ▁brooch
+- ▁барып
+- ▁banjye
+- ▁Sibil
+- ▁konstruita
+- ▁работа
+- ▁practica
+- speaking
+- مەس
+- 됐
+- ▁yavutse
+- ▁taddart
+- ▁titular
+- 黑
+- ▁Aṭas
+- ▁відаць
+- ▁flinch
+- ہر
+- ▁дзеці
+- ▁bubbling
+- ▁incurable
+- ▁далёка
+- vò
+- ▁soaring
+- ▁Kurz
+- 었다
+- ▁bashaka
+- ▁concede
+- އިން
+- ▁corso
+- ▁glänzend
+- 門
+- ▁deborah
+- ▁palavra
+- ▁بیرون
+- ▁infirmities
+- ▁проект
+- ▁margot
+- ▁иде
+- ▁ärger
+- ▁Institute
+- ▁تق
+- ▁kadar
+- ▁cress
+- 억
+- ▁ئێمە
+- 域
+- ▁ricardo
+- މަށް
+- ▁parlant
+- адносін
+- ▁continuation
+- һенә
+- ▁sendepend
+- ▁mínim
+- ▁divisió
+- ▁orchid
+- ▁défini
+- brengen
+- entorn
+- ekebwa
+- ▁выпадку
+- ▁میکند
+- ▁vuelta
+- ▁nördlich
+- ▁conmigo
+- ▁petrol
+- ▁felipe
+- ▁Olympic
+- ▁situe
+- ▁counterpart
+- ▁Кол
+- ▁troost
+- ▁purify
+- ▁armenian
+- ▁gukorera
+- ൽ
+- edició
+- ▁всех
+- ▁personaje
+- ▁entender
+- ▁contraction
+- 범
+- ▁Richtung
+- ▁upbraid
+- 号
+- ▁infirmity
+- ▁beetje
+- kette
+- ▁Francia
+- ▁거기는
+- ɛem
+- ▁estaven
+- ▁мясцова
+- ▁alessandro
+- ▁beauteous
+- ▁nuptial
+- じ
+- 勒
+- ▁назіра
+- ▁школе
+- ▁такую
+- ্রি
+- ▁Modern
+- ▁blijft
+- стаць
+- ▁беларускія
+- ▁babble
+- ▁politiko
+- wifuza
+- ▁tiaj
+- ▁பெயர
+- था
+- ▁dolphin
+- ஹ
+- ெல்லாம்
+- ▁adjective
+- ▁только
+- 宗
+- ▁Fluss
+- いた
+- ▁даследава
+- ▁runners
+- ▁Мен
+- ▁cyanjye
+- ▁génie
+- ▁Georg
+- рыста
+- லில்
+- ▁airplane
+- ▁milliner
+- ▁johannes
+- ▁bertie
+- ▁preferring
+- 런
+- ▁parsley
+- linder
+- ޖ
+- 深
+- ယ်
+- ▁accosted
+- 整
+- ▁broadside
+- 城市
+- ▁blissful
+- àstic
+- ñó
+- ▁rejection
+- accions
+- скры
+- ▁fauna
+- 만한
+- bibwa
+- ჯ
+- ▁ejército
+- ▁problème
+- 씬
+- ▁filozofi
+- ▁phyllis
+- ▁дрэва
+- 苏
+- ▁Garbí
+- gusta
+- pacient
+- ▁pollard
+- ▁йән
+- ାର
+- ▁absolv
+- wulira
+- ▁vraiment
+- mereye
+- னம்
+- ▁amorous
+- ▁apologise
+- ▁pantomime
+- ▁паўночна
+- пита
+- ▁francès
+- ▁scarcity
+- συ
+- ▁amakosa
+- rechnet
+- ▁Шмат
+- ▁haalde
+- ▁зрабіў
+- ▁chrys
+- ▁бында
+- ▁clew
+- ▁ajili
+- cepció
+- ▁fracas
+- ▁strategic
+- ▁coloca
+- ▁вуліц
+- ▁adlis
+- sidér
+- ▁ancestry
+- 算
+- ▁peintre
+- akawuka
+- ▁kwibuka
+- ▁strategi
+- ▁paraules
+- ҙер
+- atinze
+- ▁impure
+- amabuye
+- ▁белә
+- 돼서
+- ntambwe
+- ັນ
+- ▁newfoundland
+- ▁паўднёва
+- ▁upheld
+- なる
+- ▁பற்றி
+- ▁Fatima
+- ▁extens
+- лёт
+- ▁discret
+- monger
+- ▁jewelry
+- ▁Ninde
+- ▁sukceso
+- ▁잘못
+- ▁daŭre
+- ▁mythology
+- ▁tiber
+- ▁burundu
+- ▁angular
+- ென்று
+- வாக
+- ▁acusa
+- ▁mategeko
+- Office
+- investigació
+- ▁implicitly
+- urukiko
+- thwaite
+- җ
+- สิบ
+- ▁sagacious
+- ▁prophesied
+- ลูก
+- ▁multipli
+- communicati
+- ▁ҡайҙа
+- ость
+- ▁verlor
+- ɣem
+- волі
+- ▁formality
+- ▁archway
+- аҩ
+- Π
+- ▁tennyson
+- 容
+- ▁meridian
+- ▁antikva
+- yewwi
+- ▁peterkin
+- ▁pluvas
+- ▁제가
+- ▁chteau
+- creatures
+- ▁Begriff
+- зеи
+- ▁першым
+- ขอ
+- daughter
+- kitongole
+- ▁repugnance
+- ▁vielmehr
+- 實
+- ▁тауыш
+- ▁gideon
+- ▁lettuce
+- ▁intensified
+- ▁раёнах
+- ▁குழ
+- لىقى
+- ▁лиеш
+- ▁storehouse
+- ▁шчыра
+- ▁détail
+- ול
+- ▁positi
+- عاد
+- ▁عن
+- بُ
+- سلام
+- ਈ
+- ▁disbelieve
+- ▁veröffentlicht
+- ബ
+- ▁mmoja
+- ▁deportment
+- ▁cançons
+- ▁sonderbar
+- ▁definitiva
+- гляда
+- ▁schönheit
+- ▁neigh
+- ▁inutile
+- ▁دنبال
+- ▁شروع
+- agerageje
+- лічыць
+- ▁reclaim
+- ▁risa
+- ▁rabble
+- ▁언니가
+- ▁неяк
+- lusion
+- glück
+- vunaanyizibwa
+- ▁phosphor
+- ▁ethereal
+- ▁প্রতি
+- 不同
+- عام
+- umunani
+- ▁Ekde
+- 就是
+- ▁venez
+- ▁Из
+- ▁seront
+- ▁haciendo
+- ▁hauptsächlich
+- ▁magnanimous
+- ▁vacuum
+- ▁aladdin
+- வல
+- ▁pencroft
+- ▁Sinzi
+- haguruk
+- ▁morceau
+- ▁bombardment
+- ▁copp
+- ▁recul
+- ▁lernen
+- ▁convulsively
+- ▁Vietnam
+- ▁Llavors
+- ▁negligence
+- ▁meredith
+- ένα
+- ▁capturing
+- ချ
+- ▁sombrero
+- ▁kawaida
+- ▁مست
+- ▁알려
+- ▁Kopf
+- ража
+- ▁refugi
+- کردنی
+- ▁Вас
+- ▁projectile
+- ▁мець
+- ▁glace
+- source
+- ۈرۈ
+- ▁dexterous
+- ▁sukcesis
+- ▁пазней
+- မှု
+- ▁portsmouth
+- ▁lḥal
+- ▁perusal
+- étroit
+- ▁poise
+- ▁actuated
+- ▁magnetism
+- ▁일단은
+- ▁mittag
+- shooter
+- ▁oculta
+- র্থ
+- ▁Ду
+- ▁cercle
+- ▁hablaba
+- interpret
+- ▁Mwaka
+- ▁Фа
+- atelier
+- ไว้
+- ▁Schrift
+- ▁escuela
+- 中心
+- ▁kuburyo
+- ▁입고
+- bî
+- ▁korpo
+- ▁meekness
+- ▁komuna
+- ▁arrêté
+- ▁papist
+- ▁larva
+- ▁Ergebnis
+- ▁inflexible
+- ▁twitching
+- ▁бліжэй
+- ▁پێویست
+- ▁ssomero
+- ▁Аўтар
+- ▁ŝtata
+- ▁rosamond
+- ▁lleno
+- ▁pronuncia
+- ▁footstep
+- ▁grouse
+- ▁wicket
+- ▁entró
+- ▁portier
+- ▁dissuade
+- ▁prophecies
+- ▁meṛṛa
+- gląda
+- ▁plor
+- ▁особ
+- ▁brawl
+- চে
+- ▁astronomi
+- ▁twari
+- ▁Elizabeth
+- ▁célèbre
+- ▁fictitious
+- ▁sorcerer
+- 拿
+- ▁stafford
+- roofed
+- ▁Ошо
+- ▁спя
+- ▁qualify
+- ▁bedeckt
+- ▁Núria
+- ▁receptacle
+- ಣ
+- ▁cuckoo
+- ▁viscount
+- ▁Fernando
+- ▁commissary
+- 单
+- étude
+- ▁stieß
+- ▁манеш
+- ▁peces
+- ▁knox
+- ▁radar
+- ▁savoy
+- ▁cashier
+- ▁expres
+- ▁unconcerned
+- விட்டு
+- stecken
+- ▁آپ
+- ढ
+- ▁quelquefois
+- ▁продолж
+- ▁pitiable
+- ▁mataifa
+- ▁provost
+- ▁다섯
+- ▁магчымасць
+- ▁Student
+- ▁அறிவ
+- ▁Clark
+- ▁voleva
+- ▁그렇고
+- catcher
+- ▁twirl
+- ▁inhabitant
+- ▁overwhelm
+- ▁Bernard
+- ▁fung
+- ▁municipo
+- 應
+- 汉
+- ސާ
+- ▁ҙур
+- iferous
+- igiti
+- '35'
+- ayı
+- ▁turismo
+- ▁bingi
+- ▁movi
+- ▁narration
+- ▁chercha
+- بەت
+- gestanden
+- ▁viking
+- ▁província
+- 吃
+- ண்டி
+- ▁concret
+- ▁stallion
+- ▁reprodukta
+- لة
+- ▁харч
+- ahisemo
+- ▁pedestrian
+- ▁очень
+- ▁درباره
+- ▁puissance
+- ▁없으니까
+- ▁chciał
+- ▁fiŝ
+- ▁comparable
+- appeti
+- ▁лийын
+- ▁terreny
+- ▁blitz
+- ▁پێش
+- ▁fidel
+- ▁galop
+- ▁bibazo
+- ▁ماه
+- ફ
+- ครั้ง
+- ▁nevada
+- ▁oswald
+- ▁kabisa
+- يار
+- ▁racial
+- ▁wuz
+- ▁liberated
+- ▁tooneel
+- ▁hansom
+- водзіць
+- ится
+- ▁kuliko
+- ▁erkannte
+- Austràlia
+- ಷ
+- ▁ejemplo
+- ▁exchequer
+- ▁incredulity
+- ▁memphis
+- ▁unmittelbar
+- ▁Forschung
+- ▁unravel
+- 沒有
+- 학과
+- ▁unholy
+- ▁Serie
+- ▁кум
+- ▁bedaard
+- ▁argaz
+- ▁signifi
+- ҥга
+- ▁methodical
+- beleid
+- ▁tresses
+- ాల
+- ▁Bour
+- カ
+- sleeve
+- ▁пӱ
+- ▁Kä
+- amaanyi
+- wendig
+- gemerkt
+- effray
+- ▁beryl
+- ▁zaś
+- ▁Bush
+- рова
+- ▁Columbia
+- ▁возера
+- ជា
+- ▁wollt
+- amafaranga
+- ▁اومد
+- ▁قور
+- ▁narrated
+- menyesha
+- anjou
+- ▁پہ
+- ლა
+- 他的
+- ▁devolve
+- ▁памят
+- ▁parecia
+- ▁benefice
+- ▁voile
+- রো
+- ジ
+- ћ
+- ▁Massachusetts
+- ▁cimetière
+- ▁animosity
+- ▁систем
+- ▁henriette
+- ▁bryce
+- ▁беларускага
+- ▁plaid
+- ედ
+- wallow
+- ▁hartelijk
+- ▁gibraltar
+- ▁appalled
+- ▁línea
+- ▁pamela
+- ▁патрэбна
+- ਦੀ
+- ▁lanyard
+- famed
+- presión
+- ▁rebuilt
+- เพ
+- ▁graaf
+- ▁levelled
+- втор
+- ▁خوبی
+- வரை
+- алда
+- ▁bewail
+- 带
+- ▁кӱлеш
+- ▁verbergen
+- ▁அரச
+- ▁storage
+- nęła
+- ▁سۇ
+- sluiten
+- ▁rojo
+- لارنىڭ
+- ости
+- ▁porridge
+- czny
+- ▁unhealthy
+- ▁voluptuous
+- ▁доўгі
+- ▁likelihood
+- ▁incorrect
+- ▁saturated
+- ▁yelhan
+- ▁drapery
+- duid
+- ▁Arthur
+- ▁devenas
+- ছা
+- fluas
+- ▁bookseller
+- ▁schande
+- ▁Station
+- magezi
+- ยุ
+- मान
+- ▁christened
+- ▁moonlit
+- típic
+- ▁அதை
+- ▁bankrupt
+- ▁необходимо
+- ▁irresponsible
+- ▁kaminuza
+- ▁আৰু
+- ▁jenoside
+- ▁suzanne
+- sababisha
+- ▁popped
+- ▁relativament
+- Alemanya
+- ▁equity
+- मी
+- badili
+- ▁tliat
+- ▁dynamite
+- ▁Information
+- abanje
+- ▁있나
+- nscrit
+- ▁flux
+- ских
+- ▁anniversary
+- ▁magnanimity
+- ▁некаторых
+- ▁hemlock
+- ▁tevreden
+- ▁mój
+- ▁여름
+- ▁girlhood
+- ▁шунда
+- льнікі
+- ulph
+- ▁kernel
+- ▁hungary
+- ibile
+- ▁voters
+- ▁heimat
+- ▁costumbre
+- ▁apathy
+- lösen
+- ▁옛날
+- ▁gauw
+- ▁área
+- possessed
+- ▁prétend
+- ▁laurie
+- actual
+- ▁wunsch
+- kwemera
+- ▁snore
+- basaba
+- ▁절대
+- ▁flick
+- ▁stockade
+- ▁রয়েছে
+- ▁выкарыста
+- ▁Ṛebbi
+- ▁glistened
+- ▁posséd
+- ▁bengal
+- 것
+- یاری
+- ▁englische
+- ً
+- ▁freddy
+- ▁Dazu
+- échappe
+- ம்பி
+- ▁tallow
+- ▁chasseur
+- ▁proportionate
+- نوع
+- starred
+- ູ
+- ▁practitioner
+- ▁stupefied
+- ▁Zealand
+- ▁negotiate
+- 案
+- הי
+- ▁Hinter
+- ▁Gemma
+- ನ್
+- ▁ŝanĝi
+- ବା
+- ові
+- ▁platter
+- fusió
+- ▁килде
+- ▁venomous
+- ▁darcy
+- テ
+- ކޮށް
+- ▁acknowledging
+- ๊
+- 念
+- ▁màxim
+- ▁வாழ்க்கை
+- ▁주말
+- కు
+- ▁bordeaux
+- ▁бачыў
+- ▁bluster
+- ▁persuading
+- ก่อน
+- ▁knitted
+- готов
+- ▁römische
+- ▁hammond
+- ▁خۆش
+- ▁সম্
+- পি
+- bikoze
+- drukt
+- utilitzen
+- ▁plebeian
+- ▁enquiry
+- ▁javelin
+- ▁нисек
+- ▁saunders
+- ▁scorched
+- ▁башҡа
+- મા
+- ▁ўсяго
+- ▁entièrement
+- ▁fwy
+- шука
+- ▁plaything
+- ত্ত
+- ▁multitud
+- ▁voegde
+- ▁perpetrated
+- ▁vergnügen
+- ▁epidemic
+- ▁insomuch
+- ▁proksima
+- ▁deferred
+- ▁alligator
+- ▁Nigeria
+- ▁सु
+- ▁modera
+- ▁strato
+- umusore
+- ▁strapped
+- ▁maîtresse
+- ▁continuity
+- ӑн
+- വി
+- ▁rubies
+- ▁levity
+- ▁piace
+- ▁bolton
+- ▁மற்ற
+- priva
+- ▁zorgen
+- ਖ
+- ▁Although
+- ▁addicted
+- ▁conservé
+- ▁responsive
+- ▁buoyant
+- ▁signification
+- hawk
+- ▁virtual
+- علم
+- 듣
+- ▁skiff
+- ▁ziren
+- ▁durable
+- ▁الك
+- ሪ
+- ▁inarticulate
+- ▁makumyabiri
+- ▁일주일
+- nsgesamt
+- ▁dinosaur
+- ▁responsabili
+- ▁recursos
+- ▁harangue
+- ▁superhuman
+- ▁부산
+- ▁شاید
+- Straße
+- ▁vibrating
+- aglia
+- Arc
+- ▁табы
+- ิด
+- ▁Thus
+- ้ม
+- aceous
+- ▁jurist
+- ការ
+- สําหรับ
+- ▁Hodiaŭ
+- ▁blizzard
+- ▁obnoxious
+- ▁santiago
+- 排
+- ingofero
+- ▁stimulating
+- ▁heaviest
+- ▁mutilated
+- mukazi
+- ▁Despi
+- ▁finch
+- ▁сцен
+- dessous
+- ▁oordeel
+- ▁hobby
+- ▁necesita
+- ▁Grab
+- ▁fermé
+- hagarariye
+- ▁proyecto
+- ▁purchasing
+- ▁undecided
+- ▁uruhushya
+- ▁хәҙер
+- ▁tiberius
+- migabane
+- ándome
+- ▁cologne
+- 大家
+- ▁sophie
+- ▁gdyby
+- تەر
+- ▁gracie
+- ▁acaso
+- ▁Че
+- mahon
+- ▁championship
+- ▁pending
+- झ
+- ▁prophesy
+- 센
+- ▁flossie
+- ▁bɣiɣ
+- kakiiko
+- ▁cosí
+- ▁zoolang
+- овую
+- दी
+- ▁guhindura
+- ▁frigid
+- ▁penetration
+- ▁Consell
+- ▁koncert
+- ▁Estados
+- ▁Nubwo
+- ▁международно
+- ▁kenshi
+- ▁Reihe
+- ҟә
+- rashaka
+- پرسی
+- ▁Publi
+- ▁folgende
+- ▁yaliyo
+- ▁гісторыя
+- ▁halifax
+- ▁nauwelijks
+- ▁свабод
+- Uhoraho
+- ▁Daher
+- ▁huntsman
+- ▁пыр
+- ▁spesso
+- ▁faudra
+- ▁primeiro
+- ▁trembl
+- ▁primeval
+- ▁glacial
+- ▁hillock
+- ▁hervatte
+- ▁poul
+- faß
+- ▁tegenwoordig
+- ောက်
+- ▁shortened
+- ▁psic
+- ځ
+- ై
+- ▁bewitched
+- ▁uninteresting
+- ▁актыўна
+- ▁verwonderd
+- 問
+- 班
+- ▁retrieve
+- 速
+- ▁vergangen
+- größe
+- ▁რომ
+- ▁modifi
+- ifuzaga
+- مین
+- verfahren
+- ▁escucha
+- ▁아프
+- ▁vogue
+- ▁morose
+- ▁parsonage
+- 보자
+- ியது
+- ▁gazette
+- ▁immaculate
+- 議
+- 告
+- ▁betreft
+- ▁standstill
+- ▁partiality
+- ▁தூ
+- И
+- ▁commentary
+- ▁Truca
+- ▁träume
+- ▁atmospheric
+- 탄
+- اللَّ
+- ▁hulk
+- ▁hungarian
+- ▁kalif
+- ▁gewählt
+- ▁barbarism
+- biseera
+- ▁fosc
+- weichen
+- attache
+- ▁trampling
+- ▁Oxford
+- ▁vicksburg
+- vugizi
+- Àfrica
+- ▁kkooti
+- ▁scrawl
+- 领
+- deutschland
+- ▁blackmail
+- ▁Gaël
+- zaliwa
+- ▁Wilson
+- ▁qilish
+- ილ
+- läg
+- ▁Techni
+- чки
+- speise
+- sistance
+- ▁cloche
+- ▁finster
+- À
+- маған
+- َةِ
+- ▁استفاده
+- 油
+- ▁adquiri
+- ▁astronomical
+- ▁inclouen
+- ▁chandler
+- μο
+- ▁variance
+- nyola
+- ▁springfield
+- ▁və
+- attention
+- ▁Llobregat
+- ખ
+- ሉ
+- 資
+- 洛
+- ▁possono
+- ▁tattoo
+- ▁கேட்ட
+- ▁reddened
+- ▁physiological
+- ▁vijand
+- ▁innerhalb
+- ▁сдела
+- רו
+- ▁afflict
+- ▁شیر
+- ▁tassel
+- खा
+- ▁archipelago
+- ▁nekhludoff
+- ▁scrambling
+- ▁exulting
+- ▁Txell
+- ▁tambien
+- basket
+- ▁deposed
+- ▁củ
+- '21'
+- ▁яңы
+- ▁brilla
+- ▁않고
+- ▁geometr
+- ساز
+- ▁вакол
+- ▁Unió
+- ▁retorn
+- ▁indiscretion
+- ▁diocese
+- 象
+- 該
+- ▁Ireland
+- villiers
+- ▁pivot
+- ▁Holz
+- shimisha
+- ▁Ikindi
+- ▁Было
+- 할려고
+- carriage
+- ▁aŭtoro
+- ▁ауыл
+- ▁résist
+- ▁көс
+- ନ୍
+- ھى
+- ▁lorraine
+- ▁아무래도
+- ▁hicieron
+- 印
+- ▁forgery
+- ▁penrod
+- adequa
+- ▁circo
+- ▁implication
+- พระ
+- ▁процесс
+- ▁irresistibly
+- ▁nautilus
+- ▁redoubt
+- ▁admittance
+- ▁결국
+- ▁unsteady
+- ▁aubrey
+- scripció
+- spreken
+- čí
+- lund
+- ▁historique
+- ▁frustra
+- กิน
+- vacu
+- ▁bagize
+- ായി
+- ▁calcutta
+- ▁stagione
+- 嘢
+- ▁شرکت
+- ▁insinuate
+- ▁الق
+- ▁Shar
+- ▁saucy
+- ubukungu
+- ▁fuente
+- ▁libér
+- ▁domination
+- ▁Ҡо
+- particulièrement
+- ▁sawdust
+- ▁denkbeeld
+- ▁Hungario
+- quête
+- ▁된다고
+- ▁карыста
+- ▁journalism
+- ▁সম
+- চিত
+- ▁cherche
+- ▁claudius
+- ▁excavation
+- ▁магчымы
+- ▁падабаецца
+- ▁alegría
+- ▁imitating
+- ▁dmitri
+- ▁atteint
+- 重要
+- บอก
+- ▁않나
+- ▁obgleich
+- ▁alcoholic
+- ▁succor
+- eleza
+- ▁wreckage
+- bazadde
+- ▁것들
+- ayotgan
+- เรียน
+- ▁kemm
+- 剧
+- ▁எண்ண
+- ▁Jacob
+- ▁Cambridge
+- ▁кнігі
+- ▁этап
+- ▁사람들은
+- ▁modoka
+- ▁rebelled
+- dimensional
+- ▁ailment
+- ▁ўваход
+- garri
+- ▁humph
+- ▁фарма
+- ▁человек
+- ▁demasiado
+- ▁mètode
+- ▁shuffling
+- 嗎
+- ▁Tradition
+- ▁Primera
+- brücke
+- mutungo
+- енән
+- kurikiza
+- اوی
+- ▁працягу
+- ▁Cross
+- ▁expel
+- नु
+- ▁piccadilly
+- 팅
+- ▁әллә
+- 引
+- alfabet
+- ▁карцін
+- ▁سفر
+- ▁profesor
+- ▁hackney
+- mettait
+- ▁theodora
+- ▁liberté
+- மண
+- ▁prosaic
+- ▁geschah
+- ▁homesick
+- ▁cumber
+- gelesen
+- rwanya
+- ▁procede
+- ▁blackguard
+- ▁verschillende
+- ▁waarmede
+- ▁presidency
+- ▁ballast
+- ▁Finanz
+- ▁நல்ல
+- ▁семын
+- பட்ட
+- stift
+- ▁topmost
+- ▁shanty
+- ▁deftig
+- ▁prowl
+- 야겠다
+- sieur
+- ▁nestled
+- Ο
+- ▁ephraim
+- ▁Некаторыя
+- ▁އޮ
+- ▁ggwanga
+- ▁miejsc
+- ▁듣고
+- ▁muhanda
+- ▁enhavas
+- ▁enseña
+- ▁Sinshobora
+- альная
+- ▁иаа
+- ▁composició
+- ▁gewinnen
+- ▁усім
+- ▁stigma
+- ▁gelooven
+- ▁victuals
+- 캐
+- ▁indolence
+- ▁بولىدۇ
+- getrokken
+- ▁amphib
+- ▁protestations
+- ▁algú
+- ▁selva
+- рван
+- ▁말하는
+- ▁mercer
+- ▁opzicht
+- ▁아무것도
+- ▁잠깐
+- ▁প্ৰ
+- 统
+- ▁ljewin
+- ▁iḍelli
+- ▁میکنی
+- ▁Bridge
+- wirken
+- طب
+- Serv
+- iadau
+- ▁Sydney
+- ▁coneguda
+- ▁suivre
+- 留
+- 导
+- ▁хочацца
+- ▁Historic
+- ▁cherries
+- ▁niedrig
+- әрә
+- ▁reflexion
+- skew
+- ▁venido
+- ▁utah
+- பார்
+- ในการ
+- পু
+- रू
+- لەرنى
+- ▁aggregate
+- ▁symmetry
+- 康
+- ▁descobri
+- 易
+- ▁ninguno
+- その
+- aracter
+- ▁Airport
+- ▁uniformity
+- ▁windmill
+- ▁accumulate
+- ▁없어서
+- ▁shuttle
+- ▁bitaro
+- ▁gladys
+- ▁ketch
+- ▁plenum
+- ▁Kampala
+- 찬
+- َأَ
+- ▁традыцый
+- ▁giovane
+- трымлівае
+- abayizi
+- ▁palisade
+- ▁monica
+- ▁Sign
+- ▁அதிக
+- ละ
+- ▁kreskas
+- خواهی
+- σει
+- عُ
+- ddagala
+- ம்ம
+- ▁aanzien
+- ▁weldra
+- ▁Nàdia
+- ▁begegnet
+- ▁roosevelt
+- ▁technology
+- ▁время
+- 핑
+- ▁watalii
+- ▁maddened
+- ▁maddening
+- ▁Lakini
+- ▁arsenal
+- ▁һәр
+- Ш
+- ▁detach
+- ▁Questa
+- ▁жок
+- gunstig
+- ▁jammed
+- ▁cadena
+- ▁Ре
+- ▁هێ
+- ▁கண்ட
+- ▁diplomo
+- ▁vetur
+- ▁Taylor
+- ▁vecchio
+- ▁enabling
+- ใคร
+- ▁talisman
+- 术
+- ▁новага
+- zinnig
+- ▁coś
+- ▁Unidos
+- ▁schlich
+- ▁wicker
+- czek
+- нуты
+- ардын
+- ▁думае
+- ▁زیر
+- ▁Aquell
+- рэнне
+- ▁arabia
+- 칠
+- หมาย
+- ▁아무튼
+- 节
+- ▁recommence
+- ▁rosemary
+- ड़
+- ފި
+- ▁kwandika
+- ▁corporate
+- ▁tomatoes
+- ▁Abakulembez
+- дзець
+- ▁ähnliche
+- ▁килгән
+- یەکی
+- arriver
+- aspetta
+- ткан
+- ▁Gleich
+- treize
+- ▁frenzied
+- ▁penitence
+- ▁чатыры
+- ประเทศ
+- 助
+- abahanzi
+- ▁sociable
+- ▁Malgrat
+- ▁prachtig
+- ▁wulf
+- ▁скл
+- રા
+- atanzwe
+- ▁Typ
+- ▁auspicious
+- ▁icyizere
+- ▁swarthy
+- ▁grizzly
+- 區
+- コ
+- ▁вялікай
+- ವು
+- ▁utopia
+- ▁Earth
+- ▁bianca
+- ▁diablo
+- ডি
+- ▁күтәр
+- ▁uɣerbaz
+- ▁perplexing
+- ▁부모님
+- ▁최근에
+- 층
+- ▁гаварыць
+- ▁Rand
+- jointed
+- います
+- ▁basketball
+- ▁якое
+- ▁sympathetically
+- ▁clifford
+- ▁цел
+- ▁форма
+- ▁abagize
+- ▁ҡарап
+- َمْ
+- ▁detrui
+- ▁wrathful
+- gerissen
+- ળ
+- บาง
+- ጠ
+- ▁begleitet
+- ▁diminutive
+- Е
+- ▁ruffle
+- ▁fertility
+- concentr
+- ▁Privat
+- ▁respiration
+- komunik
+- ▁lovable
+- ▁rumina
+- ▁dreamily
+- 닐
+- ▁northumberland
+- ▁பெண்
+- ▁ஒன்று
+- ▁ndashaka
+- ▁barbaric
+- ▁àwọn
+- anlage
+- ▁vertelde
+- ▁많고
+- ▁Faint
+- ▁ôf
+- ▁dynamic
+- ▁erinnerung
+- ▁Һеҙ
+- ▁marchioness
+- ▁Frankfurt
+- 约
+- ▁دەکات
+- ▁espècie
+- ▁salutary
+- ahitamo
+- ▁curving
+- ▁bulge
+- ▁торған
+- ▁soared
+- habitatge
+- cursion
+- ▁templ
+- ڪي
+- ੂ
+- 떨
+- 뛰
+- 埃
+- ▁entsprechend
+- چار
+- ▁працаваў
+- ▁sonorous
+- ▁whitney
+- projekt
+- சொல்
+- บน
+- ▁dissabte
+- ▁और
+- 왕
+- ▁réalisé
+- 候
+- ▁কৰি
+- ▁відэа
+- ▁voĉo
+- کشید
+- علي
+- ▁چشم
+- ▁grandpapa
+- 프로
+- ▁venturing
+- αρ
+- hitaji
+- ▁ведаюць
+- ▁годым
+- gefahren
+- ▁모르겠다
+- ▁أنا
+- ▁الج
+- ikki
+- ▁Mexico
+- ▁gukemura
+- прыгожа
+- ▁alyosha
+- ▁armitage
+- ▁ongeluk
+- ▁tribulation
+- ▁складаецца
+- слоў
+- ▁idiom
+- wurk
+- ▁interred
+- 레이
+- ▁chatted
+- rühm
+- ▁lutte
+- himba
+- дагы
+- ▁murphy
+- ▁securities
+- ▁ئێستا
+- ▁crammed
+- ▁slackened
+- ▁virtud
+- Honor
+- ▁газет
+- աս
+- ŝtono
+- ▁fraai
+- ▁طور
+- ▁fundador
+- ▁Ғә
+- বো
+- ▁Imperio
+- انداز
+- ▁már
+- 第二
+- ▁Otto
+- ▁incongruous
+- ▁sullivan
+- 搵
+- 聽
+- 密
+- ▁кожнага
+- ▁criança
+- ▁comparti
+- ▁китә
+- populated
+- ▁Уның
+- асобных
+- ▁breng
+- ▁অব
+- ▁іншы
+- ▁tyrannical
+- цёр
+- ▁dilapidated
+- เหมือน
+- ▁défense
+- ▁forsooth
+- 運
+- acompanya
+- ▁olivier
+- ▁zenith
+- ▁축구
+- ▁ляжа
+- ▁désert
+- ▁uhuru
+- ▁разоў
+- ▁visionary
+- ▁пэўны
+- wirtschaftliche
+- ▁disagreement
+- ▁añadió
+- ▁brittany
+- ▁circumstantial
+- ▁indefatigable
+- ▁predicament
+- ▁scrooge
+- ▁থাকে
+- impanuka
+- ▁tayoga
+- ▁untimely
+- мяні
+- ▁ئوقۇ
+- نامه
+- ▁مص
+- خیر
+- ▁pittsburg
+- ▁jarred
+- ▁Broad
+- ▁Andrew
+- ▁Regional
+- کەوت
+- ▁obsolete
+- ആ
+- ▁portmanteau
+- ▁behaving
+- 哥
+- 处
+- ▁pontiff
+- ▁dinsmore
+- ▁José
+- ▁ursache
+- killer
+- ▁Graf
+- ▁bemerkt
+- ▁ripped
+- ларҙың
+- шей
+- ▁reinforced
+- ▁puerto
+- àmbit
+- bishaka
+- ▁Quintí
+- स्त
+- geleid
+- सी
+- ▁analys
+- minota
+- ▁ئەوەی
+- ▁concoct
+- ▁démon
+- umubyeyi
+- ▁নির্
+- ▁stör
+- கிறான்
+- sanyu
+- ▁ئور
+- ▁خواب
+- ▁دەتوان
+- kapu
+- ▁agamemnon
+- ▁intermittent
+- ▁overjoyed
+- ▁jamaica
+- ▁Guinea
+- ▁domestique
+- ▁possum
+- ▁fraught
+- ▁Golf
+- ▁너랑
+- ▁좋을
+- ▁snowball
+- ▁récit
+- ▁dépend
+- ▁afscheid
+- ካ
+- ▁carthaginian
+- ▁нацыянальна
+- ▁француз
+- ▁sprawling
+- 依
+- gewohnt
+- ▁Организации
+- ▁spijt
+- کێش
+- ▁consejo
+- ▁Fru
+- ▁hapless
+- ▁nannte
+- ▁vouch
+- ▁paquet
+- чэння
+- instrument
+- ▁Hü
+- гийн
+- ▁llamaba
+- ▁plicht
+- ▁casino
+- ▁discourage
+- ▁kipindi
+- ▁Urashobora
+- ▁Literatur
+- ▁muntanya
+- ஷ்
+- ▁Хаця
+- ▁commendation
+- ▁weiterhin
+- ▁Franz
+- ӗн
+- phyl
+- IS
+- ▁bathroom
+- ▁antoinette
+- ▁circonstance
+- ▁صحبت
+- ▁자체가
+- ▁Friedrich
+- ▁Stephen
+- ▁jünger
+- ▁Мала
+- ▁만났
+- ▁prueba
+- ▁일찍
+- ▁policies
+- ▁diploma
+- ▁Lewis
+- ▁greenwich
+- เอา
+- ▁circa
+- ▁Mö
+- ▁velho
+- 집이
+- ิก
+- ▁Archi
+- ભ
+- 资
+- ▁glamour
+- ಮಾ
+- aveu
+- ирован
+- ▁quartermaster
+- feier
+- ▁strenge
+- ▁가게
+- byumva
+- наход
+- ▁Zude
+- ▁prescribe
+- ▁됐어
+- viertel
+- ▁হিসেবে
+- ▁residing
+- થી
+- ▁linio
+- ▁cubic
+- piskop
+- ▁재밌어
+- rzecz
+- ▁mlima
+- ▁presumptuous
+- 員
+- 运
+- ▁taciturn
+- ▁영상
+- ▁агульна
+- ▁retract
+- ▁왔어
+- ▁одно
+- gosto
+- ▁übernahm
+- ▁Mlima
+- rupp
+- لەپ
+- ▁nuova
+- wählt
+- ▁туман
+- 줘야
+- እ
+- ▁brougham
+- ▁igikorwa
+- ▁biographer
+- ▁ukuboko
+- ▁أي
+- ▁zilizo
+- ▁arbeitet
+- ▁paddy
+- ەیەکی
+- keeping
+- ▁zweimal
+- ډ
+- ▁twitched
+- ▁reclining
+- ▁για
+- เดิน
+- ▁tradesman
+- ▁solicitous
+- fuata
+- ▁descrit
+- wundi
+- ▁chuckling
+- ▁bedacht
+- ▁balsam
+- ubuntu
+- ▁zanjye
+- ▁knack
+- ▁anselm
+- ớ
+- ▁agrikulturo
+- enseignement
+- ▁berkeley
+- ▁scabbard
+- ▁magnified
+- 肉
+- ▁상황
+- Australie
+- ▁poussa
+- ▁indecent
+- ▁frankfort
+- ▁pollute
+- ▁signifo
+- ▁Сә
+- ▁solamente
+- ▁industrio
+- ডা
+- ▁Júlia
+- ▁despicable
+- ▁굉장히
+- ▁persevering
+- ▁irksome
+- ▁сообщ
+- ▁resuming
+- ▁merkwürdig
+- ▁portentous
+- ▁clemency
+- ▁pantalon
+- ▁esperantistoj
+- ▁падобны
+- ▁Dadurch
+- ▁adultery
+- ▁Hungara
+- nnyini
+- ▁artístic
+- napatikana
+- स्थ
+- ▁tolerant
+- มีความ
+- ▁motley
+- ӹр
+- ▁شتێک
+- тельно
+- បាន
+- ▁cinnamon
+- ▁fröhlich
+- ▁scribble
+- ▁абсалютна
+- 퇴
+- ▁mtoto
+- evitch
+- ▁habría
+- scuttle
+- ▁школа
+- ▁помнік
+- ificado
+- ▁Maur
+- ▁그날
+- ீர்
+- ▁cragg
+- ▁unprecedented
+- ▁τον
+- ▁insieme
+- ▁gryce
+- ▁그래가지구
+- examen
+- ▁insistence
+- خوان
+- ▁aŭdi
+- ▁accomp
+- ▁pareció
+- ▁fretful
+- ▁licking
+- ▁مردم
+- ▁ejaculation
+- ▁Кожны
+- attività
+- ▁maurits
+- ▁vegetarian
+- ▁Шулай
+- ▁verkauft
+- ოვ
+- ▁winced
+- ▁ліст
+- ▁cyumba
+- ▁subtlety
+- ▁untold
+- ▁Ofte
+- ización
+- ▁цела
+- ▁размешчаны
+- ▁jumble
+- ▁detriment
+- ต่าง
+- ▁Rusio
+- feito
+- ▁venom
+- ▁Früh
+- ▁뒤에
+- ▁fréquent
+- ▁opportune
+- 配
+- ▁pudiera
+- 鱼
+- ▁invading
+- ভি
+- liegenden
+- ▁يول
+- ▁ziemi
+- gebroken
+- ▁wrapper
+- ▁festive
+- ▁asunto
+- ▁stessa
+- ▁Ça
+- міністра
+- ▁promontory
+- ▁unacquainted
+- ▁undesirable
+- ロ
+- 步
+- ▁spectral
+- ▁мэта
+- авыя
+- ▁сістэма
+- ▁simultaneous
+- ▁complimentary
+- leuchten
+- adulte
+- ట్
+- ▁totdat
+- لىرىنى
+- ▁gupfa
+- ӗр
+- ▁dilute
+- honnête
+- ▁disloyal
+- ▁gingerbread
+- ▁نبود
+- 周
+- ▁default
+- ▁traddles
+- fanyika
+- ▁verbreitet
+- ▁greedily
+- ▁striding
+- ▁памыл
+- ▁proverbial
+- ახ
+- ▁meteorite
+- ▁ubufasha
+- ▁existeix
+- ▁kinsmen
+- ▁Nadal
+- ▁wabivuze
+- ▁tradició
+- ▁dronk
+- gymnasium
+- ▁bismarck
+- ▁complacently
+- ▁undaunted
+- ▁ниндәй
+- 존
+- ▁vertrauen
+- ▁ralston
+- ▁surpriz
+- ▁comunicació
+- wakoze
+- ▁linguist
+- occasion
+- ▁манын
+- pflege
+- ▁notebook
+- ▁हैं
+- ▁stelde
+- emperador
+- ▁deluded
+- સા
+- ▁channing
+- ▁راست
+- ▁redakt
+- '70'
+- кавыя
+- ▁negotiation
+- ضا
+- behalten
+- ▁inexpressible
+- ▁журналіст
+- ▁ندارم
+- ▁بگیرم
+- ▁epigram
+- ▁cartwright
+- ▁дажджы
+- ▁Disney
+- では
+- ▁Georgia
+- பதி
+- ▁Hollywood
+- ▁jacqueline
+- ▁relieving
+- ▁unmistakably
+- ▁balustrade
+- 咗
+- ▁spasmodic
+- umukunzi
+- 投
+- ▁planchet
+- mateeka
+- ▁coût
+- ▁anthea
+- ubukwe
+- ▁toledo
+- ▁reconstru
+- ▁cornice
+- цярп
+- ▁Farb
+- ช่วย
+- 結
+- ▁Прычым
+- ▁whirr
+- ouille
+- ▁Welche
+- balwadde
+- ▁деле
+- ினார்
+- isirayeli
+- ញ
+- ጥ
+- ▁Salvador
+- ▁усилия
+- 幫
+- ▁überzeugt
+- 시키
+- getrieben
+- ▁labios
+- ▁geografi
+- ▁refine
+- ▁রাজ
+- ▁걔도
+- direktor
+- ▁infano
+- ▁glaubt
+- จริง
+- ▁abhorrence
+- ▁disengaged
+- ▁наогул
+- 裡
+- ▁эмес
+- ▁цэнтры
+- ▁diagonal
+- ▁dejection
+- существ
+- ▁knaap
+- ▁receding
+- ของเขา
+- ِينَ
+- ▁semblable
+- ▁voorbeeld
+- ▁compens
+- ▁virusi
+- stappen
+- درس
+- ▁sichtbar
+- wigeze
+- ▁líder
+- ▁similaj
+- gelli
+- ▁ausgezeichnet
+- ▁aŭskult
+- ▁brilliance
+- ▁néanmoins
+- ▁undergrowth
+- 睇
+- 博
+- ▁capacities
+- ▁writhed
+- ▁detroit
+- ▁다녀
+- த்தான்
+- ▁huld
+- ▁dankbaar
+- ரோ
+- ▁roubles
+- shakisha
+- ▁verliet
+- ▁рукі
+- ▁München
+- ▁retribution
+- ▁umusatsi
+- yanditswe
+- ▁relevant
+- ▁crusoe
+- byokurya
+- inyamaswa
+- ভাবে
+- ▁بدهید
+- ▁arundel
+- '22'
+- ▁magpie
+- ▁neutrality
+- ▁Ruh
+- corp
+- ންނަ
+- ▁gulden
+- government
+- ▁européen
+- ▁Möglichkeit
+- ▁birmingham
+- ▁freqüent
+- ▁hickory
+- 雞
+- เงิน
+- ▁melodrama
+- ▁merveille
+- ▁depraved
+- ▁glazed
+- ▁Ayɣer
+- ▁ofschoon
+- ▁seaweed
+- ехалі
+- ▁wielki
+- ▁munich
+- wände
+- ئان
+- ▁Bevölkerung
+- ▁bugufi
+- ▁rappresenta
+- ▁Omwana
+- ไหน
+- ▁Друг
+- ▁gemoed
+- ▁sheldon
+- ▁еҥ
+- itangazamakuru
+- ▁choc
+- เสียง
+- ▁vessant
+- ▁occupé
+- ▁بگو
+- ▁churn
+- ▁density
+- gewinn
+- ▁elevate
+- ▁رنگ
+- ▁그다음에
+- 립
+- 派
+- ▁negligent
+- ▁brocade
+- ▁Italy
+- ▁нейкія
+- icaye
+- ▁stereo
+- ▁Kugira
+- ▁wakeful
+- 蛛
+- ৎ
+- ▁Lincoln
+- ▁Venezuela
+- ▁evaporate
+- ▁eerbied
+- ▁henderson
+- ▁evacuate
+- ▁jesús
+- ▁bauern
+- ▁беларусаў
+- глыб
+- ымаш
+- fumbi
+- صف
+- ▁heimlich
+- strafe
+- ড়
+- 접
+- ▁இவ்வ
+- ▁beckoning
+- ธรรม
+- ▁cholera
+- ▁embedded
+- ▁பாட
+- 委
+- ▁살짝
+- ▁snoring
+- ▁Lleida
+- ▁irving
+- ▁тигән
+- ານ
+- ▁få
+- ▁ranald
+- ▁aviator
+- 튼
+- myitwarire
+- ሳ
+- ▁appreciative
+- ▁confiscated
+- شخص
+- プ
+- ▁pharmac
+- 儿
+- ▁rameses
+- ▁camarade
+- λά
+- ▁burgess
+- ▁Selbst
+- ▁сін
+- ▁assimilate
+- ▁vlucht
+- быз
+- ▁толын
+- ▁aspiration
+- 이까
+- ▁makabila
+- ▁neptune
+- ▁pinocchio
+- ▁unfailing
+- ▁Science
+- Itàlia
+- 千
+- 举
+- 曲
+- ▁delacour
+- ▁incentive
+- ▁cuisine
+- จัด
+- ▁Ausbildung
+- প্রা
+- ▁가봤
+- ▁அரு
+- sexual
+- ▁bedeutende
+- homicid
+- ▁betsey
+- ▁Neniu
+- ▁Southern
+- geschoss
+- ▁올해
+- ▁famoso
+- ▁کنه
+- ▁Hifadhi
+- ▁affluent
+- ▁پاکستان
+- 갑
+- ▁nitrogen
+- ราย
+- ▁재수
+- appartement
+- பொருள்
+- loque
+- ▁voyageurs
+- ▁எழுத
+- ▁Vertrag
+- นั่น
+- ▁Seiten
+- ▁ئالدى
+- ▁reparation
+- ▁georgie
+- ubushakashatsi
+- ▁регион
+- ▁грошай
+- 馬
+- 吉
+- ▁어쩔
+- ▁tebɣiḍ
+- ▁gripping
+- ▁presuppos
+- ໍ
+- atandika
+- ▁resentful
+- টো
+- स्ट
+- ▁vreemdeling
+- Ĝ
+- ತ್ತ
+- ▁Künstler
+- ▁requisition
+- ▁миңә
+- ▁volonté
+- ▁பெரிய
+- ▁omtrent
+- ▁운전
+- ▁suffolk
+- 時間
+- mitglied
+- ▁vrijheid
+- ▁muchacho
+- ▁Commission
+- ▁appena
+- ▁беҙҙе
+- ▁haastig
+- לי
+- ▁raconte
+- ଆ
+- ▁toutefois
+- ▁Museveni
+- ▁doivent
+- ▁consonant
+- ▁irgendeine
+- ▁fumbling
+- ▁astute
+- ผม
+- ▁bombay
+- ▁diminu
+- ▁gratuitous
+- ewweḍ
+- ▁mumkin
+- ▁excelled
+- kimara
+- ▁musée
+- length
+- ▁valise
+- ▁тэма
+- ուն
+- بیر
+- ▁balancing
+- ▁brunswick
+- ▁immensity
+- ▁insensibly
+- ▁molasses
+- ▁porcupine
+- ▁vagrant
+- 렌
+- ▁Benjamin
+- ▁Saɛid
+- ▁dediĉi
+- ▁Gründe
+- ▁polizei
+- ▁verdriet
+- ▁snob
+- ▁ruefully
+- ▁craze
+- ▁Pedro
+- ▁vlug
+- ▁suprema
+- သည့်
+- ▁Google
+- ▁disillusion
+- ▁expulsion
+- ▁hallucination
+- ▁rivulet
+- ▁coquette
+- ▁zigzag
+- ▁révolution
+- ▁podobn
+- ▁stimulant
+- ிருப்ப
+- ▁demeura
+- ▁dormant
+- ▁interrogation
+- ▁subtil
+- ▁башын
+- ▁мур
+- Î
+- ▁Football
+- ▁anticipating
+- ▁resplendent
+- ▁tributary
+- ▁unoccupied
+- ▁প্রথম
+- ▁gayety
+- 气
+- ្ល
+- ▁portret
+- ▁veranderd
+- escenari
+- อาจ
+- ▁umugisha
+- بور
+- byaye
+- gemeinschaft
+- ստ
+- вёў
+- ▁Habita
+- ▁mignon
+- ▁religiosa
+- ုံး
+- मध्ये
+- ▁despondency
+- ▁imprudence
+- ▁miliyoni
+- 考
+- ఉ
+- ▁Sometimes
+- ▁utalii
+- 构
+- ▁kapitän
+- posizione
+- ▁Meinung
+- ▁보러
+- ▁много
+- ▁pertany
+- szczę
+- ▁forsook
+- ishusho
+- ▁recrea
+- ▁drückte
+- ▁cuore
+- ▁клуба
+- ▁confuse
+- ▁senyal
+- ▁cincinnati
+- ▁medelijden
+- ▁midwinter
+- 듯
+- 皮
+- ▁늦게
+- ▁эксперт
+- ೆಯ
+- opération
+- ▁bonacieux
+- ▁quoting
+- ▁Scottish
+- ▁прыйшоў
+- ▁docile
+- َرَ
+- ▁tothom
+- ▁implicit
+- ▁Einzel
+- ▁charred
+- ▁страшна
+- ▁stagger
+- ▁compétition
+- ▁eensklaps
+- ▁inveterate
+- ▁savannah
+- ▁Daneben
+- ▁якога
+- ▁تغییر
+- ▁moitié
+- 記
+- ▁noisily
+- ▁simeon
+- ▁Magí
+- ▁acquis
+- ொரு
+- เคย
+- ▁geraldine
+- ▁Corona
+- 홍
+- dzę
+- ▁күрә
+- évêque
+- ▁fumbled
+- ▁humiliated
+- ▁ocasión
+- ▁repugnant
+- ▁sacrilege
+- ▁архітэктур
+- ▁паўднёвы
+- ควร
+- ▁schüttelte
+- ▁tomahawk
+- katholisch
+- ▁Schweiz
+- ▁gleichzeitig
+- ▁presiding
+- '39'
+- ▁constructive
+- บุ
+- เจ้า
+- ▁subconscious
+- ▁fecha
+- груз
+- ▁químic
+- ▁Judit
+- ▁கட்ட
+- ಶ
+- ▁tameṭṭut
+- Illinois
+- ▁Jewish
+- ▁laundry
+- ▁면접
+- 威
+- ▁combining
+- ▁maneuver
+- ▁reprezenta
+- ▁같긴
+- ▁Profes
+- ▁thrall
+- ▁Ҡу
+- ▁Проста
+- ▁electronic
+- ூர்
+- แม
+- Ҳ
+- ଜ
+- ▁Michigan
+- ▁tragedies
+- ▁Конференции
+- ▁будынак
+- ▁decorum
+- ▁dressmaker
+- ▁curiosities
+- ▁emigration
+- ▁risc
+- ▁Чо
+- 같
+- ganwa
+- ▁기분
+- ▁junio
+- ▁갔었
+- ဝ
+- ở
+- ▁demeanor
+- ▁onslaught
+- ▁sutherland
+- ▁unforeseen
+- ▁сэрца
+- ▁geschiedenis
+- ▁ئەگەر
+- ▁impetus
+- ▁हे
+- ▁Carlota
+- ▁crainte
+- änderung
+- ▁병원
+- ▁Among
+- ▁pennies
+- ▁gutsinda
+- ταν
+- ▁kirere
+- ▁régime
+- ▁الش
+- araufhin
+- ▁Gesellschaft
+- ▁indústria
+- ▁kaskazini
+- ▁Мәҙинә
+- 他们
+- ▁uchumi
+- kitiibwa
+- ▁ایران
+- ▁минең
+- ▁casualties
+- ▁guzti
+- ▁تنها
+- ▁белән
+- ▁істот
+- ▁turc
+- wandiko
+- TA
+- technic
+- ▁connaît
+- ▁executor
+- शि
+- နှင့်
+- ▁persecutor
+- փ
+- ▁Microsoft
+- язджа
+- ▁pouvez
+- ▁escapar
+- heggi
+- ▁overhaul
+- ▁yanayo
+- ▁ripened
+- examination
+- ▁skulk
+- ▁طرف
+- ▁créé
+- throat
+- 报
+- ▁incredibly
+- 탕
+- ▁zärtlich
+- ▁bellowing
+- ▁orinoco
+- ▁precedence
+- ▁coloniz
+- ▁zocht
+- kukozesa
+- முன்
+- ଯ
+- expérience
+- shidikanya
+- ▁insbesondere
+- ▁recuerdo
+- ▁uncompromising
+- ▁كېيىن
+- ▁descriptive
+- ▁elfreda
+- ▁insidious
+- ▁مشکل
+- ▁nombrosos
+- ▁pennington
+- เซ
+- ▁가격
+- ▁coronel
+- agaragaye
+- ต์
+- ▁riveted
+- ▁його
+- ▁poppy
+- assegura
+- ▁Japon
+- leistung
+- ▁revenir
+- ▁рэч
+- алтыш
+- держа
+- ▁disreputable
+- ▁mercantile
+- 짓
+- ▁insinuation
+- 知道
+- લા
+- ▁Schloss
+- esquerra
+- misión
+- ▁verdadera
+- ▁Chemin
+- ▁gurgle
+- ▁애들은
+- ปี
+- ▁Italien
+- ίν
+- ▁лёс
+- ▁Pennsylvania
+- ▁nutmeg
+- ▁ostentatious
+- 龍
+- ▁sorprend
+- 超
+- 男
+- ▁гульні
+- ▁ядерно
+- ▁imputed
+- ▁Feuer
+- ▁paraissait
+- ▁எழு
+- নো
+- ▁revue
+- ▁трав
+- ▁miranda
+- рәт
+- ▁ئۆي
+- uvchi
+- ▁buddhism
+- ▁erklären
+- ▁mémoire
+- ▁wananchi
+- ữ
+- ▁미국
+- ▁oppressor
+- ▁үҫ
+- ▁propeller
+- ▁Presenta
+- '13'
+- ▁pêche
+- wrought
+- 하겠
+- òleg
+- ▁tombstone
+- ებული
+- mukunda
+- ▁많았
+- ுகின்ற
+- ▁Allerdings
+- ▁discomfiture
+- ూ
+- ▁egotism
+- န်း
+- ▁scythe
+- ▁ekspozici
+- ▁maarufu
+- musango
+- ▁gauntlet
+- ▁cunningham
+- ▁보기
+- ▁unfavourable
+- ▁احساس
+- ▁ਵਿੱਚ
+- 菜
+- 网
+- knüp
+- ذلك
+- ▁Girona
+- ▁observatory
+- ້າ
+- espoir
+- ▁lopez
+- ▁teilweise
+- ▁sparta
+- ▁kaŝ
+- ਼
+- ▁désormais
+- ▁робіцца
+- ▁Vinyet
+- ▁برنامه
+- 成为
+- ▁fretted
+- ▁birthplace
+- ▁caça
+- ▁çok
+- ▁ئەر
+- ▁squalid
+- 氣
+- ▁tenacious
+- 眼
+- 黄
+- 条
+- ▁bracket
+- ▁sociedad
+- ▁amatwi
+- ▁necesidad
+- ▁fiesta
+- ▁كىر
+- چوو
+- ▁fanned
+- tausch
+- yigira
+- bɣi
+- ▁Plusieurs
+- 香
+- 效
+- ▁April
+- ▁mundial
+- ▁Karume
+- 할까
+- ▁Nueva
+- berri
+- ▁vreeselijk
+- ▁hierauf
+- ▁aileen
+- ▁قىز
+- ҥгы
+- ▁خون
+- ձ
+- ٻ
+- ▁arkansas
+- ▁করতে
+- 阳
+- 如果
+- ▁Тудо
+- әсәй
+- ▁разумее
+- ▁comarca
+- ሆ
+- ▁Cafodd
+- ▁mercenary
+- ▁дагэтуль
+- ▁фота
+- 決
+- 息
+- ▁friedrich
+- ▁வைத்த
+- די
+- installation
+- actiu
+- ▁perturbation
+- zaamheid
+- ▁الص
+- ▁Industrie
+- 気
+- ▁galileo
+- ▁plejparte
+- 町
+- ▁unguarded
+- ▁пӧрт
+- ▁scorching
+- 代表
+- ▁크게
+- ија
+- ▁shunned
+- ▁emitted
+- ▁دۆ
+- residencia
+- ▁corredor
+- ▁Amsterdam
+- ▁anstataŭ
+- ▁commodity
+- ▁despondent
+- ▁prostitution
+- ▁Однако
+- ▁অনু
+- ▁እና
+- ▁тиклем
+- ▁pianoforte
+- ▁daisies
+- ▁recibió
+- ▁먹었어
+- ▁referència
+- ▁cadence
+- ▁كېلى
+- ▁trafik
+- examine
+- tariki
+- ibagirwa
+- küste
+- ▁Fakte
+- ▁unavailing
+- ▁unprotected
+- ▁незалежн
+- 뉴
+- ชอบ
+- ▁vandaag
+- 房
+- ▁sneeze
+- ിച്ച
+- ▁kwinjira
+- ▁torturing
+- děl
+- ▁erlaubt
+- ▁coração
+- ▁confiance
+- ▁hummed
+- ▁Alicio
+- ▁왔다
+- ▁ferai
+- ▁выхад
+- ▁unendlich
+- ▁걸어
+- ಚ
+- ▁credulity
+- ▁guillotine
+- ▁중학교
+- ▁commencing
+- ▁Pendant
+- 论
+- ▁archives
+- plasm
+- mahirwe
+- ▁kobiet
+- ▁fehlen
+- алтын
+- против
+- bafasha
+- ▁garibaldi
+- ▁rodriguez
+- ▁josiah
+- myitozo
+- ▁kipps
+- ▁Fahrzeug
+- ▁deffir
+- ▁Darüber
+- ▁Ĉar
+- ▁championnat
+- ▁آدم
+- ▁decorative
+- huriye
+- ▁ontario
+- ▁urchin
+- ҳои
+- ▁Шә
+- ▁consecration
+- ధ
+- ▁ўсю
+- 致
+- ▁aylmer
+- ▁símbol
+- ▁формы
+- ્યા
+- ▁طول
+- ▁gawaine
+- ▁foreground
+- ▁debout
+- ▁kredas
+- מו
+- μπ
+- ▁кид
+- Escola
+- ▁convulsed
+- べ
+- 标
+- 园
+- ▁Charlotte
+- ▁motiva
+- ▁danube
+- ▁evangelical
+- ▁가져
+- ές
+- fric
+- には
+- ▁revoir
+- ubwami
+- 해봐
+- ▁hojas
+- 축
+- ▁расійскі
+- লী
+- ▁immemorial
+- ▁прамавугольн
+- กล่าว
+- ▁seizure
+- ▁ointment
+- ▁sultry
+- ▁اینکه
+- Alf
+- ▁deftly
+- ▁كېتى
+- जे
+- архі
+- пові
+- ujju
+- ▁Howard
+- เมือง
+- ▁제주도
+- ▁Demokrat
+- ▁귀찮
+- 論
+- ޓް
+- inigeze
+- ▁нәмә
+- ▁cosmic
+- ๆ
+- ▁barcelona
+- ▁binyuze
+- ▁można
+- 낮
+- ▁inestimable
+- ▁lucifer
+- absor
+- ▁dundee
+- ▁bovendien
+- なく
+- ▁encima
+- ▁keizer
+- ▁savior
+- ▁repulse
+- ერი
+- ޗ
+- ▁uncontrollable
+- ▁песня
+- 魚
+- 托
+- 照
+- ▁чакаць
+- ▁mitjans
+- ▁جدید
+- förmig
+- ▁lemonade
+- वी
+- onekana
+- ▁promoting
+- sicherheit
+- ▁magistrat
+- могай
+- ▁سنگ
+- ▁rebuild
+- ▁كۆپ
+- ជ
+- ▁Städte
+- ▁miscellaneous
+- émotion
+- ▁bietet
+- sistir
+- ▁있어야
+- arquitectura
+- ▁Entscheidung
+- ▁unbeliever
+- ▁cucumber
+- ▁znowu
+- ▁untidy
+- ▁exterminate
+- биз
+- mufasha
+- ▁вылуча
+- ▁granddaughter
+- ▁ҡатын
+- scribing
+- ▁Kreuz
+- Б
+- ▁공부를
+- ▁exeter
+- ್ರ
+- ▁composite
+- ɛeddi
+- ▁scepticism
+- چاق
+- پەر
+- addled
+- креп
+- ▁pueda
+- ▁pessimist
+- ▁propietari
+- ▁sprinkling
+- ▁ambiguous
+- 登
+- ▁jsou
+- ▁얘가
+- ▁좋다
+- isumbuye
+- ▁hybrid
+- kompani
+- ▁kidney
+- ässig
+- czyć
+- ▁podido
+- ▁Point
+- ▁Jung
+- confidence
+- ▁خودش
+- ▁chapeau
+- ▁leidenschaft
+- สิ่ง
+- ▁Менавіта
+- эканамічн
+- ▁annihilation
+- ▁concierge
+- ▁consummation
+- ▁комплекс
+- 곱
+- 體
+- 修
+- ▁numerical
+- ▁gangway
+- ▁전공
+- aeroport
+- കാ
+- stukken
+- ▁Austria
+- ▁бачу
+- рҙар
+- ▁reverted
+- ▁exaggerate
+- ั้น
+- ▁impoverish
+- Μ
+- ▁trudged
+- ▁Samuel
+- ▁darrell
+- ▁volition
+- ▁priestess
+- océan
+- ▁아침
+- ▁blyth
+- ▁whiff
+- بخش
+- ▁disciplina
+- நாட்ட
+- ួ
+- ▁Tanganyika
+- ▁thenceforth
+- ▁behauptet
+- 학점
+- ▁pekee
+- выкарыстоўваецца
+- ୍ୟ
+- citizens
+- ▁Pacific
+- Empire
+- emploi
+- higgins
+- ▁кенә
+- ▁Republican
+- ▁opaque
+- ・
+- ഡ
+- ứ
+- ▁gepatroj
+- ▁unprofitable
+- 湖
+- ▁satirical
+- ▁decisió
+- ▁Liberal
+- itorero
+- verhältnis
+- ▁банк
+- incorpor
+- ▁flauw
+- গুলি
+- ▁chauvelin
+- ▁concurrence
+- ▁onderzoek
+- ▁delamere
+- ▁음악
+- ալ
+- ▁بولغان
+- ▁다닐
+- ▁голас
+- ▁macbeth
+- ▁ordentlich
+- kontroll
+- ্ট
+- solvent
+- ▁flaunt
+- قەت
+- ▁이상한
+- ▁distribució
+- ▁pietro
+- రి
+- ▁Sarah
+- バ
+- tamaduni
+- ڳ
+- ▁gekregen
+- 粉
+- 切
+- 模
+- ▁Guerau
+- ไฟ
+- ▁vuol
+- ▁справы
+- ▁römer
+- ▁enthalten
+- ▁glutton
+- ▁bonfire
+- ▁vicarage
+- ▁awning
+- ▁lucian
+- ▁Trump
+- ▁Chap
+- ▁pebble
+- ▁Дарэчы
+- 궁
+- ▁gobernador
+- ▁verbaasd
+- ▁دېگەن
+- せ
+- ▁cicely
+- ▁modèle
+- ▁apologetically
+- ▁dobbin
+- ▁skipped
+- ▁adalah
+- ogràfic
+- ▁Britio
+- ▁budding
+- ▁горш
+- ▁gloat
+- ρό
+- มือ
+- ▁klage
+- њ
+- verständig
+- ▁empresari
+- schlüssel
+- ▁bountiful
+- ▁providential
+- ▁Elektro
+- ▁егет
+- 索
+- ▁archibald
+- ▁trophies
+- ▁bluish
+- ▁다니는
+- 였는데
+- ▁historiador
+- ▁여기서
+- ▁Bwana
+- qeɣ
+- ▁poète
+- ▁Mbere
+- splitting
+- พล
+- δε
+- ▁rooster
+- ▁Book
+- ໃນ
+- ▁fuzzy
+- ▁senhomiĝis
+- ▁літаратуры
+- ▁sustenance
+- 台湾
+- ▁чыгун
+- ▁হয়ে
+- ▁pronunciation
+- ▁greyhound
+- ோர்
+- ▁envelop
+- ▁План
+- práv
+- eneral
+- ▁discouraging
+- ▁magnifique
+- ▁decipher
+- ▁imposible
+- ▁verzocht
+- ▁perverted
+- ▁пісаць
+- ▁chloe
+- ▁ivanovitch
+- cyclist
+- ▁trolley
+- ▁работу
+- போய்
+- ▁truppen
+- ▁hindoo
+- ▁bicaye
+- ▁stärker
+- ▁embarc
+- haufen
+- ▁Messi
+- ▁раман
+- ▁edziĝ
+- 여서
+- ▁glorie
+- ▁endémique
+- ▁insupportable
+- ▁unterstützt
+- ubutabera
+- ▁Quatre
+- ▁ساخت
+- ▁نفر
+- ▁இரண்டு
+- ▁gladiator
+- beherrsch
+- ▁ottoman
+- ▁fakto
+- ▁gedanke
+- ▁کلی
+- ński
+- unternehmen
+- característiques
+- ▁поддержк
+- ▁marjory
+- ▁unterschiedlich
+- ▁غذا
+- ▁arbeid
+- ▁housing
+- ▁influencia
+- ▁contemplative
+- ▁동네
+- ▁молод
+- ▁overhear
+- kandidat
+- ่าง
+- ▁Шул
+- ▁satisfac
+- ▁blazon
+- قانداق
+- ▁britànic
+- ▁dispelled
+- ▁unseemly
+- 견
+- autrefois
+- iyemeje
+- ாட்சி
+- กลับ
+- 社会
+- ▁bereikt
+- ▁inverted
+- ▁olympia
+- ▁diffuse
+- handkerchief
+- ▁narrator
+- казан
+- ▁dilluns
+- ▁gedurende
+- ▁hubbard
+- ▁unmolested
+- ▁worcester
+- 랐
+- ▁Karriere
+- ▁futility
+- ▁ĝardeno
+- ▁mathilde
+- ▁faubourg
+- ကျ
+- ▁condicions
+- ▁தெரிய
+- ▁alabama
+- ▁behandelt
+- ▁ibimenyetso
+- իր
+- ▁toleration
+- ▁bemerken
+- вания
+- plätze
+- ▁khaki
+- ವಾ
+- ▁enclose
+- ubaha
+- ▁osob
+- ▁replacing
+- 乡
+- 晚
+- ▁ddunit
+- ▁alcove
+- ▁évolue
+- ▁quadrat
+- ▁بله
+- нюю
+- ၏
+- Journal
+- ፍ
+- ▁Following
+- ▁indiscreet
+- ▁Можна
+- vögel
+- aleshores
+- 室
+- актив
+- ▁Kraft
+- ▁foothold
+- ▁가장
+- ▁malheureuse
+- жение
+- ▁kühn
+- esmenta
+- зілі
+- ▁Fußball
+- ▁более
+- ▁advising
+- ▁нізкі
+- ▁ràpidament
+- ▁soupçon
+- ▁egiten
+- amitié
+- ▁Biblio
+- ▁Ɣef
+- ▁керек
+- әсәк
+- ▁метад
+- ▁вуз
+- ရာ
+- stufe
+- ▁bishoboka
+- ▁Facebook
+- ▁Funktion
+- ▁depravity
+- 李
+- 甲
+- ▁schrecken
+- ▁petrified
+- ▁gaudi
+- ▁bulging
+- gâ
+- ▁parfum
+- ▁ончал
+- ▁Congress
+- ▁foretell
+- ▁economist
+- ramutse
+- ▁Renat
+- ថ
+- ▁acclamation
+- สร้าง
+- umuhanzi
+- ▁gräfin
+- ▁novembre
+- 病
+- 风
+- ▁gibbon
+- ▁nombroses
+- ंड
+- ▁obwohl
+- ▁இருப்ப
+- ▁europäischen
+- ▁crucial
+- ▁tristan
+- 些
+- ▁presidi
+- ▁күрен
+- ▁чэмпіянат
+- ▁aggression
+- ▁compatible
+- ▁immaterial
+- ▁licentious
+- ▁intusschen
+- 权
+- ▁Alfred
+- ▁mansoul
+- ▁gewoonlijk
+- écouta
+- ▁нашага
+- stärke
+- ▁Baldaŭ
+- ▁комп
+- ▁suicid
+- zugehen
+- ္
+- ▁imbecile
+- ▁многіх
+- ▁спецыяліст
+- ▁adolescent
+- ▁dergleichen
+- ▁despotic
+- entrenador
+- วด
+- ▁gebrek
+- ژە
+- ▁Gemein
+- ▁група
+- ▁flughaveno
+- ingénieur
+- ใหญ่
+- ▁fabrizio
+- ▁Stück
+- ▁hortense
+- ވީ
+- ▁execu
+- ▁throbbed
+- ▁수능
+- 現在
+- ▁hinüber
+- ▁fanaticism
+- এ
+- ▁Strecke
+- ▁Karere
+- ▁haraka
+- квіт
+- າຍ
+- หัว
+- ▁abstinence
+- ▁inequality
+- ▁многа
+- 祖
+- ▁crevasse
+- ▁Bulgari
+- ▁Soviet
+- ▁lavinia
+- ▁mwalimu
+- ▁strawberry
+- ▁tinkling
+- ▁Those
+- ▁жыве
+- десят
+- sasula
+- ▁hakurikijwe
+- ▁будаўніцтва
+- ▁상관없
+- 屋
+- ▁wafted
+- ▁chakula
+- ▁leeward
+- ▁mascot
+- ▁faßte
+- ▁carácter
+- ▁میشود
+- ▁Думаю
+- ▁undulating
+- ▁wigwam
+- ▁placard
+- ▁fruktoj
+- 没有
+- ▁yolks
+- ▁туҡта
+- ▁모든
+- ▁разгляда
+- ▁mistook
+- ▁innkeeper
+- ▁башланы
+- ▁mitigate
+- போல்
+- َعَ
+- ▁montaña
+- ▁habrá
+- ▁flirtation
+- ▁எனக்கு
+- ቅ
+- ▁embodiment
+- ▁impregnable
+- ▁thimble
+- 徑
+- ▁verändert
+- מי
+- ▁imperio
+- ▁prisonnier
+- ▁належ
+- ▁Japanio
+- قص
+- ▁antipathy
+- 꼈
+- เครื่อง
+- ▁empirical
+- ▁sigurd
+- ▁gukorwa
+- ▁guhabwa
+- ▁Antwort
+- ▁suele
+- ▁zinazo
+- ▁tayri
+- ▁mépris
+- ▁보면서
+- ▁conosce
+- ▁ofereix
+- алізава
+- ▁avanza
+- beddel
+- anyuze
+- เสีย
+- drž
+- ▁incalculable
+- ▁infatuation
+- ▁reticence
+- ▁genoemd
+- ▁Taiwan
+- ▁petulant
+- लाई
+- ្ត
+- નો
+- ▁emigr
+- technisch
+- 졌어
+- ▁avesse
+- ▁Omusajj
+- ▁stumm
+- ▁Petr
+- versammlung
+- ▁contempl
+- ြ
+- lícules
+- makimbirane
+- ▁hacienda
+- ▁irrepressible
+- ▁mafaranga
+- ▁obsequious
+- ▁Ніхто
+- 永
+- 計
+- 森
+- ▁spectrum
+- ▁animaux
+- ▁maanden
+- ▁Pollando
+- ▁gatanu
+- ouverture
+- ▁organisé
+- ▁aumenta
+- ▁princo
+- ▁காட்ட
+- ▁член
+- ▁beliebt
+- ▁malsano
+- нню
+- Anglaterra
+- মধ্যে
+- ▁unparalleled
+- ▁дрэнна
+- ▁galahad
+- جێ
+- ▁moistened
+- ▁hallowed
+- ▁даведа
+- алогіі
+- ▁conservatory
+- ▁있긴
+- стэр
+- ▁Luego
+- ▁climatic
+- ▁herzlich
+- ▁eerlijk
+- sinnig
+- vuelve
+- өнө
+- ▁reçoit
+- ഭ
+- 㗎
+- ▁imputation
+- ▁Hispanio
+- 欧
+- gedragen
+- ▁hispana
+- hänge
+- ▁Brücke
+- апярэ
+- ▁carreus
+- ▁suceso
+- ▁brittle
+- πά
+- ▁چهار
+- Claude
+- ▁religioso
+- ଅ
+- ▁dimecres
+- பொழுது
+- ▁seguretat
+- ▁unheeded
+- 低
+- ▁turnbull
+- 担任
+- ▁draperies
+- ▁хацелася
+- ▁stürzte
+- dzić
+- Europe
+- ึก
+- wigisha
+- ィ
+- ខ
+- ▁chipmunk
+- ▁phronsie
+- ▁sensuous
+- ▁힘든
+- ▁ప్ర
+- ▁espalda
+- ▁erscheint
+- ▁hungaroj
+- ▁হয়
+- ▁вӱд
+- ▁놀러
+- prowad
+- ▁steamship
+- ▁entomb
+- ▁officiel
+- gratul
+- ▁канцы
+- знаў
+- გან
+- ▁ٿي
+- yereka
+- 힘
+- ▁gratuit
+- ▁versteht
+- ▁لباس
+- ▁cudgel
+- ▁barometer
+- ▁inglese
+- ނަށް
+- ▁acclaim
+- ▁optimism
+- ▁چیست
+- ▁bourgeoisie
+- செய்
+- twixt
+- kulima
+- ▁fortify
+- ▁দল
+- ম্প
+- ▁Petit
+- ▁gnome
+- ပါသည်။
+- ▁Abatutsi
+- ▁experiencing
+- ▁omnipotent
+- ▁існуе
+- ▁стаіць
+- ▁byanjye
+- ▁buhoro
+- ▁나올
+- нструкц
+- ▁substitution
+- ▁administrator
+- ▁explode
+- ▁außerordentlich
+- ▁controversial
+- ▁rouletabille
+- ▁desselben
+- 底
+- ▁malmulte
+- ▁हु
+- ▁stagnant
+- ▁Außen
+- ▁نگاه
+- ▁València
+- ▁выглядзе
+- oficina
+- ▁asocio
+- ffordd
+- ▁evasion
+- ihuta
+- ▁nachricht
+- ульт
+- ▁aufmerksam
+- マ
+- ▁aangenaam
+- ▁appliances
+- ṛeɣ
+- ▁smeek
+- ▁Sinshaka
+- ▁verklaar
+- organització
+- 해가지구
+- Α
+- ዎች
+- ▁regeneration
+- 빡
+- ▁Thank
+- ▁debajo
+- ต้องการ
+- 唔好
+- ▁beaufort
+- ▁carrots
+- ▁sacrist
+- ▁squatted
+- ▁premiojn
+- ▁frappa
+- ▁авто
+- accepta
+- mererwa
+- ໍາ
+- profesoro
+- ಆ
+- ▁Mashariki
+- ▁cavendish
+- ▁divendres
+- ▁Puerto
+- 堡
+- ▁delvile
+- ▁pungent
+- sickness
+- ▁facetious
+- ▁brightening
+- ▁młod
+- ▁harcourt
+- exerce
+- ▁தோன்ற
+- ساڭ
+- ம்பு
+- چین
+- American
+- ▁monaĥejo
+- ▁expiration
+- ▁quadruped
+- 츠
+- ▁Lösung
+- 律
+- 私
+- ▁prolific
+- ▁shefford
+- ▁effusion
+- ▁três
+- ▁tamurt
+- ▁besaß
+- ▁scarred
+- ▁сыҡты
+- adigan
+- ▁ordain
+- ▁சீ
+- ikirere
+- espérance
+- ୀ
+- ự
+- ▁Acuɣeṛ
+- ▁sorpresa
+- ▁vicissitudes
+- ▁entweder
+- ▁alchemy
+- 师
+- ▁lillian
+- ▁athwart
+- ▁typewriter
+- ▁Yavuze
+- 所有
+- ▁movable
+- ▁hombro
+- ▁établi
+- ▁bereft
+- ▁spook
+- ▁Quart
+- esclave
+- ောင်း
+- ▁தர
+- rühren
+- চি
+- ▁ejerc
+- 科学
+- ▁distemper
+- byiringiro
+- undzwanzig
+- ▁Memorial
+- ▁efficacious
+- ▁equanimity
+- ▁unsuspected
+- verwaltung
+- ▁Quixot
+- ▁katolika
+- ▁worauf
+- ▁varietat
+- ▁amherst
+- ▁unbecoming
+- ▁хлеб
+- ▁hamsini
+- ▁그걸로
+- ▁Olw
+- ్ర
+- ▁wechselte
+- hirika
+- ▁stifle
+- ބާ
+- europäische
+- 혼
+- ▁morocco
+- ▁protruding
+- muwendo
+- ▁lisbon
+- สิ่งที่
+- ▁بدون
+- ▁sleet
+- ▁сімвал
+- ▁filtr
+- ▁один
+- ޢ
+- Índia
+- ਨੂੰ
+- ▁recapture
+- ▁حو
+- ▁establir
+- 住喺
+- ▁overtuigd
+- ▁figurative
+- ▁heißen
+- kkiriza
+- ▁revint
+- rócił
+- insectes
+- ▁Hamilton
+- ▁Zwischen
+- ▁preoccupation
+- ▁rousseau
+- ▁thackeray
+- 쇼
+- ▁carnival
+- ▁dimarts
+- icyumweru
+- ▁наведва
+- ข้อ
+- shyushy
+- scritto
+- noonye
+- 젠
+- większ
+- ▁pouvaient
+- ▁unfavorable
+- 观
+- ▁சோ
+- دۇر
+- ▁converge
+- stärkt
+- ходзіў
+- ാണ്
+- 준다
+- 깔
+- ▁nshingano
+- ▁secrétaire
+- ▁wszystkie
+- ▁είναι
+- ▁할머니
+- ▁pembroke
+- ▁카톡
+- 맘
+- ▁يەنە
+- läufig
+- дзясят
+- ▁vengono
+- สาม
+- ▁poesia
+- ▁dowry
+- wärme
+- ոն
+- ▁Бел
+- ველ
+- უნ
+- ▁glaci
+- ህ
+- ▁clumsily
+- ▁compunction
+- ▁romianoj
+- ▁гульня
+- ▁lecoq
+- 张
+- ▁헤어
+- Д
+- 港
+- 將
+- 争
+- 似
+- 许
+- 쪼
+- ଶ
+- ഇ
+- ム
+- 足
+- 링
+- 麻
+- 帝
+- ਡ
+- 质
+- 隆
+- ӯ
+- 불
+- ផ
+- ភ
+- 값
+- 種
+- 沒
+- ષ
+- 贝
+- ဒ
+- ഹ
+- 职
+- 놔
+- 泰
+- 亲
+- 题
+- 탁
+- 覺
+- 牙
+- 염
+- 哈
+- 농
+- 防
+- ຸ
+- 见
+- 推
+- 철
+- 松
+- ဂ
+- 友
+- 版
+- 예
+- ず
+- 讓
+- ዳ
+- 렵
+- 播
+- 叫
+- 皇
+- 량
+- 침
+- 春
+- 尾
+- チ
+- 蛋
+- ଡ
+- 飲
+- 兵
+- П
+- ଥ
+- 组
+- 갖
+- ڙ
+- 介
+- 诺
+- 样
+- 富
+- 护
+- 픈
+- ď
+- 票
+- 蒙
+- 參
+- 别
+- 텐
+- 錢
+- 깝
+- 쓴
+- 척
+- パ
+- В
+- 宣
+- 꿈
+- 雨
+- 洋
+- ష
+- 跟
+- 암
+- 连
+- 達
+- 旅
+- 섭
+- 鲁
+- 界
+- 除
+- 项
+- ắ
+- 万
+- 历
+- 늦
+- ψ
+- '&'
+- ओ
+- ಧ
+- 辖
+- 察
+- 奇
+- 精
+- 久
+- ຖ
+- 状
+- ờ
+- 닭
+- 蕨
+- 규
+- 겁
+- Ա
+- 餐
+- 련
+- 実
+- 记
+- 消
+- 嚟
+- Σ
+- 황
+- 麵
+- 킹
+- ـ
+- 対
+- グ
+- 郡
+- ᱩ
+- ޙ
+- 쳤
+- 嘉
+- 授
+- '%'
+- 務
+- 伦
+- 复
+- ஃ
+- 率
+- 边
+- ઈ
+- 良
+- 남
+- Ṛ
+- ಇ
+- ጋ
+- 澳
+- 兴
+- ዋ
+- 警
+- 咩
+- 設
+- 곡
+- 爱
+- ಎ
+- 變
+- 麼
+- 효
+- 架
+- ニ
+- 找
+- 蟹
+- 卫
+- У
+- 航
+- 独
+- 声
+- 纪
+- എ
+- శ
+- デ
+- ဲ
+- ሚ
+- 素
+- 强
+- 豆
+- 액
+- 완
+- Ӱ
+- 梅
+- 톡
+- ഴ
+- 껴
+- 먼
+- ໂ
+- 內
+- 竹
+- 報
+- ャ
+- 奖
+- 击
+- 愛
+- ଷ
+- 朗
+- ଧ
+- 创
+- 쎄
+- 例
+- 塘
+- 難
+- 훈
+- 施
+- 닌
+- 認
+- 企
+- 跑
+- 姆
+- ਧ
+- 寺
+- ቶ
+- 济
+- ធ
+- ஈ
+- 停
+- 装
+- ほ
+- ໄ
+- ウ
+- 널
+- メ
+- 另
+- ӳ
+- ך
+- 问
+- 덜
+- 給
+- ଭ
+- 莫
+- 营
+- 믿
+- 強
+- 返
+- 隻
+- 决
+- 俄
+- ഉ
+- ၊
+- 恩
+- 깜
+- 瑞
+- ռ
+- 迪
+- 注
+- ಥ
+- 連
+- 办
+- 덕
+- 턴
+- 笑
+- 库
+- 롤
+- 典
+- ਫ
+- ઇ
+- 姐
+- 흥
+- 낼
+- 假
+- 떡
+- ẽ
+- 升
+- 紅
+- 총
+- 伯
+- 離
+- ኛ
+- ޤ
+- 曼
+- ៉
+- 꿀
+- 佛
+- ဟ
+- ュ
+- 较
+- ճ
+- 닥
+- 뿌
+- ឹ
+- 害
+- 席
+- 艺
+- キ
+- 從
+- 热
+- Ž
+- ణ
+- 習
+- 酒
+- 农
+- ŷ
+- ಭ
+- 亞
+- 媽
+- 엄
+- ោ
+- ặ
+- 꿔
+- 红
+- ঙ
+- 끔
+- ښ
+- 待
+- 짤
+- ሮ
+- 쌓
+- 웬
+- 試
+- 丁
+- 選
+- ᱟ
+- ఆ
+- ଳ
+- 唐
+- 뻐
+- 鬼
+- 啊
+- ű
+- ొ
+- 術
+- 興
+- 잤
+- 承
+- ఇ
+- 벽
+- 끌
+- 컨
+- 喜
+- 낸
+- 득
+- 背
+- Ĉ
+- 엽
+- ږ
+- 耳
+- భ
+- ൊ
+- 督
+- 控
+- 准
+- 셋
+- 數
+- 止
+- オ
+- 찰
+- 열
+- ồ
+- 렇
+- ブ
+- ፈ
+- 믄
+- 條
+- 롱
+- 什
+- 差
+- 述
+- 圍
+- 夜
+- 顿
+- 並
+- 戏
+- 쌍
+- 句
+- 温
+- 范
+- છ
+- 探
+- サ
+- 守
+- 拍
+- 跌
+- ຢ
+- ਭ
+- 善
+- 据
+- ઉ
+- ጣ
+- 树
+- 쟁
+- 녀
+- 람
+- ഫ
+- ຽ
+- 帶
+- 令
+- 馆
+- 雪
+- 賣
+- ୟ
+- 测
+- ሊ
+- 狗
+- 副
+- 备
+- 给
+- ჰ
+- 즐
+- 突
+- օ
+- 盟
+- Ε
+- ‍
+- 橫
+- 혜
+- 統
+- 医
+- Ṣ
+- 뜻
+- 协
+- 层
+- 玩
+- ţ
+- 막
+- 乌
+- 话
+- 孩
+- 丹
+- 射
+- 离
+- 녹
+- 陆
+- 试
+- þ
+- ൈ
+- 튀
+- 庭
+- 浪
+- 뻔
+- ଚ
+- 組
+- 銀
+- へ
+- 宁
+- ည
+- 켜
+- 樂
+- 創
+- ՝
+- 處
+- ժ
+- 终
+- 且
+- 坐
+- 臺
+- 헤
+- ሎ
+- 双
+- 팬
+- 졸
+- 랜
+- 午
+- ビ
+- 写
+- 短
+- 证
+- Í
+- 념
+- 萬
+- 让
+- 涌
+- 單
+- 飞
+- ੍
+- ゃ
+- 沃
+- 季
+- 싱
+- 쭉
+- ợ
+- 几
+- ჭ
+- 楼
+- 植
+- 빈
+- 耶
+- 志
+- 极
+- 깐
+- 转
+- ஸ
+- 橋
+- 弟
+- 响
+- ଇ
+- 觀
+- 移
+- 封
+- 夏
+- 划
+- 瓜
+- 費
+- 켓
+- 均
+- የ
+- 轉
+- 藏
+- 君
+- 솔
+- 싼
+- హ
+- ഥ
+- 観
+- ါ
+- 怎
+- 船
+- 随
+- 푸
+- 云
+- 랬
+- 島
+- ừ
+- 章
+- 微
+- 맡
+- 權
+- 聲
+- 环
+- 풍
+- 養
+- Ú
+- ଉ
+- Đ
+- 骨
+- ଛ
+- 端
+- ں
+- 털
+- ៃ
+- 既
+- 爸
+- 虫
+- 录
+- 겹
+- 뀌
+- 롯
+- 親
+- 干
+- 评
+- 仍
+- ਥ
+- 뿐
+- 確
+- 婆
+- 잉
+- 師
+- 約
+- エ
+- 邦
+- 웨
+- ೈ
+- ॅ
+- 炸
+- 義
+- 埔
+- ໃ
+- 送
+- ђ
+- 规
+- 폭
+- 洪
+- 終
+- 빙
+- 錯
+- 桥
+- 户
+- ฤ
+- 休
+- 革
+- Η
+- 奶
+- 屯
+- ๋
+- 땡
+- 苦
+- 含
+- 调
+- ሩ
+- 華
+- ょ
+- 質
+- Ọ
+- 散
+- ڌ
+- Γ
+- 核
+- 冷
+- Ә
+- 蒂
+- 칼
+- 辦
+- 急
+- ổ
+- 냉
+- 银
+- 免
+- 発
+- 续
+- 吸
+- 類
+- 롭
+- 值
+- 冠
+- 习
+- 炒
+- 众
+- 么
+- 款
+- 遗
+- ሁ
+- 細
+- 费
+- 唯
+- 旗
+- 若
+- 哋
+- 胡
+- 멘
+- 張
+- 坦
+- 슷
+- 슈
+- 엑
+- ぎ
+- 埋
+- 宋
+- 降
+- 킨
+- ޚ
+- ᱮ
+- 調
+- 視
+- 婚
+- 럴
+- 编
+- 右
+- 帕
+- 增
+- 탔
+- 홀
+- 產
+- 互
+- 넓
+- 킬
+- 抗
+- む
+- ݙ
+- 샤
+- 摩
+- ຶ
+- څ
+- 螺
+- 尚
+- 刻
+- 盛
+- ଁ
+- ኖ
+- 떼
+- 죄
+- 透
+- 围
+- ੜ
+- ቱ
+- 網
+- 冰
+- 呼
+- आ
+- Ò
+- ঝ
+- 징
+- 聞
+- 館
+- 黎
+- 許
+- 予
+- 细
+- ナ
+- 雄
+- 概
+- ៅ
+- 備
+- 飛
+- 肯
+- 糕
+- 젤
+- ご
+- 搭
+- ၀
+- 咪
+- ダ
+- 規
+- 흔
+- ݨ
+- ङ
+- 岸
+- আ
+- 玉
+- 股
+- 율
+- 樓
+- ዝ
+- 謝
+- 雲
+- 굴
+- 畫
+- 礼
+- 똑
+- 즌
+- ね
+- 延
+- 继
+- 陈
+- Δ
+- 健
+- 熱
+- 序
+- 显
+- 較
+- 酸
+- 救
+- ဉ
+- 緊
+- 读
+- 弗
+- 兒
+- 團
+- ̍
+- 눠
+- 냥
+- ズ
+- 準
+- 盾
+- ቃ
+- び
+- 숨
+- 콩
+- Г
+- 认
+- ቢ
+- 衣
+- 侧
+- 앉
+- ̣
+- 破
+- 딜
+- 號
+- 꽃
+- 흐
+- 鐘
+- 址
+- 쓰
+- 泊
+- 菊
+- 夠
+- 退
+- 责
+- 옆
+- 乜
+- 仁
+- ŋ
+- ٔ
+- 番
+- 隊
+- 닝
+- 軍
+- 词
+- 컴
+- 毒
+- ሲ
+- 슨
+- ऊ
+- 極
+- 職
+- 렀
+- ェ
+- 轻
+- ຜ
+- 紫
+- 燒
+- 滿
+- 宫
+- 媒
+- 앙
+- 컬
+- 適
+- 追
+- 异
+- げ
+- 씀
+- 策
+- 谷
+- ョ
+- ฆ
+- 洗
+- 判
+- 攻
+- ዊ
+- 韩
+- 仲
+- 采
+- ឡ
+- +
+- 標
+- 廣
+- 雅
+- 负
+- 蝶
+- 吧
+- 末
+- 練
+- 營
+- 板
+- 佬
+- Ó
+- 付
+- 될
+- 醫
+- 庆
+- 确
+- ќ
+- 远
+- ઝ
+- 際
+- 遊
+- 盖
+- ዘ
+- 赤
+- 泽
+- 纽
+- 归
+- ポ
+- 콜
+- 籍
+- 享
+- 則
+- 델
+- 압
+- 콘
+- 署
+- 矛
+- 舰
+- ଂ
+- 亡
+- 材
+- 杀
+- င
+- 宝
+- 遇
+- 旺
+- 童
+- 茨
+- 똥
+- 뚫
+- 쌀
+- 乾
+- 黃
+- 몬
+- 씻
+- 映
+- 赫
+- 态
+- ଖ
+- 걘
+- 벤
+- 춤
+- 申
+- ጊ
+- 俾
+- 供
+- 批
+- 征
+- 奈
+- ಒ
+- ᱱ
+- 앤
+- 們
+- 益
+- எ
+- 령
+- 遠
+- 寫
+- ఎ
+- 汽
+- 왠
+- 폴
+- Ҫ
+- 忙
+- ミ
+- 菲
+- 낀
+- 哲
+- ः
+- 랄
+- ઘ
+- 优
+- ঞ
+- 專
+- 荃
+- 럭
+- 캠
+- 維
+- 률
+- 봄
+- 웹
+- 폐
+- 届
+- 舅
+- 隔
+- 럼
+- ķ
+- 깊
+- 덴
+- Մ
+- ឺ
+- 墓
+- 溪
+- ネ
+- 舞
+- ఫ
+- 努
+- ツ
+- 뮤
+- 痛
+- 驻
+- 杨
+- ử
+- 옥
+- Ĵ
+- 秘
+- 續
+- 刺
+- 仙
+- ঢ
+- 엘
+- 객
+- 岩
+- 쁠
+- 艾
+- Ż
+- థ
+- 볶
+- Հ
+- 怕
+- 答
+- 池
+- ベ
+- 藤
+- 눌
+- 葵
+- $
+- 랩
+- セ
+- 뷰
+- 智
+- Â
+- 짱
+- ڀ
+- 爆
+- 符
+- 룸
+- 윤
+- 템
+- 혈
+- 充
+- 积
+- अ
+- 咀
+- 땅
+- 벨
+- 泡
+- ᱢ
+- 켰
+- 血
+- 丽
+- ዲ
+- 柏
+- 삭
+- ٺ
+- 捷
+- 卷
+- 댓
+- 姓
+- 딘
+- 餅
+- 聯
+- 激
+- 齊
+- 杜
+- 変
+- 慢
+- 討
+- 헐
+- 默
+- ケ
+- 兜
+- 泳
+- 眾
+- 冬
+- 宜
+- ฎ
+- 剛
+- 津
+- 聚
+- 멍
+- 唱
+- ۍ
+- 벗
+- 셀
+- ះ
+- 峰
+- 홉
+- 卢
+- 韦
+- 巨
+- ০
+- ጉ
+- 尖
+- 咖
+- ؛
+- 릿
+- ੌ
+- 節
+- 妳
+- 컵
+- ዜ
+- 傳
+- 乎
+- 某
+- Ё
+- ऱ
+- 텔
+- ሞ
+- ڄ
+- 놈
+- 엠
+- 刘
+- 優
+- ソ
+- 穆
+- 僕
+- ऑ
+- Ś
+- 雀
+- 坑
+- 潮
+- 轨
+- 拜
+- 满
+- 验
+- 쟤
+- 픽
+- 载
+- 杰
+- ఒ
+- 嶺
+- ᱚ
+- 낭
+- 荷
+- 貢
+- 迎
+- 谢
+- 暗
+- 삶
+- 總
+- 幕
+- 价
+- 吞
+- Έ
+- ၁
+- 趣
+- ଼
+- ỏ
+- 밍
+- 벅
+- 借
+- 牌
+- 릉
+- ሌ
+- 甘
+- 鐵
+- 튜
+- 價
+- 翼
+- 哪
+- Ս
+- ័
+- 댄
+- 圆
+- ቦ
+- 倒
+- 歡
+- 援
+- 珠
+- 豬
+- ぐ
+- 堅
+- 균
+- 沿
+- 糖
+- 避
+- 摄
+- 알
+- 関
+- ハ
+- ኑ
+- 펜
+- ಫ
+- 桃
+- 丰
+- 輕
+- 灯
+- 滑
+- 헌
+- 妹
+- 描
+- 贵
+- 腳
+- 協
+- 暴
+- 毕
+- ዛ
+- 井
+- 積
+- 춘
+- 秀
+- 难
+- 杯
+- 灵
+- ボ
+- 葡
+- 钱
+- ピ
+- 엉
+- 穿
+- 途
+- ಉ
+- 負
+- Ն
+- 잃
+- ᱠ
+- 断
+- 震
+- 闻
+- 拥
+- இ
+- 紀
+- 豪
+- 삐
+- 枝
+- 鸟
+- 홈
+- 孙
+- ף
+- 춥
+- 睡
+- 旋
+- 態
+- 药
+- 擔
+- 舊
+- 罪
+- ủ
+- 챙
+- ਯ
+- 柱
+- 漫
+- 麦
+- 胜
+- ဗ
+- ģ
+- ჟ
+- 섞
+- ằ
+- 薄
+- 旁
+- 严
+- 静
+- ៊
+- 빛
+- 導
+- 畔
+- 幸
+- 朱
+- ሽ
+- 況
+- 걷
+- 싹
+- 흘
+- 疑
+- 余
+- 藓
+- 缺
+- 预
+- ኮ
+- ガ
+- 频
+- 虎
+- 桶
+- 錄
+- 題
+- 乘
+- 掉
+- 操
+- 靠
+- 监
+- 構
+- ៍
+- ฬ
+- ぶ
+- 萄
+- ਏ
+- 껄
+- 닫
+- 뒷
+- 占
+- 딱
+- ዚ
+- 陵
+- 좌
+- 略
+- 殿
+- 须
+- 皆
+- 햄
+- 례
+- 养
+- ጀ
+- ᱫ
+- ഒ
+- 翠
+- 誰
+- 蘭
+- 찜
+- 喇
+- 阶
+- 芬
+- 圖
+- 쏘
+- 촬
+- 柳
+- 둥
+- 힌
+- 役
+- 旧
+- Ì
+- 恐
+- 染
+- 析
+- 检
+- ዓ
+- ዙ
+- 束
+- 兹
+- 楚
+- 鄉
+- ঃ
+- 밝
+- ฑ
+- 預
+- 単
+- ៀ
+- ỉ
+- 곧
+- 昂
+- 孔
+- 洞
+- 厂
+- 软
+- 圾
+- 垃
+- Қ
+- 腹
+- 盘
+- 弹
+- 섬
+- 滨
+- 宇
+- ペ
+- 齐
+- 甚
+- 灰
+- ။
+- ޞ
+- 齋
+- 묻
+- 틱
+- 佐
+- 盤
+- 御
+- 讲
+- ୂ
+- 寄
+- 彩
+- 證
+- 餸
+- 醒
+- 큐
+- 聖
+- 危
+- 菌
+- அ
+- ঐ
+- 낳
+- Җ
+- ళ
+- ひ
+- モ
+- 泛
+- 请
+- 힐
+- Ł
+- 粒
+- 顺
+- 辑
+- 鼠
+- 덩
+- 辛
+- Э
+- ฐ
+- ዎ
+- 項
+- 넷
+- 첨
+- 팩
+- 露
+- 吴
+- ́
+- 庙
+- 臣
+- 郎
+- 症
+- 混
+- Β
+- ሜ
+- 桑
+- ಖ
+- 搞
+- 꿨
+- 須
+- 敦
+- 歴
+- 願
+- 隶
+- 腊
+- ᱾
+- 탑
+- 撞
+- 简
+- 밴
+- ጎ
+- 舉
+- 廷
+- 觉
+- 佳
+- 鹿
+- ឆ
+- ワ
+- 흰
+- 淡
+- 執
+- 幼
+- 迫
+- 顶
+- 仅
+- 磡
+- 泉
+- «
+- ዶ
+- ፊ
+- 꽂
+- 떴
+- 툰
+- 羅
+- 昌
+- 坡
+- 帮
+- ሃ
+- ቡ
+- 칭
+- 妻
+- 获
+- 꼐
+- 珍
+- ٠
+- 驗
+- 슴
+- 仕
+- 吓
+- 乙
+- 董
+- 땠
+- 몽
+- 엇
+- 困
+- 佩
+- 艇
+- 違
+- 順
+- 锦
+- Ν
+- 劳
+- 晋
+- 핀
+- 戈
+- 課
+- 괴
+- 둔
+- ֆ
+- ഞ
+- 码
+- 势
+- 横
+- 週
+- ১
+- 랙
+- 握
+- 乳
+- 斗
+- 炮
+- Х
+- 唇
+- 끗
+- 램
+- ഏ
+- 盡
+- 秋
+- 络
+- 緩
+- 칸
+- 鮮
+- 渡
+- 崎
+- 套
+- 败
+- 梁
+- 납
+- 휘
+- 浙
+- 荣
+- 兄
+- 熙
+- 簡
+- 壁
+- 農
+- 压
+- 々
+- 異
+- 鞦
+- 틴
+- 厅
+- 融
+- 霍
+- 翻
+- 낙
+- 韆
+- 꼴
+- 컸
+- 핫
+- 斷
+- ホ
+- 况
+- 轮
+- 识
+- 뜬
+- °
+- 湯
+- Ṭ
+- 톤
+- 浮
+- 産
+- 汗
+- 临
+- 兼
+- ၍
+- 렁
+- 忍
+- 犯
+- 迷
+- З
+- ૃ
+- 협
+- 尊
+- 甜
+- 또
+- 銅
+- 꼼
+- 측
+- 탐
+- 怪
+- 折
+- 畀
+- 跳
+- 響
+- 働
+- 残
+- 玛
+- Ф
+- 換
+- ڇ
+- 喎
+- 貴
+- 팡
+- 퀴
+- 咸
+- 頂
+- 逃
+- 附
+- 按
+- 뼈
+- 찐
+- 缽
+- 宿
+- 勝
+- 粗
+- 깃
+- 盪
+- 領
+- 魏
+- 纲
+- 섯
+- 護
+- 亮
+- 需
+- 깎
+- ಈ
+- ቻ
+- 讀
+- 驚
+- 抽
+- 鑼
+- 熟
+- 输
+- 啡
+- 썰
+- 傷
+- 訊
+- 陣
+- 紙
+- 忘
+- 伴
+- 呈
+- 执
+- 蛇
+- 没
+- 凤
+- ዕ
+- Ե
+- ਘ
+- ᱤ
+- 捕
+- 奏
+- 际
+- 勢
+- 惠
+- Щ
+- ᱧ
+- 禮
+- ᱛ
+- 뒀
+- 庄
+- 爵
+- ୱ
+- 晏
+- ఈ
+- ኒ
+- 喝
+- 낄
+- 혔
+- 瑟
+- „
+- 앨
+- 暖
+- 遭
+- 距
+- 囉
+- 弱
+- 换
+- ቤ
+- 꽁
+- 팠
+- 놀
+- 혹
+- 扎
+- ጅ
+- 亂
+- ऐ
+- ዱ
+- 샵
+- ቁ
+- ゲ
+- 닮
+- 液
+- Ј
+- 裝
+- 説
+- ዩ
+- 컷
+- 祥
+- 床
+- 袋
+- 雖
+- 圈
+- 琴
+- 贡
+- ノ
+- 끓
+- 階
+- 蝦
+- 鲜
+- Բ
+- 렴
+- 嘴
+- 짬
+- 冲
+- 긍
+- ቆ
+- 턱
+- 庫
+- 復
+- 芒
+- 財
+- 徒
+- 冈
+- ギ
+- 杉
+- 刀
+- Õ
+- 偏
+- 릇
+- 钟
+- 製
+- 掌
+- ଠ
+- 啱
+- 陽
+- 핸
+- '|'
+- ଫ
+- 덤
+- 쿨
+- 蓝
+- 售
+- 楽
+- 納
+- 빤
+- 밑
+- 級
+- 낚
+- 吕
+- 夢
+- 戲
+- 针
+- ቹ
+- ኪ
+- 劃
+- ゆ
+- ヤ
+- 练
+- 献
+- 겸
+- 뚱
+- ざ
+- ៏
+- 寶
+- ฒ
+- 厄
+- 柴
+- 脉
+- 廳
+- Є
+- ਝ
+- 끈
+- 얀
+- 樹
+- 腦
+- ፋ
+- 쿄
+- ぬ
+- Ԑ
+- 磅
+- 앴
+- 據
+- 補
+- 烈
+- 続
+- 固
+- 讯
+- 枪
+- 硬
+- ဏ
+- ᱭ
+- 갚
+- 斑
+- 腐
+- 胞
+- ဇ
+- 펴
+- 昨
+- 晒
+- 載
+- 脑
+- 篇
+- 尤
+- 苔
+- Կ
+- 멜
+- 碼
+- 戴
+- 丝
+- 招
+- 险
+- 奉
+- 羽
+- 侯
+- ቋ
+- 戰
+- 흡
+- 沉
+- ቲ
+- 뚜
+- 恒
+- 湾
+- 堆
+- 逐
+- 绝
+- 잘
+- ሱ
+- 쩍
+- ጨ
+- 姑
+- 隨
+- 勤
+- 袭
+- 繁
+- ァ
+- 乔
+- 찔
+- 函
+- 涅
+- 吹
+- 赖
+- 悪
+- 弄
+- 殖
+- 研
+- 侵
+- ሬ
+- 岡
+- 葉
+- ၂
+- ẫ
+- 責
+- 덮
+- 粥
+- 歩
+- 燈
+- 挑
+- உ
+- 잖
+- ڃ
+- 뽀
+- 혁
+- Я
+- 闆
+- 覚
+- 갠
+- 뺐
+- 얇
+- 測
+- 梯
+- 诸
+- অ
+- 잊
+- 貨
+- 厚
+- 欢
+- 补
+- ザ
+- 갤
+- 棒
+- 爭
+- 绿
+- 雙
+- 裁
+- 臨
+- ਛ
+- 患
+- 禁
+- 閉
+- 迁
+- 错
+- 抵
+- ኤ
+- 葬
+- •
+- 評
+- 땄
+- 벼
+- 哭
+- 漢
+- 坊
+- 适
+- 犹
+- 蜜
+- 宾
+- 够
+- 镜
+- ѓ
+- 棋
+- 趕
+- 減
+- 诗
+- 《
+- 疫
+- 밖
+- 苣
+- 靖
+- Ê
+- ഈ
+- 阪
+- 涉
+- ፡
+- ०
+- 왼
+- ञ
+- 藥
+- 》
+- 寒
+- 怀
+- 랫
+- ٰ
+- 닦
+- 卒
+- 邻
+- 莲
+- Ӹ
+- ഖ
+- 振
+- 歷
+- 凯
+- 杂
+- 牧
+- 禾
+- 椒
+- ഓ
+- 팝
+- 坪
+- 汁
+- 肚
+- 齿
+- 释
+- 鳞
+- ኔ
+- 곰
+- 뚝
+- 띠
+- 씹
+- 잇
+- ృ
+- ሙ
+- 貼
+- 熊
+- 甸
+- 豐
+- 莎
+- ፕ
+- 屈
+- 聊
+- 泥
+- ١
+- ୃ
+- 広
+- 댔
+- 빅
+- 짠
+- Ř
+- ሀ
+- ゴ
+- 顧
+- 麥
+- 坚
+- 欲
+- ሻ
+- Ҙ
+- 濟
+- 訴
+- 担
+- 羊
+- 売
+- 経
+- 障
+- 묘
+- 替
+- ଓ
+- ൺ
+- ሶ
+- ‟
+- 븐
+- 릭
+- 巧
+- 様
+- 累
+- 绍
+- 裏
+- 迹
+- 块
+- 听
+- 跨
+- 盐
+- 茂
+- 叙
+- 勸
+- ৷
+- 岁
+- 묵
+- 칙
+- 耐
+- 训
+- 扶
+- 脱
+- 培
+- 溶
+- 购
+- 徽
+- 겉
+- 敢
+- 欣
+- 顔
+- 买
+- 墨
+- 挥
+- 鳍
+- 획
+- 夺
+- 竟
+- 겪
+- 둬
+- 숭
+- 쫄
+- 뜯
+- 择
+- 샘
+- 愿
+- ൃ
+- 租
+- 舍
+- 尽
+- 疗
+- Ό
+- ፓ
+- 煮
+- 뀐
+- 估
+- 紧
+- ኢ
+- 묶
+- 캡
+- 檢
+- 醬
+- 繼
+- 翰
+- 巡
+- џ
+- 価
+- 녕
+- 殺
+- 涼
+- 溫
+- 伤
+- 阻
+- Θ
+- 꾼
+- 뽕
+- 碗
+- 読
+- 鼓
+- 勞
+- 逆
+- 毎
+- 絕
+- 뷔
+- ኝ
+- 랭
+- 흑
+- Λ
+- 桂
+- 徐
+- 竞
+- 敏
+- ሄ
+- ỗ
+- 땀
+- 寿
+- 扮
+- 穴
+- ଘ
+- 랍
+- 붕
+- 諗
+- 搬
+- 畢
+- 筆
+- ఏ
+- ฏ
+- ぞ
+- づ
+- 贞
+- 昭
+- 崇
+- 径
+- 氧
+- Ĥ
+- 犬
+- 귤
+- 썹
+- 尋
+- 促
+- 尺
+- 証
+- 械
+- 墙
+- 郊
+- ໆ
+- 層
+- 렉
+- 锡
+- 契
+- 験
+- 箱
+- 긋
+- ሕ
+- Ի
+- 芝
+- 絲
+- 宅
+- 阵
+- 逊
+- 춰
+- 쪘
+- 乱
+- 敬
+- 赵
+- 쭈
+- 疾
+- 頓
+- 笔
+- 矿
+- 込
+- 늙
+- 닿
+- 힙
+- 狀
+- 碑
+- 梦
+- 财
+- 辐
+- 렛
+- 촉
+- 팟
+- 潜
+- 伸
+- 陶
+- 箭
+- 措
+- ឋ
+- 础
+- ̀
+- 蚊
+- 偶
+- 헷
+- 蚤
+- 郑
+- 勇
+- 奴
+- 召
+- 税
+- 眠
+- ễ
+- 꽉
+- 냅
+- 펙
+- 荔
+- 輸
+- 叔
+- 覆
+- 宏
+- 歲
+- 浦
+- 焦
+- ቴ
+- ዮ
+- 앱
+- 좁
+- Դ
+- 範
+- 採
+- ォ
+- 闭
+- 絶
+- 稱
+- 炎
+- 藻
+- 욱
+- 剑
+- 慣
+- 漏
+- 帰
+- 裂
+- 衡
+- 涯
+- ፖ
+- 傾
+- ဦ
+- ጃ
+- 衰
+- 쉴
+- 쩐
+- 灭
+- 燕
+- 薯
+- ץ
+- 뇌
+- 辽
+- 孟
+- ጓ
+- 쫙
+- 埗
+- 漸
+- ೃ
+- 秦
+- ጫ
+- 귄
+- 뺄
+- 웅
+- 쨌
+- 캔
+- 嘛
+- 鳥
+- 污
+- 卻
+- 幅
+- 插
+- 筒
+- 灾
+- 忠
+- 疆
+- 戦
+- 릎
+- ጭ
+- ヒ
+- 曬
+- 榮
+- 塊
+- 晶
+- 遍
+- 课
+- 宪
+- 仪
+- 嚴
+- 殊
+- 댕
+- 띄
+- 샐
+- 蔡
+- 废
+- 译
+- 赞
+- 顾
+- 鞋
+- 碧
+- ሥ
+- 뺏
+- 쫓
+- 究
+- 魔
+- 隐
+- 颜
+- 慧
+- 멋
+- ǹ
+- 긁
+- 룹
+- Χ
+- 讨
+- 誉
+- 応
+- 賺
+- 鴨
+- 갓
+- Գ
+- 戶
+- 側
+- 妇
+- 險
+- 듬
+- 靜
+- 苗
+- 邨
+- 織
+- '*'
+- ኩ
+- 맹
+- 썸
+- 쵸
+- 톱
+- 昆
+- 猪
+- 鼻
+- 陸
+- 鏡
+- 饰
+- 娘
+- 榜
+- 磁
+- 却
+- Ά
+- Ӧ
+- ሴ
+- 썩
+- 溝
+- 衝
+- 壞
+- 屏
+- 销
+- ຣ
+- ϊ
+- 덥
+- 爾
+- 傅
+- 篮
+- 货
+- 汇
+- ૌ
+- 폼
+- 廠
+- 缘
+- 曹
+- 獲
+- 仓
+- 颁
+- 雕
+- 臭
+- 審
+- 뱅
+- 沖
+- 亨
+- 触
+- 郭
+- 肥
+- ຟ
+- 굽
+- 쌌
+- 藝
+- 丘
+- 狼
+- 赏
+- ஆ
+- 灘
+- 叉
+- ૉ
+- 靈
+- 껌
+- 흠
+- 绪
+- 샌
+- 剩
+- 蒸
+- 娜
+- 拔
+- ឱ
+- 蹄
+- ઠ
+- ௌ
+- 劇
+- 苑
+- 룩
+- 뺀
+- 额
+- 丸
+- 贯
+- 阴
+- 団
+- ਐ
+- 믹
+- ᱞ
+- ᱨ
+- 纹
+- 콤
+- 慮
+- 獨
+- 骑
+- 勋
+- 抱
+- 猫
+- 枚
+- ៈ
+- 衛
+- '='
+- 茄
+- 겐
+- 껍
+- 慈
+- 淨
+- 訪
+- 详
+- 拟
+- 峡
+- ౌ
+- Į
+- 敗
+- 柯
+- 掂
+- 爬
+- 濕
+- 芳
+- 笼
+- ឈ
+- ẩ
+- 陕
+- 맙
+- 缅
+- ૈ
+- 舒
+- 堪
+- 孤
+- 钢
+- 碘
+- 談
+- 吗
+- 寝
+- 粵
+- ২
+- ሸ
+- 鳳
+- 찢
+- 逝
+- 屬
+- 蒲
+- 邀
+- 卵
+- 丈
+- 蚜
+- ፎ
+- 裔
+- 뜩
+- Џ
+- ቂ
+- 밟
+- 잼
+- 쪄
+- 팍
+- 핵
+- 垂
+- 栖
+- 轴
+- 剂
+- 茜
+- 拆
+- ٬
+- ဓ
+- 啫
+- 擇
+- 煩
+- 脚
+- 舌
+- 붓
+- 쇠
+- 綠
+- ఖ
+- 侍
+- 麗
+- 勃
+- 寻
+- 卖
+- ဥ
+- 멈
+- 緒
+- 裕
+- 쥐
+- 쫍
+- Ո
+- 坏
+- 屎
+- 虾
+- 刑
+- 刷
+- 즘
+- 錦
+- 宙
+- ଞ
+- 맵
+- 쌈
+- ೌ
+- 엊
+- 喔
+- 婦
+- 昔
+- 針
+- 签
+- 诉
+- 刚
+- 晴
+- 妙
+- ៗ
+- ᱜ
+- Φ
+- 値
+- 숫
+- 윗
+- 擊
+- 饭
+- 匹
+- 拖
+- 監
+- −
+- 閒
+- 넉
+- 넨
+- Պ
+- 腸
+- 启
+- 薪
+- 邮
+- 伐
+- ふ
+- 倍
+- ၉
+- 勉
+- 돋
+- Ι
+- ጽ
+- 寧
+- 柔
+- 懷
+- 軟
+- 潘
+- 凡
+- 玻
+- 陷
+- 転
+- 購
+- 쾌
+- 鹽
+- 乗
+- 耀
+- ڍ
+- 吖
+- 뤄
+- 玄
+- 飽
+- 迈
+- 肝
+- 蛾
+- 撒
+- 硕
+- 堇
+- 沈
+- 胸
+- 陳
+- 낯
+- 듀
+- 臉
+- 押
+- 棄
+- 凍
+- 访
+- 册
+- ኦ
+- 療
+- 챔
+- 允
+- 纸
+- ኋ
+- 풋
+- ٢
+- 胆
+- 貸
+- 뻥
+- 쭐
+- 伞
+- 壓
+- 伏
+- 貌
+- 颗
+- 烏
+- 쩨
+- 岐
+- 麟
+- 尿
+- 扒
+- 訓
+- 郁
+- 췄
+- ਠ
+- 뭉
+- 쩡
+- 펌
+- 悲
+- 祝
+- 壳
+- 岭
+- 贸
+- 伟
+- 夷
+- 锋
+- 祭
+- १
+- ឬ
+- 旦
+- 윈
+- Ñ
+- 뻤
+- 쉐
+- 텍
+- 혐
+- 訂
+- 腩
+- 藍
+- 倫
+- 妈
+- 惡
+- 霸
+- 孝
+- 减
+- ¥
+- 歸
+- 額
+- 윙
+- 찝
+- 텀
+- 宮
+- 遺
+- 멤
+- 煲
+- 潭
+- 烧
+- 煙
+- 몇
+- 헬
+- 扁
+- 줌
+- 览
+- 滋
+- 怒
+- 搜
+- 碎
+- 複
+- 己
+- Տ
+- ‽
+- 仆
+- 닉
+- 봇
+- 앗
+- Ġ
+- 窗
+- 滕
+- 猛
+- 综
+- 珊
+- 误
+- Ɣ
+- 柄
+- 젝
+- ၄
+- 鍾
+- 騙
+- 닷
+- 략
+- 엎
+- ૅ
+- 虱
+- 銷
+- 黨
+- 添
+- 廉
+- 辆
+- È
+- 穗
+- 탱
+- ਓ
+- 縣
+- 宽
+- 坎
+- 亥
+- 币
+- 鬆
+- Ô
+- ẻ
+- 賽
+- 궈
+- 볍
+- 딪
+- ഘ
+- 攪
+- 瓶
+- 炭
+- 阁
+- 扩
+- 串
+- 铜
+- Ӓ
+- ޏ
+- 梨
+- Å
+- ၅
+- 讷
+- 룰
+- 割
+- 築
+- 翁
+- 啟
+- 遵
+- 刊
+- 璃
+- 翔
+- 驱
+- 伍
+- 幻
+- ヴ
+- ৪
+- 辣
+- ヘ
+- 圓
+- 梗
+- 댁
+- 셉
+- 셜
+- 쌩
+- 帽
+- 怖
+- 顯
+- 俗
+- 夕
+- 键
+- Ụ
+- 孫
+- 꼽
+- 빴
+- 奔
+- 瓣
+- 藩
+- 拒
+- 恶
+- Ԥ
+- ၈
+- ኘ
+- 뱃
+- 칵
+- ጡ
+- ፒ
+- 偷
+- 踏
+- 桐
+- 憶
+- 靚
+- 嫌
+- 疏
+- 隧
+- 厦
+- 茅
+- 貝
+- 寨
+- 绕
+- ÿ
+- ̃
+- 捨
+- 썬
+- 쿼
+- 툭
+- 푼
+- ൗ
+- 渐
+- 鳴
+- 퉁
+- 脫
+- 癌
+- 腿
+- 愉
+- 哺
+- 檸
+- 拓
+- ぼ
+- 荆
+- 狸
+- 뿔
+- Þ
+- 哂
+- 浩
+- 爪
+- 辅
+- 岳
+- 濃
+- 荒
+- 陪
+- 姨
+- 鬧
+- 绘
+- 廟
+- 禄
+- 빔
+- 빽
+- 坛
+- 즉
+- 詞
+- 恭
+- 彈
+- 审
+- 催
+- ぱ
+- ဤ
+- 蝽
+- 뺑
+- ỹ
+- 謂
+- 둑
+- 壮
+- 鸡
+- 魂
+- 純
+- 乃
+- 両
+- 樱
+- औ
+- Վ
+- 彻
+- 粟
+- 갱
+- 쎈
+- 탠
+- 蓬
+- 獎
+- 袖
+- ጦ
+- ৯
+- 廢
+- 껏
+- 쥬
+- ґ
+- 絡
+- ಏ
+- ᱴ
+- 輪
+- 姊
+- 衫
+- 填
+- 恢
+- 燃
+- 辞
+- 斜
+- 円
+- 恋
+- 塑
+- 俺
+- 欠
+- 鞍
+- Ħ
+- ౦
+- 掛
+- 깡
+- 탓
+- 谓
+- 龟
+- 屿
+- 谱
+- 瞬
+- 缩
+- 掃
+- 烟
+- 忆
+- 賀
+- 肺
+- 輩
+- 꼰
+- 헛
+- 盲
+- 蟲
+- 샷
+- ޠ
+- Ρ
+- 寂
+- 陰
+- 弁
+- 賢
+- 弘
+- 扬
+- 豹
+- 腺
+- ỳ
+- €
+- 엥
+- 撤
+- 箕
+- 仰
+- 従
+- 釋
+- 摇
+- 毁
+- 큼
+- ឃ
+- 秒
+- 蒿
+- ጂ
+- 嚇
+- 맻
+- Ք
+- 攞
+- 껀
+- 뜰
+- Օ
+- 勵
+- 蕾
+- 戻
+- 伝
+- 鰂
+- 窩
+- 惜
+- ᱥ
+- 迅
+- 빗
+- 흉
+- Թ
+- 订
+- 拳
+- 衔
+- 鼎
+- 猎
+- 蘿
+- 橙
+- 辰
+- 砂
+- 삿
+- ٩
+- 뉘
+- 햇
+- 煎
+- 碟
+- 損
+- 披
+- 衍
+- 爛
+- 萧
+- 札
+- І
+- ়
+- ឌ
+- 饮
+- 拼
+- 遣
+- Լ
+- 膜
+- 撐
+- 姿
+- 吻
+- 棠
+- 沟
+- 詢
+- 阜
+- 밭
+- 傻
+- 截
+- 曉
+- 爽
+- 쩜
+- 핏
+- 擁
+- 稻
+- 忽
+- 磨
+- 吾
+- 葱
+- 株
+- 蜂
+- 庞
+- ቪ
+- ዞ
+- 虹
+- ৫
+- ỷ
+- 맣
+- 賞
+- 凝
+- 肩
+- 歧
+- Ɛ
+- 帆
+- ԑ
+- ၆
+- 앵
+- ۰
+- ၃
+- 뀔
+- 贏
+- 驾
+- 慶
+- 抓
+- 嫁
+- ஒ
+- 脷
+- 廈
+- 悄
+- 増
+- 葛
+- Ҷ
+- ቷ
+- 뱀
+- ′
+- 悠
+- 莉
+- 筲
+- 暂
+- 丞
+- 陀
+- 掘
+- ぜ
+- ኙ
+- ⠀
+- 摸
+- 굿
+- 퀄
+- 훅
+- 匯
+- 崗
+- 妥
+- 祠
+- 満
+- 蛙
+- 豉
+- 頻
+- 芸
+- 湿
+- 誕
+- ৩
+- 懂
+- 隱
+- 卑
+- 郵
+- 編
+- 肃
+- 劣
+- 档
+- 瘤
+- 盆
+- 冒
+- 弃
+- 彭
+- 稅
+- 깥
+- ፅ
+- 넥
+- 쇄
+- 쏟
+- 푹
+- 籌
+- 亭
+- 芹
+- 邓
+- 薇
+- ୌ
+- 壽
+- 蜱
+- 凌
+- 膽
+- 륵
+- 扇
+- 멸
+- 歐
+- ಘ
+- 喊
+- 赢
+- 雜
+- 珀
+- 礎
+- 撚
+- 恰
+- 藉
+- 驶
+- 못
+- 땜
+- 桌
+- 蒜
+- ሠ
+- 섹
+- 図
+- 駛
+- 誠
+- 圭
+- 鍋
+- 勾
+- 纯
+- 谋
+- ւ
+- 茛
+- ቸ
+- 贺
+- ፉ
+- 猴
+- 眉
+- 퀘
+- Չ
+- 吊
+- 鱲
+- 瞓
+- 囊
+- 腰
+- 矶
+- 赶
+- 蚬
+- 妃
+- 揮
+- 융
+- 틈
+- ኃ
+- 损
+- 뇨
+- 빚
+- 찡
+- 侄
+- 仿
+- 浅
+- 邵
+- 券
+- 忧
+- 润
+- 雍
+- 汤
+- 遮
+- 丧
+- ヨ
+- 髮
+- 냄
+- 횟
+- Ő
+- ጆ
+- 淘
+- 귈
+- 므
+- Ֆ
+- 櫃
+- 琳
+- 虚
+- 织
+- 阮
+- 循
+- 辨
+- 姻
+- 您
+- ຝ
+- 棕
+- ኬ
+- ‚
+- 崩
+- 暑
+- 飾
+- 깼
+- 칩
+- 펑
+- Ռ
+- ᱵ
+- 슐
+- 寞
+- 巷
+- 圳
+- 谈
+- 抜
+- 杭
+- 筋
+- 摘
+- 挪
+- 狄
+- 跃
+- 鹰
+- 汝
+- 맞
+- 般
+- 茎
+- Ə
+- ਊ
+- 咋
+- 泣
+- 丟
+- 翅
+- 蔥
+- 駅
+- 脏
+- 矩
+- 闽
+- 疲
+- 牡
+- 韓
+- 槭
+- 嗱
+- ڼ
+- 暫
+- 淚
+- 蓋
+- 뎅
+- 뢰
+- 찼
+- 績
+- 誤
+- 瑚
+- 魯
+- 旨
+- 蛤
+- 쩔
+- 競
+- 紹
+- ৬
+- 狂
+- 냠
+- 쏠
+- 潛
+- 栗
+- 遲
+- 纵
+- 棉
+- 隋
+- 囲
+- 丛
+- Ō
+- 喀
+- 曜
+- 棘
+- 냬
+- 奧
+- 脆
+- 逛
+- 젖
+- 録
+- 悬
+- 魷
+- 晨
+- 狮
+- 桔
+- 胎
+- 脊
+- 倉
+- 页
+- ໌
+- ፃ
+- ỡ
+- 帳
+- 팽
+- 嚕
+- ẳ
+- 咕
+- 烤
+- 抚
+- 椅
+- 抑
+- 粹
+- ឿ
+- 擦
+- 焼
+- 夾
+- 脂
+- ৮
+- ኞ
+- ዴ
+- 哩
+- 擺
+- 睛
+- 憲
+- 怡
+- 滩
+- 蕃
+- 矮
+- 卜
+- 惟
+- 賓
+- 稳
+- 敌
+- 蔷
+- 枢
+- Ζ
+- 숲
+- Ї
+- ઊ
+- 굔
+- 띵
+- 뭣
+- 섰
+- 옴
+- 펼
+- 擅
+- 稀
+- 兽
+- 慎
+- 浓
+- 氟
+- 娱
+- 簽
+- 芋
+- 貓
+- ゼ
+- 乖
+- 揀
+- 젓
+- 閃
+- 룬
+- 쉰
+- 処
+- 姜
+- 凱
+- 吐
+- 災
+- 逸
+- 頃
+- 樟
+- 籠
+- 覽
+- 販
+- 携
+- 卿
+- 艘
+- 滚
+- 奮
+- 欖
+- 遷
+- ޟ
+- 楊
+- 淳
+- Ẹ
+- 훔
+- 卓
+- 詐
+- 顆
+- 漁
+- 劫
+- 廊
+- 녔
+- 储
+- 廿
+- 굶
+- 喫
+- 뱉
+- 燥
+- 狭
+- 浴
+- 碳
+- 渠
+- 铃
+- 斐
+- 獅
+- 琉
+- 胶
+- 乏
+- 叛
+- 뭐
+- 冕
+- 帯
+- ଢ
+- ਢ
+- ၌
+- 咬
+- 惑
+- 貿
+- 赴
+- 괌
+- 멕
+- 츄
+- 횡
+- 稚
+- 蠻
+- 励
+- 捐
+- 祈
+- 遂
+- 杆
+- ங
+- 荨
+- 拠
+- 滅
+- ᱷ
+- ኳ
+- 汰
+- 탭
+- 팁
+- 밸
+- 얄
+- 汀
+- 褐
+- 晩
+- 虛
+- 诚
+- 醉
+- 辉
+- 毫
+- 虑
+- 蒋
+- 稿
+- 冯
+- 検
+- 靴
+- 瑙
+- 蝠
+- 贾
+- Շ
+- 寓
+- 糟
+- 뻘
+- 훌
+- 鎮
+- 幣
+- 违
+- 阔
+- 稍
+- 牵
+- 効
+- 鉴
+- 逻
+- 戟
+- 揚
+- 腔
+- Ã
+- 繳
+- ĺ
+- ٨
+- 폈
+- 콕
+- 聰
+- 丑
+- 岗
+- 舟
+- 愈
+- 狱
+- 罕
+- 绩
+- 笠
+- 甥
+- 収
+- 億
+- 傲
+- 壶
+- 蕉
+- Ӑ
+- ‐
+- ٤
+- ፤
+- 렬
+- 및
+- 옹
+- 칫
+- 哀
+- 縮
+- 碌
+- 鎖
+- 鹹
+- 夥
+- 巫
+- 灌
+- 弥
+- 沼
+- 锁
+- Ā
+- ʿ
+- ઑ
+- 緣
+- 랗
+- 숏
+- 肌
+- 啤
+- 頼
+- 慰
+- 虽
+- 崖
+- 溜
+- 拱
+- 諸
+- 氯
+- ユ
+- 询
+- 倾
+- 喉
+- ៌
+- २
+- 깅
+- 꽈
+- 쑥
+- 옵
+- 잎
+- 袁
+- 澄
+- 酋
+- 募
+- 挂
+- 痕
+- 纤
+- 屠
+- 践
+- 榕
+- 逼
+- 辭
+- 膠
+- 蛱
+- ۂ
+- ઢ
+- 輝
+- 锥
+- 됩
+- 蘇
+- 蝴
+- 鹃
+- Է
+- 娃
+- 嶼
+- 淋
+- 夸
+- 蓮
+- 蠢
+- 帐
+- 醇
+- 龄
+- 壬
+- 貫
+- 漂
+- 菇
+- 黐
+- ⁄
+- 梵
+- 돔
+- 킥
+- 텅
+- ଵ
+- 狐
+- 戀
+- 託
+- 獻
+- 寸
+- 递
+- 丙
+- 儀
+- 龜
+- 萼
+- 凸
+- Ή
+- ޘ
+- ఓ
+- 奪
+- 皿
+- 넛
+- 똘
+- 룡
+- 쁨
+- 왤
+- 짰
+- 툴
+- ፀ
+- 佔
+- 昏
+- 쿵
+- ഠ
+- 甫
+- 惹
+- 蘋
+- 撰
+- 撥
+- 磷
+- 纜
+- 尉
+- 鹅
+- 牆
+- 遥
+- 옛
+- Ø
+- Ý
+- ՛
+- 餓
+- 鸿
+- 첼
+- 퍽
+- 훠
+- 雑
+- 浸
+- 狠
+- 赐
+- 鋪
+- 扣
+- 诞
+- 乞
+- 赚
+- 禧
+- 堵
+- ሂ
+- 堤
+- 昼
+- 鵝
+- 겄
+- 맑
+- 쓱
+- 얹
+- 옳
+- 첩
+- 匈
+- 秩
+- 夹
+- 鹤
+- 甩
+- 遅
+- 邪
+- 扑
+- 闪
+- 壤
+- 慌
+- 綜
+- 隙
+- 綫
+- 縄
+- 佑
+- 觸
+- ঈ
+- 攰
+- 鹦
+- 욘
+- 軒
+- 阅
+- 깄
+- 꺾
+- 벚
+- 틸
+- 剪
+- 幽
+- 彦
+- 歇
+- 氢
+- 浜
+- 徵
+- 悉
+- 棚
+- 黒
+- 酷
+- 掠
+- 弓
+- 鞭
+- ឯ
+- 兔
+- 洁
+- 漆
+- 餘
+- 갸
+- 녜
+- 맺
+- 텝
+- 퓨
+- 兆
+- 懶
+- 誌
+- 喙
+- 殼
+- 螢
+- 諾
+- 鞘
+- 徙
+- 鄰
+- 盒
+- 纺
+- 荐
+- 赠
+- 垒
+- 婴
+- 胺
+- ឧ
+- ሏ
+- 嘞
+- 怨
+- 颠
+- 뭇
+- 펀
+- 筑
+- 芽
+- 塗
+- 寮
+- 伪
+- 墩
+- 菠
+- 宴
+- 慕
+- 链
+- 敷
+- 盈
+- 戚
+- 柚
+- 輔
+- ٥
+- ၇
+- ሺ
+- 廚
+- 糊
+- 蔓
+- 겟
+- 뛸
+- 벳
+- 숱
+- 찹
+- 媛
+- 淺
+- 漠
+- 玫
+- 皱
+- 蓉
+- 賭
+- 鳗
+- 卯
+- 坂
+- 慘
+- 穩
+- 瑰
+- 奎
+- 氮
+- 褲
+- 叠
+- 锤
+- '['
+- 爲
+- 恆
+- ̄
+- Զ
+- ९
+- 喂
+- 渣
+- 萤
+- ጸ
+- 搶
+- 蔗
+- 薩
+- 鑊
+- 铺
+- 늬
+- 솜
+- 爺
+- 贊
+- 頸
+- 宛
+- 硫
+- 贤
+- 耕
+- 贼
+- 嘈
+- 騎
+- 凭
+- 峽
+- 惯
+- 芥
+- ٣
+- 厭
+- 徳
+- 撃
+- 椰
+- 檔
+- 곽
+- 렙
+- 쭤
+- 펐
+- 艳
+- 跡
+- 骗
+- 戸
+- 縱
+- 姚
+- 邸
+- 幹
+- 悼
+- 誓
+- 伽
+- 않
+- ຼ
+- 巾
+- 厘
+- 噏
+- 얌
+- Ċ
+- Ɓ
+- ఐ
+- 仗
+- 営
+- 妖
+- 鳅
+- 벙
+- 숍
+- 얜
+- 욜
+- 읍
+- Ҩ
+- 滷
+- 繞
+- 闲
+- 掩
+- 逢
+- 饼
+- 扭
+- 攝
+- 栋
+- 茹
+- 済
+- 雾
+- 沐
+- Խ
+- ሐ
+- ቫ
+- 揸
+- 淮
+- 猜
+- 禅
+- 胃
+- 껑
+- 됨
+- 팸
+- 肖
+- 辺
+- 씌
+- 嬤
+- 屆
+- 旭
+- 铭
+- 僅
+- 聘
+- 俊
+- 鸣
+- 绳
+- 纷
+- 胁
+- 躍
+- 绣
+- 殘
+- 欄
+- 澤
+- 削
+- 阎
+- 찮
+- Ң
+- ٦
+- 徹
+- 棟
+- 裸
+- 깠
+- 늫
+- 뻗
+- 셈
+- 柿
+- 煤
+- 矢
+- 庸
+- 酰
+- 滴
+- 詳
+- 噚
+- 芯
+- 魁
+- 弯
+- 崔
+- 圧
+- 犀
+- ̂
+- \
+- Ь
+- ፌ
+- 礁
+- 蘆
+- 钦
+- 괄
+- 렷
+- 샴
+- 셰
+- 抹
+- 牽
+- 妮
+- 嘗
+- 宰
+- 廖
+- 莊
+- 俘
+- 凉
+- 迦
+- 仇
+- 缓
+- 걱
+- 査
+- 儲
+- Ҭ
+- Ջ
+- 労
+- 毀
+- 糙
+- ৭
+- ಐ
+- ឲ
+- 冚
+- 厲
+- 疼
+- 偉
+- 儒
+- 牲
+- 讚
+- 呃
+- 碱
+- 萝
+- 逗
+- 彰
+- 砖
+- 肿
+- 訳
+- 僧
+- 咨
+- 挙
+- 溯
+- 瀑
+- 碍
+- 尝
+- 汪
+- 檬
+- 赌
+- 菁
+- ५
+- 抄
+- 斤
+- 襯
+- 닙
+- 쯔
+- 퀸
+- 텨
+- 蜗
+- 貪
+- 귓
+- 엮
+- ಞ
+- 憎
+- 滇
+- 酥
+- 娶
+- 孕
+- 捉
+- 晓
+- 樽
+- 蜘
+- 閱
+- 雇
+- 筹
+- 癸
+- 酶
+- 儘
+- 睦
+- 膨
+- Œ
+- 勁
+- 恨
+- 拎
+- 苞
+- 貧
+- 隣
+- 갇
+- 넜
+- 댈
+- 릏
+- ઞ
+- 凹
+- 歉
+- 謠
+- 躲
+- ぽ
+- 嗣
+- 滾
+- 汐
+- 喱
+- 贴
+- 権
+- 烷
+- 児
+- 菩
+- 辈
+- 騰
+- ୈ
+- 肋
+- 킷
+- ఘ
+- ᱡ
+- 勘
+- 叻
+- 鍊
+- 퐁
+- ヌ
+- 媳
+- 稣
+- 鲍
+- 僱
+- 榴
+- 芙
+- 悟
+- 拘
+- 瓷
+- 仑
+- 霖
+- 饱
+- 棺
+- 滤
+- 赋
+- 渊
+- 央
+- 蕊
+- 蹟
+- 搖
+- 秉
+- 껐
+- Ë
+- 戒
+- 揭
+- 瞭
+- 趨
+- 軽
+- 깍
+- 륙
+- 뻑
+- 쌔
+- 콧
+- 픔
+- 蔭
+- 迴
+- 帅
+- 硝
+- 舱
+- 踢
+- 嘆
+- 忌
+- 渔
+- 伙
+- 盗
+- 蹈
+- 锐
+- 颇
+- 窮
+- 醜
+- 祯
+- 腕
+- 尹
+- £
+- 凰
+- 喪
+- 嬲
+- 潔
+- '~'
+- ঋ
+- 債
+- 芭
+- 薹
+- 짚
+- 펭
+- 毅
+- 汶
+- 糧
+- 蜡
+- 敵
+- 枫
+- 橡
+- 凶
+- 呆
+- 碰
+- 薛
+- 涵
+- 姬
+- 柬
+- 贫
+- 撮
+- 豊
+- 庶
+- 纬
+- Ș
+- ሯ
+- ᱣ
+- 嬰
+- 孢
+- 潤
+- 酵
+- 閣
+- 굵
+- 덧
+- 짖
+- 쿡
+- 燭
+- 驼
+- 冊
+- 厥
+- 愁
+- 楠
+- 濑
+- 診
+- 谁
+- 铸
+- 頁
+- 岑
+- 晃
+- 瘦
+- 肾
+- 噶
+- 窑
+- 戌
+- 摆
+- 窓
+- 坟
+- 涂
+- 炉
+- 壊
+- 檗
+- 毋
+- 餵
+- 壇
+- 専
+- 脾
+- 霧
+- 鬚
+- ŕ
+- ऋ
+- 戊
+- 脈
+- Ҵ
+- ۱
+- ጤ
+- 廁
+- 烯
+- 蓼
+- 럿
+- 샹
+- 싯
+- 쬐
+- 펫
+- 헝
+- 흙
+- 墅
+- 賴
+- 喷
+- 榄
+- 佈
+- 咒
+- 扰
+- 豫
+- 罂
+- 硤
+- 継
+- 莓
+- 餃
+- ゥ
+- 宠
+- 罐
+- ಓ
+- ಠ
+- ዥ
+- 弦
+- 菓
+- 봅
+- 왓
+- 톨
+- Փ
+- 滘
+- 瑜
+- 紐
+- 綿
+- 逮
+- 酮
+- 熄
+- 笛
+- 繫
+- 繰
+- 氛
+- 瀝
+- 陨
+- 歳
+- 殷
+- 奠
+- 奸
+- 酬
+- 杖
+- 钠
+- 浆
+- 狩
+- 좋
+- 険
+- 擬
+- Υ
+- 椎
+- ፐ
+- 瘟
+- 렐
+- 쭝
+- ڦ
+- 旱
+- 衆
+- 삽
+- 摧
+- 罩
+- 鬥
+- 徴
+- 抬
+- 挖
+- 噴
+- 揺
+- 梭
+- 棍
+- 腾
+- 谊
+- 鲤
+- 侠
+- 枕
+- 耗
+- 많
+- 斛
+- 詩
+- 蔔
+- 亿
+- 椭
+- 盔
+- 胀
+- ८
+- ഐ
+- ဧ
+- 惊
+- 拾
+- 肢
+- 谥
+- 줍
+- 剥
+- 愚
+- 槍
+- 縫
+- 绵
+- 迟
+- 鋼
+- 颖
+- 눕
+- 딲
+- 멧
+- 숟
+- 쏴
+- 癮
+- 篤
+- 鳃
+- 鹉
+- 컹
+- 厳
+- 獄
+- 扫
+- 括
+- 槟
+- 焗
+- 賊
+- 钾
+- 坝
+- 払
+- 獵
+- 罹
+- 墟
+- 绥
+- 雌
+- 沫
+- 魅
+- 梳
+- 抛
+- 榆
+- 潟
+- 脸
+- 悔
+- 杏
+- 랴
+- Æ
+- Ƙ
+- ۖ
+- ३
+- 傍
+- 惧
+- 譯
+- 鄧
+- 霉
+- 鳄
+- 껜
+- 꿍
+- 둡
+- 륜
+- 뮬
+- 슝
+- 슥
+- 윽
+- 줏
+- 짼
+- 嫩
+- 栄
+- 炆
+- 庚
+- 颓
+- 巢
+- 幌
+- 荠
+- 撇
+- 苯
+- 暨
+- 肇
+- 鍵
+- 颂
+- 寡
+- ஏ
+- ᱼ
+- 琼
+- 寬
+- 挽
+- 襲
+- 坤
+- 橘
+- 洒
+- 贈
+- 钝
+- ဌ
+- 虐
+- Û
+- ᱦ
+- ᱯ
+- 惰
+- 擴
+- 昇
+- 蒐
+- 蠔
+- 겜
+- 뛴
+- 왁
+- 짭
+- 펄
+- ጻ
+- 壯
+- 汕
+- 贪
+- 軌
+- 闸
+- 駆
+- 栽
+- 氨
+- 硅
+- 瑪
+- 栏
+- 豁
+- 槐
+- ଝ
+- Ұ
+- ६
+- 哇
+- 憂
+- 绒
+- 깬
+- 돗
+- 뵈
+- 삘
+- 쏙
+- 얗
+- 웰
+- 즙
+- 캉
+- 펠
+- ዌ
+- 弊
+- 併
+- 噉
+- 冶
+- 綁
+- 刁
+- 磐
+- 鳩
+- 蜥
+- 卸
+- 歪
+- 芦
+- '}'
+- Ũ
+- Ŭ
+- ǔ
+- ᱰ
+- 鮑
+- 륨
+- 쩰
+- ѐ
+- ٧
+- ᱹ
+- 勿
+- 孖
+- 悅
+- 扉
+- 朴
+- 榧
+- 橄
+- 竇
+- 苹
+- 낑
+- 늑
+- 돕
+- 뒹
+- 붐
+- 뿜
+- 엣
+- 첸
+- 팥
+- 乸
+- 馨
+- 麒
+- 妄
+- 趾
+- 毗
+- 衞
+- 沪
+- 勳
+- 膝
+- 辩
+- 奋
+- 拡
+- '@'
+- 塩
+- 椿
+- 铠
+- 傑
+- 壺
+- 霞
+- ቬ
+- 剖
+- 懐
+- 柑
+- 渴
+- ጁ
+- 厨
+- 곈
+- 왈
+- ଃ
+- 妾
+- 棱
+- 琦
+- 聪
+- 蚓
+- 锅
+- 鸦
+- 擾
+- 斥
+- 梓
+- 瘋
+- 绶
+- 荊
+- 荡
+- 霜
+- 恕
+- 抖
+- 殉
+- ～
+- 峭
+- 涛
+- 陌
+- 罢
+- 迄
+- 憧
+- 抢
+- 绑
+- 缴
+- 蝇
+- 秃
+- 떤
+- 鮭
+- 匙
+- 垣
+- ぷ
+- 恥
+- 钩
+- Ū
+- 庇
+- 痴
+- 눅
+- 뤘
+- 륭
+- 뾰
+- 샜
+- 쌰
+- 잣
+- Ξ
+- Ӗ
+- 茸
+- 髪
+- 庵
+- 礙
+- 荫
+- 侦
+- 贷
+- 颱
+- 憾
+- 択
+- 眺
+- 窄
+- 薬
+- 鑽
+- 詹
+- 趋
+- 녁
+- Ը
+- 攀
+- 攸
+- 菱
+- 賤
+- Ύ
+- Ω
+- ۀ
+- Ị
+- ‒
+- 咳
+- 埠
+- 宵
+- 뜸
+- 렘
+- 롬
+- 젯
+- 쭘
+- 킵
+- ঔ
+- 膳
+- 鴻
+- 扯
+- 沽
+- 羌
+- 踩
+- 猶
+- 帜
+- 畜
+- 驳
+- 嗲
+- 尻
+- 蝉
+- 誘
+- 粮
+- 阀
+- 飓
+- 窟
+- 総
+- 劑
+- 縁
+- 嗌
+- 湧
+- 飼
+- 慳
+- 鈴
+- 駐
+- '#'
+- Ɗ
+- Ҽ
+- Ճ
+- 吏
+- 屌
+- 枱
+- 爐
+- 謀
+- 鹏
+- 뚤
+- 텃
+- Յ
+- 唷
+- 喳
+- 杞
+- 爹
+- 謹
+- ഔ
+- 兀
+- 饶
+- 颈
+- 庐
+- 聴
+- 挡
+- 淆
+- 轰
+- 禽
+- 诱
+- 炼
+- ဿ
+- 脅
+- 艦
+- 薑
+- 槲
+- ቨ
+- 갯
+- 꾹
+- 붉
+- 캄
+- 쾅
+- Ĕ
+- Ī
+- Ҿ
+- ሟ
+- 峻
+- 諮
+- 贬
+- 鉄
+- 쟀
+- 釣
+- 懼
+- 撻
+- 牟
+- 肠
+- 昧
+- 鲈
+- 侨
+- 湘
+- 铎
+- 厢
+- 履
+- 债
+- 悦
+- 粤
+- ឥ
+- 沸
+- 曝
+- 净
+- 厉
+- 搏
+- 浏
+- 淇
+- 獐
+- 禍
+- ఛ
+- ቧ
+- 弾
+- 欺
+- 蛄
+- 씁
+- 캘
+- ४
+- ሹ
+- ቭ
+- 呱
+- 啥
+- 嚿
+- 嬸
+- 洽
+- 蟻
+- 诵
+- 豚
+- 颅
+- 駕
+- 깁
+- 뻣
+- 핍
+- 倡
+- 擎
+- 檯
+- 唤
+- 惱
+- 憑
+- 罵
+- 譽
+- 悶
+- 涩
+- 邑
+- 襄
+- 翌
+- 犁
+- 踪
+- 酱
+- 帘
+- 淀
+- 侶
+- 渭
+- 蹤
+- 黙
+- ሾ
+- ጪ
+- 撈
+- 畅
+- 盧
+- 齢
+- 솥
+- 쨍
+- 쫀
+- 튕
+- Ļ
+- Њ
+- 眨
+- 籽
+- 肽
+- 舖
+- 钙
+- 벡
+- 쫑
+- 큘
+- 핬
+- 헨
+- 壹
+- 濾
+- 蔵
+- 冤
+- 啖
+- 繩
+- 衙
+- 詰
+- '{'
+- 挺
+- 朵
+- 坞
+- 醋
+- 尘
+- 鄭
+- 挝
+- 陡
+- 劍
+- 淄
+- 缆
+- 罰
+- 曳
+- 隠
+- Ć
+- Ӡ
+- ७
+- ᱲ
+- 噌
+- 揼
+- 矣
+- 룻
+- 앰
+- 횐
+- 兇
+- 燉
+- 錶
+- 攬
+- 笏
+- 肤
+- 剤
+- 嫡
+- 屡
+- 缀
+- 囚
+- 懿
+- 镁
+- 祀
+- 趁
+- 鸭
+- 叩
+- 讼
+- 颌
+- 叮
+- 髓
+- 艰
+- 贩
+- 鴛
+- 鴦
+- 框
+- 邱
+- 袂
+- 拯
+- 濤
+- 萍
+- 韭
+- 妨
+- 薈
+- ዬ
+- 壩
+- 澎
+- 瞧
+- 繪
+- 羞
+- 胖
+- 谭
+- 鸢
+- 겔
+- 랠
+- 맸
+- 몫
+- 뵙
+- 삥
+- 쏭
+- 앚
+- 읏
+- 팜
+- ۲
+- ਔ
+- 剿
+- 卧
+- 峯
+- 彎
+- 臂
+- 蓄
+- 跤
+- 驅
+- 뎌
+- ―
+- 咏
+- 屑
+- 拋
+- 熔
+- 萊
+- 遙
+- 騒
+- 俱
+- 嗜
+- 緑
+- 缠
+- 螃
+- 谨
+- 喘
+- 趙
+- 霹
+- 哦
+- 莺
+- 邢
+- 憬
+- 氷
+- 酯
+- 滞
+- 黔
+- 噬
+- 俞
+- 濱
+- ޛ
+- ፔ
+- 暮
+- 밋
+- 뽈
+- Ў
+- ጄ
+- Ấ
+- 侣
+- 唸
+- 桦
+- 泌
+- 稔
+- 莹
+- 钻
+- 갛
+- 꿋
+- 슛
+- 쑤
+- 얕
+- 엌
+- Ҟ
+- ־
+- 傘
+- 呵
+- 妝
+- 嵩
+- 慾
+- 暈
+- 沧
+- 禹
+- 蝕
+- 흩
+- 嘩
+- 盞
+- 笨
+- 魄
+- 償
+- 垩
+- 駁
+- 嘎
+- 枯
+- 茫
+- 谦
+- 匿
+- 跋
+- 镰
+- ‧
+- 氰
+- 臘
+- 輯
+- ぴ
+- 馳
+- Ă
+- ͘
+- ঊ
+- ቄ
+- 瑶
+- 盼
+- 薦
+- 酿
+- ಃ
+- ១
+- 冀
+- 茱
+- 阱
+- 랖
+- 섀
+- 얽
+- 엿
+- 윌
+- 잌
+- 찻
+- 챠
+- Ґ
+- ẵ
+- 婿
+- 擠
+- 曆
+- 鷹
+- 沛
+- 濫
+- 祿
+- 蜀
+- 釐
+- 闊
+- 沮
+- 濯
+- 拨
+- 訝
+- 柒
+- 珂
+- 蚁
+- 诊
+- 劈
+- 腓
+- 鄂
+- 劲
+- 拂
+- 砵
+- 溃
+- 摔
+- 肆
+- 蛳
+- ဍ
+- 呂
+- 彙
+- 樑
+- 飘
+- 鴉
+- ‹
+- 暇
+- 芩
+- 鰻
+- 齡
+- Ĩ
+- ΐ
+- ѝ
+- ۓ
+- ቺ
+- ›
+- 殃
+- 譜
+- 곁
+- 깰
+- 뗄
+- 잭
+- 촥
+- 캣
+- 펍
+- 
+- ០
+- 嘲
+- 憐
+- 戇
+- 掙
+- ฌ
+- 悩
+- 淒
+- 聂
+- 剌
+- 匝
+- 辯
+- 謊
+- 鉛
+- 鲷
+- 娄
+- 罚
+- 苋
+- 衷
+- 膀
+- 裴
+- 赣
+- 縦
+- 僚
+- 卦
+- 坳
+- 鱿
+- 绎
+- 铅
+- 窝
+- 妓
+- 岔
+- 乒
+- 扱
+- 溢
+- 讽
+- 膚
+- 傣
+- 懒
+- 猿
+- 輋
+- 丫
+- 穎
+- 笪
+- 蓟
+- 倚
+- 匠
+- 匪
+- 砍
+- 莆
+- 莞
+- 诏
+- ː
+- Ժ
+- ሣ
+- 螞
+- 釀
+- 갭
+- 눴
+- 닳
+- 띡
+- 맴
+- 숯
+- 쐈
+- 씽
+- 킴
+- ዣ
+- 奕
+- 砌
+- 諒
+- 錫
+- 雰
+- 喻
+- 岷
+- 灑
+- 炳
+- 爍
+- 絵
+- 炊
+- 礦
+- 跆
+- 輿
+- 螯
+- 朔
+- 蔽
+- 唆
+- 塌
+- 馴
+- 纠
+- 螂
+- 铝
+- 懸
+- ゾ
+- 猕
+- 稽
+- 姥
+- 吟
+- 廓
+- 肪
+- 攤
+- ፏ
+- 哄
+- 巻
+- 猬
+- 藐
+- 闵
+- 캥
+- Ē
+- ޣ
+- 咽
+- 婷
+- 榔
+- 漬
+- 瓊
+- 窰
+- 鼩
+- 꿇
+- 넬
+- 됬
+- 떳
+- 릅
+- 봬
+- 뽐
+- 삑
+- 삔
+- 죙
+- 퀵
+- 텼
+- 홋
+- 휙
+- 揾
+- 斧
+- 獸
+- 膏
+- 诈
+- 郝
+- 핥
+- 훑
+- ഊ
+- 堺
+- 懵
+- 擰
+- 烃
+- 瘾
+- 瞻
+- 舜
+- 萌
+- 蚌
+- 辟
+- 훗
+- 孚
+- 揉
+- 紛
+- 罷
+- 頑
+- 偽
+- 斩
+- 糸
+- 慨
+- 捲
+- 璧
+- 筷
+- 睿
+- 翡
+- 苷
+- 蝙
+- 冢
+- 佼
+- 岚
+- 胤
+- 讳
+- 佣
+- 邹
+- 乓
+- 钉
+- 寵
+- 帥
+- 榈
+- 尧
+- 槽
+- ఞ
+- 唉
+- 牠
+- 茵
+- ۳
+- 乍
+- 幡
+- 撲
+- 樊
+- 绯
+- 銃
+- 꽝
+- 뗐
+- 썪
+- 앓
+- 컥
+- Ź
+- Ҥ
+- ሑ
+- Ở
+- 琪
+- 竖
+- 糯
+- 缕
+- 腫
+- 蛭
+- 酗
+- 顏
+- 걀
+- 괘
+- 밉
+- 빕
+- 웜
+- 쟈
+- 휠
+- 唧
+- 敘
+- 疊
+- 癲
+- 缝
+- 賦
+- 鄞
+- 冻
+- 絞
+- 寅
+- 懋
+- 濒
+- 烛
+- 睜
+- 蛮
+- 谕
+- 皺
+- 厌
+- 吵
+- 漲
+- 汚
+- 韵
+- 劉
+- 砲
+- ぺ
+- 溲
+- 阐
+- 柜
+- 藨
+- 龐
+- 淹
+- 吼
+- 荚
+- 喬
+- 藜
+- 萸
+- Ձ
+- ఊ
+- 滯
+- 麽
+- 꿉
+- 잰
+- Ց
+- ՞
+- ሷ
+- ቾ
+- 厕
+- 嘥
+- 娥
+- 悖
+- 斬
+- 曰
+- 棵
+- 畏
+- 畿
+- 肴
+- 腥
+- 酉
+- 깽
+- 낌
+- 뽂
+- 켄
+- 쿤
+- 헉
+- ڱ
+- 峨
+- 朋
+- 枉
+- 皖
+- 駿
+- 斋
+- 熬
+- 纏
+- 襪
+- 趟
+- 釘
+- 陥
+- 骆
+- 喚
+- 羚
+- 俯
+- 兑
+- 歯
+- 渝
+- 彗
+- 抒
+- 鲇
+- 喵
+- 沁
+- 虔
+- 雳
+- 泪
+- 穷
+- 蔬
+- 拐
+- 摺
+- 舂
+- 帖
+- 铉
+- 憩
+- 咧
+- 譲
+- 嘟
+- 籲
+- 釜
+- 韌
+- Ď
+- Ķ
+- ۴
+- ۵
+- ۶
+- 倦
+- 偕
+- 塵
+- 掣
+- 暉
+- 瀨
+- 爷
+- 粋
+- 菸
+- 賠
+- 鶴
+- 곶
+- 끽
+- 둣
+- 떵
+- 맷
+- 셌
+- 솟
+- 츤
+- 푠
+- 呔
+- 嗽
+- 嘔
+- 拗
+- 泄
+- 煽
+- 챌
+- 捱
+- 浑
+- 鲸
+- 兮
+- 凑
+- 嚨
+- 幢
+- 栉
+- 楔
+- 萎
+- 荧
+- 跪
+- 锈
+- 畴
+- 譚
+- 镖
+- 斌
+- 吠
+- 篷
+- 醚
+- 钓
+- 闹
+- 饲
+- 晤
+- 蔚
+- 庁
+- 栓
+- 侃
+- 掷
+- 眶
+- 狙
+- 菝
+- 嬉
+- 尸
+- 註
+- 隍
+- 哉
+- 欽
+- 蜆
+- 雁
+- ಔ
+- ɔ
+- Ծ
+- ዦ
+- 崛
+- 揪
+- 盎
+- 箬
+- 艷
+- 鲅
+- 꺠
+- 낵
+- 냔
+- 눔
+- 뎀
+- 뎃
+- 몄
+- 뵀
+- 숑
+- 씸
+- 옇
+- 쨈
+- 탬
+- 俭
+- 疯
+- 竣
+- 締
+- 鹭
+- 壟
+- 嫂
+- 嵌
+- 檀
+- 筱
+- 簪
+- 谍
+- 骤
+- 谅
+- 騮
+- 騷
+- 娅
+- 汛
+- 軀
+- 侬
+- 哎
+- 挫
+- 皎
+- 侏
+- 繹
+- 镶
+- 荞
+- 嫔
+- 癡
+- 哮
+- 璋
+- Ḥ
+- 좀
+- 葜
+- 闯
+- 旬
+- 泵
+- 窒
+- 絨
+- 髀
+- ሦ
+- 壘
+- 汾
+- ቼ
+- 啄
+- 琅
+- 窿
+- 葳
+- 맏
+- 빰
+- 촘
+- ̈
+- ጌ
+- ឍ
+- 惘
+- 掰
+- 柺
+- 洱
+- 矾
+- 磚
+- 缨
+- 腧
+- 訣
+- 誡
+- 둠
+- 맬
+- 샾
+- 썽
+- 젬
+- 쥴
+- 冗
+- 剃
+- 吳
+- 眞
+- 篆
+- 胚
+- 脹
+- 栩
+- 沢
+- 祕
+- 祟
+- 肅
+- 铀
+- 龚
+- 凳
+- 剝
+- 忒
+- 撕
+- 紥
+- 纱
+- 虞
+- 鏈
+- 駄
+- 簇
+- 鑑
+- 堰
+- 毓
+- 窃
+- 谬
+- 轩
+- 姉
+- 晰
+- 闇
+- 闘
+- 聋
+- 烹
+- 牺
+- 俳
+- 茴
+- 刃
+- 嘯
+- 窦
+- 㩒
+- 捧
+- 耆
+- 傢
+- 沂
+- 秆
+- 删
+- 柵
+- 梧
+- 蚕
+- 驸
+- 齒
+- ઔ
+- ḿ
+- 睏
+- 苟
+- 갬
+- 촛
+- Ņ
+- ؔ
+- ኣ
+- ጧ
+- ᱪ
+- 偿
+- 匡
+- 叹
+- 峒
+- 栾
+- 棲
+- 脖
+- 赦
+- 鲡
+- 뗀
+- 뮌
+- 쉼
+- 쎘
+- 얏
+- 췌
+- 툼
+- 튠
+- ഛ
+- 揦
+- 漳
+- 烘
+- 祺
+- 鳖
+- 刹
+- 韧
+- 飄
+- 轎
+- 『
+- 皂
+- 蝎
+- 骼
+- 劾
+- 拦
+- 榻
+- 靶
+- 斟
+- 渥
+- 笃
+- 廂
+- 穀
+- 诫
+- 賬
+- 揣
+- 匣
+- 猩
+- 』
+- 啸
+- 赎
+- 蝨
+- 縛
+- 焰
+- 舆
+- 鄣
+- 閂
+- 鬃
+- 哼
+- 嚼
+- 懲
+- 熠
+- 戎
+- 歹
+- 荛
+- 邁
+- 陂
+- 鳕
+- Ώ
+- 偈
+- 囡
+- 趴
+- 閘
+- 텁
+- ሔ
+- ኗ
+- ፆ
+- 嘍
+- 屁
+- 揈
+- 灼
+- 畸
+- 稠
+- 笙
+- 籮
+- 遴
+- 驟
+- 麝
+- 麪
+- 꺄
+- 낡
+- 늠
+- 듭
+- 몹
+- 믈
+- 밧
+- 밲
+- 뼛
+- 샛
+- 섣
+- 屜
+- 翊
+- 耍
+- 苇
+- 葫
+- 辜
+- 酌
+- 鉤
+- 霎
+- 頗
+- 끙
+- 룐
+- 弛
+- 怠
+- 珉
+- 戳
+- 紋
+- 谟
+- 豈
+- 辱
+- 鄙
+- 钥
+- 彬
+- 慷
+- 潑
+- 葶
+- 躺
+- 冥
+- 澜
+- 裘
+- 鯊
+- 怯
+- 袍
+- 裙
+- 䁪
+- 佃
+- 勧
+- 弧
+- 拚
+- 溴
+- 绸
+- 钧
+- 煌
+- 祷
+- 釉
+- 瓯
+- 堕
+- 鹑
+- ゅ
+- 壆
+- 簕
+- 惨
+- 灶
+- 祉
+- 鹘
+- 剷
+- 擒
+- 梾
+- 蝗
+- 丼
+- 啞
+- 籃
+- 缶
+- 蟬
+- 遜
+- 錐
+- 훼
+- ©
+- ۚ
+- ۸
+- ጩ
+- ጾ
+- 嗡
+- 捡
+- 歼
+- 氹
+- 洮
+- 穹
+- 籬
+- 醛
+- 闢
+- 鶯
+- 麓
+- 굘
+- 궜
+- 낱
+- 룽
+- 뺨
+- 뽁
+- 싣
+- 쎌
+- 웁
+- 짙
+- 춧
+- 팎
+- 핼
+- ဩ
+- 厝
+- 娟
+- 弼
+- 彌
+- 擋
+- 檳
+- 欸
+- 盜
+- 篙
+- 緬
+- 蘅
+- 蠟
+- 貂
+- 躬
+- 朕
+- 蚵
+- ២
+- 箍
+- 锻
+- 佗
+- 咲
+- 熏
+- 紗
+- 鲨
+- 敲
+- 舶
+- 銘
+- 怦
+- 恍
+- 撳
+- 劝
+- 贰
+- 髒
+- 彝
+- 淑
+- 艱
+- 伺
+- 绅
+- 谎
+- 翘
+- 瑾
+- 苍
+- 蟑
+- 瞄
+- 赁
+- 迭
+- 荸
+- 婪
+- 曖
+- 溉
+- 绛
+- ဠ
+- 埼
+- 擲
+- 涧
+- 瀬
+- 稼
+- 訟
+- ఠ
+- 啶
+- 柽
+- 泮
+- 泻
+- 渾
+- 滌
+- 纂
+- 羹
+- 臟
+- 鹌
+- 헹
+- ¬
+- ۃ
+- ઍ
+- ዑ
+- ዢ
+- ៧
+- 啩
+- 弈
+- 簿
+- 糰
+- 缸
+- 荼
+- 郓
+- 饋
+- 鹂
+- 곗
+- 궐
+- 딥
+- 떄
+- 봔
+- 슉
+- 슘
+- 쯧
+- 챈
+- 톰
+- ઃ
+- ಊ
+- ጮ
+- 嘜
+- 夯
+- 屙
+- 徊
+- 徘
+- 晝
+- 洵
+- 漱
+- 诛
+- 빳
+- 샬
+- 刈
+- 巒
+- 羲
+- ᱬ
+- 彪
+- 拌
+- 挟
+- 焉
+- 砸
+- 菅
+- 霄
+- 掖
+- 晟
+- 饒
+- 噪
+- 孵
+- 浣
+- 猟
+- 襟
+- 攋
+- 疣
+- 謢
+- 剣
+- 瞳
+- 苓
+- 吁
+- 愕
+- 烦
+- 瞒
+- 羟
+- 谣
+- 噹
+- 沾
+- 潞
+- 聆
+- 啼
+- 濂
+- 痒
+- 顕
+- 墜
+- 滝
+- 颶
+- 墮
+- 捽
+- 褒
+- 坠
+- 囝
+- 聳
+- 邏
+- 垫
+- 貯
+- 垢
+- 埕
+- 寰
+- 帚
+- 滙
+- 綺
+- 葎
+- 馅
+- 桤
+- 俚
+- 绰
+- 郷
+- ƴ
+- ϋ
+- ዷ
+- 瑕
+- 腍
+- 셧
+- ʽ
+- Ί
+- ѕ
+- ኚ
+- ዐ
+- 妍
+- 屉
+- 屍
+- 捞
+- 櫻
+- 焊
+- 牴
+- 珧
+- 璇
+- 脓
+- 蒺
+- 蓣
+- 蕪
+- 虧
+- 裤
+- 隅
+- 靡
+- 鲹
+- 鸞
+- 黯
+- 굼
+- 긱
+- 꾀
+- 댐
+- 딤
+- 뗘
+- 멩
+- 멱
+- 빢
+- 뿅
+- 웍
+- 윷
+- 잦
+- 쟨
+- 쫘
+- 폿
+- 햐
+- 헥
+- ಛ
+- 伎
+- 咯
+- 栎
+- 眷
+- 镑
+- 唑
+- 唞
+- 畑
+- 鸻
+- 亏
+- 凿
+- 洼
+- 琢
+- 谏
+- 鎂
+- 崁
+- 脳
+- 漿
+- 炙
+- 祸
+- 轄
+- 顽
+- 奢
+- 妒
+- 涡
+- 涸
+- 钰
+- 鬣
+- 噓
+- 癱
+- 蚀
+- 贿
+- 寇
+- 粘
+- 倪
+- 箔
+- 摊
+- 睐
+- 娴
+- 渲
+- 谐
+- 橇
+- 脐
+- 鼹
+- 晾
+- 搣
+- 楝
+- 渚
+- 纖
+- 耽
+- 逹
+- 叭
+- 斉
+- 紳
+- 飩
+- Ľ
+- 墾
+- 捍
+- 輾
+- ٗ
+- ៖
+- ៣
+- ㄟ
+- 乂
+- 侮
+- 咆
+- 崙
+- 帛
+- 懈
+- 擸
+- 晉
+- 暹
+- 檻
+- 綱
+- 羧
+- 菏
+- 衅
+- 輻
+- 钇
+- 鯖
+- 걜
+- 냑
+- 넙
+- 똔
+- 띨
+- 쉘
+- 쏜
+- 옅
+- 짹
+- 휩
+- Ț
+- ᱝ
+- 濛
+- 糾
+- 莪
+- 鈕
+- 볐
+- 얍
+- 좆
+- 凋
+- 宸
+- 彤
+- 燴
+- 缉
+- 袜
+- 酪
+- 鳐
+- 叨
+- 媲
+- 晕
+- 淫
+- 籤
+- 紮
+- 芮
+- 蟾
+- 覓
+- 踐
+- 轿
+- 鈔
+- 鳢
+- 睾
+- 纔
+- 腈
+- 谴
+- 隼
+- 黏
+- 檐
+- 蛟
+- 誇
+- 靼
+- 岖
+- 岱
+- 愤
+- 炀
+- 羁
+- 逞
+- 鞑
+- 鸠
+- 浔
+- 钛
+- 鲮
+- 岬
+- 焚
+- 鬱
+- 珅
+- 蕴
+- 咎
+- 戮
+- 酚
+- 鲳
+- 왜
+- 賃
+- 僭
+- 鮫
+- 寥
+- 绡
+- 匂
+- 嶄
+- 掀
+- 歎
+- 捋
+- 砚
+- 肘
+- 犠
+- 睹
+- 覧
+- 賛
+- 洸
+- 纾
+- 菀
+- 逾
+- 镧
+- 蕗
+- 邕
+- ̔
+- 叢
+- 彷
+- 恵
+- ^
+- ×
+- ʺ
+- Љ
+- ӊ
+- Ӳ
+- 仮
+- 佰
+- 噃
+- 噸
+- 圻
+- 奀
+- 扔
+- 挠
+- 搗
+- 浛
+- 滔
+- 炫
+- 耘
+- 膩
+- 臻
+- 苈
+- 螳
+- 蠅
+- 裹
+- 铨
+- 頒
+- 궤
+- 넹
+- 덱
+- 땔
+- 볕
+- 뵌
+- 뵐
+- 쉭
+- 쌋
+- 옌
+- 윳
+- 젼
+- 쫒
+- 챘
+- 첵
+- 틋
+- 휜
+- ۹
+- ഃ
+- 恤
+- 揖
+- 樸
+- 茲
+- 詭
+- 赔
+- 遞
+- 鄢
+- 骏
+- 깻
+- 윰
+- ൌ
+- 奄
+- 姍
+- 愧
+- 扼
+- 溺
+- 珙
+- 蕎
+- 阙
+- 涟
+- ぇ
+- 兌
+- 吩
+- 咐
+- 憤
+- 渋
+- 瑛
+- 簷
+- 耸
+- 胰
+- 膊
+- 蘑
+- 辗
+- 酞
+- 霆
+- 鲱
+- 牢
+- 缔
+- 缚
+- 酔
+- 崴
+- 捩
+- 撫
+- 舄
+- 莼
+- 嘻
+- 塾
+- 桜
+- 涿
+- 苻
+- 泓
+- 盅
+- 蹦
+- 哨
+- 喧
+- 栅
+- 炕
+- 舎
+- 艙
+- 蕈
+- 誼
+- 謁
+- 饪
+- 搔
+- 拝
+- 塁
+- 浒
+- 驿
+- 謙
+- 褶
+- 捜
+- 砟
+- 膈
+- 煦
+- 遐
+- 捺
+- 蘊
+- 疮
+- 竜
+- 贱
+- 稃
+- 罟
+- 湳
+- 圏
+- 遏
+- 鯛
+- 嘘
+- 攜
+- 湊
+- 瑩
+- 臓
+- 芪
+- 雋
+- 偵
+- 巣
+- 惦
+- 桉
+- 渉
+- 祐
+- 驴
+- 骥
+- 魑
+- ޡ
+- ଈ
+- 剔
+- 暝
+- 梢
+- 甬
+- 禪
+- 竺
+- ŏ
+- ؑ
+- ዉ
+- 佥
+- 俸
+- 凄
+- 娑
+- 惶
+- 槤
+- 涷
+- 漾
+- 炽
+- 烽
+- 珏
+- 痾
+- 筐
+- 糞
+- 罄
+- 羈
+- 耿
+- 腭
+- 萃
+- 蔻
+- 谳
+- 贻
+- 迩
+- 邾
+- 閩
+- 骠
+- 겅
+- 곪
+- 넴
+- 뇽
+- 딧
+- 똠
+- 묽
+- 쐬
+- 옐
+- 웩
+- 쨰
+- 캅
+- Ģ
+- ቮ
+- 刪
+- 啜
+- 泼
+- 琛
+- 稲
+- 腮
+- 芷
+- 萱
+- 蔣
+- 蠱
+- 鰲
+- 黧
+- 댑
+- 숄
+- 캬
+- 콥
+- 탤
+- ๅ
+- 庾
+- 睫
+- 笋
+- 耧
+- 脇
+- 菰
+- 鑰
+- 隸
+- 雛
+- 顷
+- 骰
+- 呻
+- 墊
+- 惕
+- 膺
+- 曇
+- 湄
+- 錨
+- ヶ
+- 臃
+- 俬
+- 啉
+- 瀚
+- 烙
+- 珈
+- 癒
+- 臀
+- 蕭
+- 锑
+- 閪
+- 孪
+- 躝
+- 拷
+- 燦
+- 陇
+- 骸
+- 魇
+- 瞩
+- 朧
+- 綴
+- 渺
+- 胥
+- 禛
+- 挤
+- 柝
+- 腋
+- 漩
+- 窯
+- 蒨
+- 郦
+- 铬
+- 矫
+- 倬
+- 烁
+- 輛
+- 酢
+- 膛
+- 艮
+- 琐
+- 穏
+- 侥
+- 拢
+- 撩
+- 翩
+- 镊
+- 雏
+- 搽
+- 锌
+- 弩
+- 漓
+- 蛹
+- 嘧
+- 炜
+- 箜
+- 篌
+- 鉱
+- 彿
+- 擘
+- 绢
+- 吽
+- 懇
+- 訕
+- 攏
+- 桓
+- 縉
+- 魍
+- 魎
+- 鹼
+- ʾ
+- Ւ
+- ۗ
+- ఢ
+- ฯ
+- ኡ
+- ጹ
+- ፑ
+- 丶
+- 亢
+- 偃
+- 啰
+- 囌
+- 圩
+- 塚
+- 嫲
+- 恪
+- 昱
+- 棗
+- 瑧
+- 璟
+- 瓏
+- 畐
+- 痰
+- 痱
+- 笈
+- 粳
+- 紓
+- 胯
+- 腌
+- 腎
+- 臧
+- 菉
+- 蟠
+- 謄
+- 豂
+- 躉
+- 逅
+- 遁
+- 邂
+- 郫
+- 鋒
+- 鋸
+- 鑄
+- 闺
+- 雉
+- 頌
+- 鱷
+- 鲃
+- 鶏
+- 鼱
+- 귿
+- 꼿
+- 꿩
+- 듦
+- 딛
+- 띃
+- 숴
+- 짢
+- 쨋
+- 챕
+- 촤
+- 츰
+- 튈
+- 팰
+- 혓
+- §
+- ଋ
+- ዟ
+- ጢ
+- 僖
+- 尷
+- 愣
+- 杵
+- 楷
+- 潺
+- 潼
+- 祁
+- 縻
+- 翟
+- 驕
+- 냇
+- 닛
+- 샥
+- 쩝
+- 콰
+- ኅ
+- 亜
+- 媚
+- 尬
+- 戍
+- 渤
+- 燙
+- 犧
+- 疹
+- 篩
+- 钜
+- 齁
+- 丕
+- 呤
+- 嗆
+- 嘌
+- 挨
+- 柃
+- 氫
+- 鏟
+- 韶
+- 廬
+- 敖
+- 湛
+- 澡
+- 碾
+- 藁
+- 鈍
+- 鲑
+- 嬌
+- 痺
+- 铵
+- 堉
+- 撓
+- 楞
+- 渗
+- 濠
+- 禺
+- 绞
+- 賈
+- 顛
+- 朦
+- 璨
+- 蝸
+- 鰭
+- 喃
+- 垦
+- 孀
+- 瞰
+- 銳
+- 歓
+- 闌
+- 坯
+- 夭
+- 杷
+- 枇
+- 渎
+- 撑
+- 梠
+- 駭
+- 蚂
+- 阕
+- 掲
+- 쫌
+- 醸
+- 茁
+- 邙
+- 糬
+- 勐
+- 垮
+- 噗
+- 蜻
+- 洩
+- 涙
+- 軸
+- 鍛
+- ဋ
+- 霊
+- 囍
+- 恬
+- 淪
+- 癖
+- 舐
+- 蜃
+- 踎
+- 锯
+- 隘
+- 餛
+- 驰
+- 敝
+- 濁
+- 肛
+- 臼
+- 轸
+- 鑫
+- 陋
+- 쇽
+- "\x97"
+- Ƴ
+- ஔ
+- 跣
+- 헀
+- Ќ
+- ؓ
+- ઋ
+- Ả
+- ぁ
+- 刽
+- 塢
+- 墳
+- 姦
+- 徬
+- 扛
+- 抿
+- 摞
+- 摟
+- 昴
+- 暧
+- 桨
+- 桿
+- 楓
+- 殳
+- 氾
+- 潰
+- 烂
+- 瑭
+- 璜
+- 瘀
+- 瞌
+- 砰
+- 竅
+- 竿
+- 緻
+- 縈
+- 縝
+- 胪
+- 裾
+- 褪
+- 誅
+- 跖
+- 邈
+- 郿
+- 銜
+- 铰
+- 镌
+- 镓
+- 隷
+- 頬
+- 颍
+- 鳌
+- 鼐
+- 곯
+- 끕
+- 놉
+- 놋
+- 늄
+- 늪
+- 뗬
+- 랏
+- 앳
+- 읊
+- 잴
+- 캇
+- 켈
+- 켔
+- 튄
+- 튿
+- 푯
+- 훤
+- Ћ
+- ဈ
+- 朽
+- 杠
+- 秤
+- 舔
+- 蚶
+- 靂
+- 鱈
+- 鱔
+- 긌
+- ─
+- 勻
+- 廃
+- 歆
+- 甄
+- 祚
+- 簾
+- 绫
+- 踊
+- 闖
+- 驹
+- 鹛
+- ઐ
+- ಢ
+- 圃
+- 幔
+- 揩
+- 濡
+- 琶
+- 盏
+- 钴
+- 卉
+- 噜
+- 婺
+- 崧
+- 湍
+- 瀾
+- 竊
+- 翱
+- 苕
+- 蒼
+- 褚
+- 餚
+- 迂
+- 俨
+- 攔
+- 滥
+- 犒
+- 氓
+- 虏
+- 俏
+- 疙
+- 瘩
+- 綦
+- 駝
+- 騭
+- 鯉
+- 姒
+- 帧
+- 浇
+- 痢
+- 饥
+- 窥
+- 纭
+- 愫
+- 筛
+- 舗
+- 鲣
+- 惩
+- 釈
+- 琮
+- 淖
+- 胱
+- 铲
+- 旷
+- 碲
+- 狍
+- 渓
+- 燮
+- 揹
+- 镒
+- 垌
+- 邳
+- ㄧ
+- 堀
+- 丢
+- 挣
+- 揗
+- 摂
+- 罅
+- 褓
+- 锷
+- 麾
+- 覲
+- 粿
+- 肄
+- 蜓
+- 噩
+- 濺
+- 矗
+- 翦
+- 饕
+- 鸮
+- ಋ
+- 岿
+- 沔
+- 煉
+- 鬘
+- ឪ
+- 亀
+- 婉
+- 孭
+- 暄
+- 枋
+- 溞
+- 睥
+- 磯
+- 谜
+- 嗤
+- 陞
+- Ė
+- ̕
+- ۷
+- ଔ
+- ᱶ
+- 倩
+- 僵
+- 恙
+- 汙
+- 漚
+- 潲
+- 疵
+- 粽
+- 鬍
+- 썅
+- ­
+- ®
+- Ŵ
+- ̓
+- ٪
+- ୦
+- ៦
+- ᱸ
+- Ủ
+- ỵ
+- 仏
+- 刨
+- 咇
+- 嚀
+- 娇
+- 嫦
+- 宦
+- 幺
+- 怵
+- 恳
+- 慤
+- 懦
+- 戛
+- 撷
+- 旣
+- 晖
+- 暅
+- 暱
+- 柞
+- 楸
+- 樞
+- 毯
+- 渦
+- 瀉
+- 灞
+- 爨
+- 琥
+- 瓢
+- 癢
+- 睨
+- 瞞
+- 砝
+- 秧
+- 筠
+- 箪
+- 篡
+- 粼
+- 縷
+- 翹
+- 菔
+- 蓆
+- 蕩
+- 衿
+- 袱
+- 谙
+- 豌
+- 跛
+- 辍
+- 逍
+- 逕
+- 鉅
+- 銮
+- 颐
+- 颙
+- 餒
+- 鵪
+- 鶉
+- 갰
+- 곌
+- 꺽
+- 꿰
+- 뎁
+- 땋
+- 롷
+- 룔
+- 멓
+- 뼌
+- 뼘
+- 뿍
+- 셍
+- 쉑
+- 슁
+- 쑈
+- 욤
+- 윅
+- 읔
+- 쭌
+- 챗
+- 촐
+- 켠
+- 콸
+- 큭
+- 팻
+- 햅
+- 嗨
+- 桧
+- 樺
+- 煞
+- 爰
+- 瑤
+- 祗
+- 緘
+- 莘
+- 鈉
+- 鵬
+- 吝
+- 吶
+- 鱸
+- 롸
+- 凈
+- 峙
+- 毙
+- 疤
+- 稷
+- 蓓
+- 蜴
+- 讣
+- 辙
+- 黛
+- 恃
+- 掏
+- 榨
+- 羨
+- 茉
+- 荽
+- 躁
+- 嫻
+- 泞
+- 畳
+- 腱
+- 蜊
+- 骚
+- 忻
+- 擢
+- 簧
+- 骁
+- 侈
+- 咙
+- 巅
+- 葆
+- 釗
+- 鬠
+- 锺
+- 鋆
+- 卤
+- 粪
+- 镂
+- 糴
+- 饿
+- 鼆
+- 锰
+- 鲢
+- 鸥
+- 亵
+- 晗
+- 焙
+- 筥
+- 锴
+- 镀
+- 騾
+- 麋
+- 砕
+- 驯
+- 拇
+- 燶
+- 蝮
+- 迥
+- 揃
+- 噁
+- 胴
+- 挿
+- 熨
+- 舵
+- 鋭
+- 閨
+- 懺
+- 倧
+- 馈
+- 鲗
+- 鸵
+- 呟
+- 孺
+- 姫
+- 璀
+- 憊
+- 饗
+- 冧
+- 忟
+- 挈
+- 晌
+- 梘
+- 沆
+- 滁
+- 謎
+- 跚
+- 雯
+- 餼
+- 佇
+- 妆
+- 撿
+- 竄
+- 豎
+- 뤼
+- ఔ
+- ៥
+- 捌
+- 澹
+- 皋
+- 렜
+- 쇳
+- 숀
+- Ŋ
+- Ѕ
+- Ъ
+- ॊ
+- ፥
+- ៤
+- →
+- 刮
+- 吨
+- 咚
+- 咱
+- 唬
+- 嗚
+- 嚐
+- 奂
+- 孥
+- 屐
+- 巩
+- 弋
+- 怜
+- 恼
+- 憫
+- 懞
+- 扳
+- 拃
+- 拈
+- 捏
+- 掗
+- 掸
+- 摅
+- 擱
+- 柘
+- 槻
+- 氦
+- 沱
+- 洶
+- 涕
+- 淥
+- 湃
+- 溧
+- 滄
+- 牀
+- 牦
+- 牻
+- 玘
+- 玟
+- 琰
+- 璽
+- 瓒
+- 畋
+- 痤
+- 瘡
+- 盯
+- 硎
+- 禀
+- 穫
+- 窜
+- 篠
+- 繋
+- 舫
+- 蓑
+- 藿
+- 蚺
+- 諦
+- 貲
+- 赍
+- 踱
+- 踴
+- 踹
+- 蹲
+- 躇
+- 躊
+- 轍
+- 遼
+- 遽
+- 铕
+- 锶
+- 阆
+- 頰
+- 骂
+- 骊
+- 鮪
+- 鰈
+- 鰜
+- 鹳
+- 黍
+- 겡
+- 깟
+- 끍
+- 넋
+- 넵
+- 뉠
+- 뉩
+- 늉
+- 뎠
+- 뒨
+- 땁
+- 띤
+- 뭄
+- 쌘
+- 쑨
+- 엡
+- 쳔
+- 켐
+- 콱
+- 큔
+- 텄
+- 펩
+- 퓸
+- 핌
+- 홧
+- 휑
+- Ψ
+- ሼ
+- 劖
+- 喐
+- 拄
+- 枣
+- 瀏
+- 牯
+- 狹
+- 砷
+- 莒
+- 蕙
+- 딕
+- 폄
+- Ğ
+- ḅ
+- 駱
+- 俐
+- 僑
+- 婢
+- 澍
+- 絆
+- 繕
+- 薏
+- 诃
+- 遢
+- 邺
+- 閑
+- 飢
+- 饵
+- 鵲
+- 藕
+- 咿
+- 噎
+- 孰
+- 敕
+- 栢
+- 櫸
+- 涪
+- 澱
+- 甾
+- 羰
+- 蹴
+- 顰
+- 桢
+- 佤
+- 嗶
+- 奨
+- 嫉
+- 櫚
+- 瀕
+- 砻
+- 稈
+- 蜕
+- 閏
+- 傩
+- 喹
+- 昙
+- 棣
+- 淝
+- 玖
+- 誊
+- 镕
+- 阊
+- 隴
+- 馮
+- 哌
+- 炔
+- 肓
+- 諺
+- 賜
+- 邋
+- 閲
+- 饌
+- 娛
+- 蟒
+- 嚷
+- 剰
+- 寢
+- 捶
+- 斎
+- 犸
+- 竭
+- 铋
+- 锆
+- 鯨
+- 澀
+- 拧
+- 灏
+- 蟥
+- 颚
+- 锗
+- 砾
+- 钮
+- 譴
+- 疇
+- 铤
+- 廻
+- 邰
+- 噄
+- 捻
+- 藹
+- 踵
+- 鉢
+- 铢
+- 塀
+- 鐙
+- 蒴
+- 㪐
+- 吱
+- 圪
+- 掅
+- 矲
+- 禿
+- 脍
+- 蛻
+- 蜈
+- 蠹
+- 豇
+- 龛
+- 淵
+- 爿
+- 胿
+- 蚣
+- 㩿
+- 卞
+- 叱
+- 哣
+- 嵐
+- 撼
+- 昕
+- 殴
+- 沦
+- 滂
+- 瑆
+- 糋
+- 荥
+- 萋
+- 譬
+- 賒
+- 飨
+- Ÿ
+- ḫ
+- 嗅
+- 孜
+- 摷
+- 澧
+- 筍
+- 箇
+- 荘
+- 讎
+- 趷
+- 퓰
+- ሿ
+- 잿
+- 뚠
+- 뺌
+- 쌉
+- 쌨
+- ˮ
+- ̧
+- Ӯ
+- Ղ
+- ಝ
+- ዡ
+- ጵ
+- ፁ
+- ṅ
+- ⁠
+- √
+- ♫
+- ♭
+- 䰧
+- 佚
+- 俎
+- 俠
+- 厴
+- 吭
+- 呗
+- 唥
+- 唪
+- 喽
+- 嗒
+- 嗷
+- 囑
+- 坭
+- 垓
+- 奓
+- 岘
+- 崆
+- 巉
+- 幟
+- 彫
+- 徭
+- 悒
+- 惋
+- 惮
+- 愜
+- 慚
+- 憚
+- 戯
+- 掕
+- 揞
+- 摻
+- 擳
+- 斃
+- 旌
+- 昊
+- 昺
+- 椤
+- 楦
+- 榛
+- 樵
+- 橞
+- 欉
+- 歙
+- 歛
+- 泾
+- 濞
+- 炬
+- 犊
+- 犷
+- 狒
+- 獠
+- 珩
+- 琚
+- 畲
+- 瘁
+- 瘌
+- 皈
+- 皓
+- 皚
+- 眯
+- 睄
+- 睬
+- 瞅
+- 硞
+- 碉
+- 祆
+- 綾
+- 肨
+- 芎
+- 苴
+- 茯
+- 蒡
+- 蓁
+- 蔊
+- 薷
+- 蘸
+- 虓
+- 蛀
+- 蜇
+- 蜍
+- 蟀
+- 蟋
+- 蠶
+- 訐
+- 詡
+- 詬
+- 誨
+- 譖
+- 谚
+- 貶
+- 賄
+- 蹌
+- 蹣
+- 蹺
+- 躼
+- 輟
+- 迾
+- 郃
+- 錬
+- 钒
+- 铷
+- 锚
+- 锹
+- 镥
+- 隕
+- 靭
+- 韻
+- 頷
+- 頹
+- 餮
+- 饅
+- 馩
+- 鳚
+- 鳶
+- 鸽
+- 黾
+- 鼬
+- 겋
+- 괭
+- 궂
+- 꽌
+- 꽥
+- 넝
+- 뇸
+- 덨
+- 땟
+- 떫
+- 떽
+- 뚸
+- 렝
+- 밈
+- 뻬
+- 뽜
+- 삣
+- 쉈
+- 쌜
+- 썜
+- 쐐
+- 쓕
+- 씰
+- 읒
+- 읜
+- 잽
+- 젱
+- 죗
+- 짊
+- 쮸
+- 쯘
+- 쯩
+- 칡
+- 캤
+- 켤
+- 쿱
+- 퀀
+- 탯
+- 퉤
+- 팼
+- 헙
+- 횔
+- 힝
+- 荀
+- 룟
+- 휸
+- 佟
+- 坷
+- 塱
+- 壢
+- 扈
+- 抾
+- 泱
+- 琵
+- 谛
+- 鏊
+- 鑿
+- 钕
+- 顎
+- 駒
+- 倘
+- 箴
+- ଐ
+- 嗦
+- 嶋
+- 搁
+- 曙
+- 洎
+- 煜
+- 牘
+- 禱
+- 缪
+- 軚
+- 镐
+- 闰
+- 陲
+- 雒
+- 颊
+- 鸨
+- 麺
+- 黚
+- 黜
+- 宕
+- 挛
+- 摯
+- 沅
+- 謬
+- 錆
+- 颤
+- 暢
+- 冼
+- 勖
+- 勺
+- 啅
+- 孱
+- 幂
+- 斝
+- 橼
+- 珥
+- 踉
+- 鎭
+- 钯
+- 镆
+- 顼
+- 餡
+- 馭
+- 鲽
+- 鲿
+- 鹋
+- 鹩
+- 玲
+- 磊
+- 铂
+- 闫
+- 坍
+- 壕
+- 沤
+- 狡
+- 繊
+- 轔
+- 韋
+- 枷
+- 肼
+- 拮
+- 氡
+- 籐
+- 臥
+- 謾
+- 跷
+- 馔
+- 馗
+- 焕
+- 揽
+- 瘓
+- 蜣
+- 贅
+- 瘠
+- 沌
+- 걔
+- 餌
+- 嵴
+- 浄
+- 炱
+- 倖
+- 臆
+- 欒
+- 烨
+- 纥
+- 胭
+- 趖
+- 郛
+- 钳
+- 骹
+- 仞
+- 搾
+- 擂
+- 棧
+- 楣
+- 殻
+- 瘿
+- 簫
+- 裳
+- 霓
+- 昐
+- 栈
+- 蛍
+- 觚
+- 逵
+- 髻
+- 夔
+- 溥
+- 滲
+- 臍
+- 釵
+- 鈿
+- 閥
+- 鲾
+- 鹊
+- 唂
+- 寃
+- 戥
+- 烊
+- 碇
+- 稟
+- 缇
+- ៨
+- 匐
+- 奚
+- 懊
+- 曷
+- 楂
+- 螨
+- 趺
+- 镍
+- 펨
+- ±
+- ̋
+- ໊
+- ṯ
+- 嗯
+- 넒
+- 뙤
+- 뤠
+- ኸ
+- ዪ
+- ጐ
+- ጿ
+- ៩
+- ṉ
+- ∆
+- ●
+- 䆀
+- 佮
+- 佻
+- 俅
+- 儈
+- 剡
+- 劏
+- 厍
+- 咦
+- 咹
+- 啓
+- 喼
+- 嘸
+- 嚢
+- 圑
+- 塭
+- 妬
+- 庹
+- 弍
+- 徇
+- 忿
+- 悚
+- 惇
+- 戕
+- 戽
+- 拙
+- 拭
+- 捹
+- 掞
+- 掟
+- 掳
+- 摙
+- 柢
+- 桩
+- 桮
+- 梱
+- 梼
+- 榷
+- 樁
+- 橐
+- 殯
+- 毡
+- 氚
+- 汲
+- 涨
+- 淞
+- 淸
+- 渌
+- 溼
+- 煨
+- 熒
+- 玮
+- 璁
+- 璈
+- 疚
+- 疟
+- 瞑
+- 砒
+- 硃
+- 禡
+- 禦
+- 穌
+- 窠
+- 筜
+- 筼
+- 箓
+- 篳
+- 聼
+- 舢
+- 舨
+- 茔
+- 莴
+- 莽
+- 萺
+- 蓠
+- 蓺
+- 蔫
+- 虬
+- 蛏
+- 蛞
+- 蜢
+- 蝓
+- 螣
+- 袈
+- 裟
+- 諫
+- 诠
+- 谶
+- 賰
+- 趌
+- 轳
+- 辘
+- 酝
+- 鈣
+- 鉗
+- 鉸
+- 銭
+- 鍚
+- 鐡
+- 鐸
+- 顱
+- 飜
+- 饑
+- 馒
+- 駟
+- 鮟
+- 鯪
+- 鱇
+- 鳔
+- 鸾
+- 鹪
+- 굥
+- 귐
+- 꾜
+- 낏
+- 녈
+- 놥
+- 뉸
+- 댜
+- 떈
+- 떰
+- 롄
+- 맀
+- 먀
+- 멏
+- 멨
+- 몀
+- 뭍
+- 밌
+- 뱄
+- 뵤
+- 뻠
+- 숩
+- 쌂
+- 쑹
+- 씅
+- 앝
+- 옙
+- 좃
+- 줴
+- 짔
+- 쩠
+- 쬬
+- 칬
+- 켸
+- 쿰
+- 탰
+- 텟
+- 퇸
+- 팹
+- 폔
+- 퐈
+- 푤
+- 훙
+- 휀
+- 힉
+- 떔
+- ޥ
+- †
+- 罨
+- 혠
+- Ư
+- 冉
+- 恽
+- 擷
+- 昀
+- 桟
+- 泗
+- 泸
+- 璐
+- 盃
+- 讶
+- 锂
+- 鬢
+- 뾱
+- ּ
+- ऍ
+- ഢ
+- ኺ
+- 傃
+- 嚥
+- 芰
+- 蚝
+- 龢
+- 俣
+- 儍
+- 兖
+- 厩
+- 吋
+- 垄
+- 娯
+- 娼
+- 挞
+- 敞
+- 熹
+- 珒
+- 笹
+- 罔
+- 翎
+- 膦
+- 苳
+- 虜
+- 账
+- 钵
+- 俩
+- 儂
+- 劊
+- 呜
+- 庠
+- 挾
+- 斕
+- 梣
+- 櫈
+- 瀟
+- 癆
+- 碩
+- 绾
+- 荏
+- 褦
+- 諷
+- 鑠
+- 钞
+- 铯
+- 隹
+- 騅
+- 骢
+- 鲀
+- 倭
+- 傀
+- 儡
+- 巽
+- 捆
+- 柠
+- 汴
+- 縞
+- 萘
+- 蛎
+- 錠
+- 颉
+- 帷
+- 疃
+- 嗟
+- 洄
+- 灿
+- 猾
+- 硖
+- 祜
+- 紊
+- 衢
+- 诀
+- 躯
+- 醍
+- 醐
+- 鼾
+- 镫
+- 晁
+- 郸
+- 啻
+- 邯
+- 枠
+- 橢
+- 淌
+- 湮
+- 瘍
+- 襞
+- 髣
+- 綻
+- 佯
+- 唛
+- 嘿
+- 壌
+- 徨
+- 氙
+- 漪
+- 硼
+- 礫
+- 綢
+- 蒞
+- 蛰
+- 辇
+- 铳
+- 鰐
+- 擤
+- 猗
+- 姣
+- 孽
+- 搦
+- 杮
+- 枥
+- 柰
+- 棖
+- 煖
+- 箋
+- 綵
+- 胍
+- 苁
+- 苧
+- 茑
+- 莩
+- 薔
+- 衬
+- 衲
+- 諜
+- 賂
+- 邃
+- 铆
+- 隈
+- 甕
+- ኰ
+- ⋯
+- 傖
+- 兎
+- 凊
+- 嗝
+- 坩
+- 塙
+- 櫥
+- 沓
+- 痹
+- 糉
+- 莸
+- 薟
+- 蠕
+- 踅
+- 鉎
+- 鲫
+- 麭
+- ǫ
+- Ѹ
+- Ր
+- ḏ
+- ♯
+- 剾
+- 甌
+- 絚
+- 跩
+- 蹬
+- 빻
+- 쾰
+- 퀭
+- ḷ
+- 됴
+- 틔
+- "\x93"
+- "\x94"
+- Ð
+- Ę
+- Ʋ
+- ǎ
+- Ǹ
+- ȃ
+- ȏ
+- Ʉ
+- ̌
+- ̱
+- ۛ
+- ॄ
+- ॔
+- ୧
+- ቿ
+- ኟ
+- ዒ
+- ዖ
+- ጳ
+- ጴ
+- ጺ
+- ឮ
+- ḩ
+- Ḫ
+- ṇ
+- ↔
+- ≥
+- ⻣
+- 䱽
+- 亩
+- 亳
+- 佘
+- 俟
+- 俦
+- 僻
+- 冴
+- 剎
+- 匆
+- 叟
+- 呉
+- 呦
+- 呷
+- 唈
+- 啵
+- 喩
+- 喰
+- 嘰
+- 塬
+- 壱
+- 壸
+- 妊
+- 婊
+- 婕
+- 婵
+- 岀
+- 嶙
+- 巔
+- 巿
+- 廪
+- 弸
+- 恺
+- 悗
+- 悸
+- 戾
+- 扻
+- 抆
+- 揳
+- 搓
+- 撬
+- 斡
+- 昃
+- 曱
+- 栻
+- 椋
+- 楋
+- 橿
+- 檜
+- 歕
+- 殄
+- 沥
+- 浚
+- 淤
+- 淩
+- 渇
+- 濘
+- 烚
+- 焘
+- 燜
+- 燧
+- 猁
+- 猄
+- 猖
+- 猞
+- 獗
+- 珐
+- 琨
+- 瑋
+- 甴
+- 疍
+- 痣
+- 痧
+- 癀
+- 眵
+- 眸
+- 瞎
+- 瞪
+- 砥
+- 礑
+- 祢
+- 禰
+- 稞
+- 箩
+- 簍
+- 糍
+- 糜
+- 紡
+- 緲
+- 縹
+- 繄
+- 罠
+- 羣
+- 聿
+- 脛
+- 脯
+- 艋
+- 芜
+- 苎
+- 荇
+- 菴
+- 萦
+- 薮
+- 蜷
+- 蠣
+- 褸
+- 訛
+- 詮
+- 諧
+- 譎
+- 谒
+- 貘
+- 贛
+- 贮
+- 躋
+- 郯
+- 鄄
+- 鄱
+- 酣
+- 醪
+- 醺
+- 鎌
+- 铪
+- 锣
+- 闔
+- 闳
+- 霾
+- 靳
+- 鞄
+- 鞅
+- 餋
+- 馏
+- 馥
+- 驥
+- 骧
+- 髴
+- 鯇
+- 鱗
+- 鲲
+- 鲻
+- 鳎
+- 鴿
+- 鷄
+- 鷯
+- 鸝
+- 鼙
+- 龠
+- 갉
+- 갼
+- 겊
+- 겝
+- 곘
+- 껸
+- 꼄
+- 꼳
+- 끅
+- 냘
+- 넚
+- 놘
+- 눟
+- 뉜
+- 늣
+- 댤
+- 덫
+- 뎄
+- 됙
+- 듈
+- 땍
+- 뜀
+- 롹
+- 륑
+- 맜
+- 멎
+- 뭠
+- 뭡
+- 뭥
+- 뮴
+- 밎
+- 밨
+- 뱁
+- 뷴
+- 빘
+- 뻰
+- 뿟
+- 뿡
+- 삻
+- 샅
+- 샨
+- 셤
+- 셩
+- 셴
+- 솎
+- 쉣
+- 쉿
+- 슌
+- 슾
+- 썻
+- 쎅
+- 쎔
+- 쏵
+- 쏸
+- 쒯
+- 앎
+- 얉
+- 얐
+- 엤
+- 옜
+- 옫
+- 읃
+- 졍
+- 좍
+- 좐
+- 죵
+- 줜
+- 쥔
+- 쥘
+- 짆
+- 짯
+- 쨀
+- 쨉
+- 쩧
+- 쭙
+- 챂
+- 촙
+- 췻
+- 츨
+- 칟
+- 켕
+- 큄
+- 큣
+- 퉜
+- 핟
+- 햑
+- 헴
+- 헸
+- 힛
+- ῖ
+- 땃
+- 띔
+- 읎
+- 킁
+- "\x96"
+- ǰ
+- ഋ
+- 咄
+- 媠
+- 漕
+- 潦
+- 舺
+- 阇
+- 굣
+- 뭅
+- 푀
+- ʙ
+- 뱌
+- 䟴
+- 伶
+- 嚮
+- 揇
+- 敛
+- 椴
+- 槓
+- 殆
+- 犍
+- 瑄
+- 眩
+- 碁
+- 耙
+- 膣
+- 蔺
+- 诰
+- 赃
+- 赈
+- 赉
+- 踮
+- 钼
+- 鸯
+- 鹬
+- 尪
+- 倯
+- 匀
+- 匮
+- 吡
+- 啍
+- 啕
+- 啧
+- 啾
+- 嗄
+- 噛
+- 嚎
+- 妤
+- 娠
+- 媾
+- 嫣
+- 嵋
+- 彧
+- 悽
+- 拶
+- 掬
+- 摳
+- 摹
+- 擏
+- 旳
+- 枹
+- 柩
+- 汞
+- 獺
+- 珰
+- 珲
+- 瑳
+- 璉
+- 畝
+- 痉
+- 稜
+- 篋
+- 缮
+- 耦
+- 芃
+- 虻
+- 貞
+- 赂
+- 蹋
+- 酐
+- 鋤
+- 鏢
+- 鏰
+- 钐
+- 陝
+- 鹫
+- 坨
+- 覃
+- 〇
+- 傉
+- 哑
+- 嗓
+- 屹
+- 崭
+- 涞
+- 痪
+- 瘫
+- 粛
+- 縊
+- 苛
+- 蜉
+- 蜒
+- 蜿
+- 诡
+- 遡
+- 鵜
+- 齧
+- 虢
+- 颧
+- 咥
+- 杓
+- 桡
+- 簔
+- 綽
+- 飪
+- 诅
+- 딨
+- ၎
+- 덟
+- 얻
+- 젊
+- 옮
+- ٘
+- ଙ
+- 꽤
+- 짧
+- Ḍ
+- 훨
+- Ǧ
+- 굉
+- 걍
+- 첫
+- 굳
+- Ẓ
+- 浊
+- 聡
+- 荁
+- 뭘
+- ໋
+- 傕
+- 撸
+- 汜
+- 洙
+- 疽
+- 瘧
+- 癩
+- 纮
+- 蝥
+- 觐
+- 讃
+- 豔
+- ⿸
+- 疒
+- 硩
+- 伛
+- 姝
+- 庖
+- 徛
+- 攑
+- 枭
+- 烫
+- 睁
+- 矯
+- 竈
+- 绽
+- 翃
+- 聶
+- 胼
+- 膻
+- 芍
+- 荪
+- 蛸
+- 螟
+- 蠓
+- 谯
+- 跻
+- 蹒
+- 軋
+- 郏
+- 錘
+- 闍
+- 阖
+- 髎
+- 鮎
+- 鸲
+- 괜
+- 鱟
+- ṁ
+- 䒏
+- 刣
+- 勍
+- 匏
+- 咾
+- 哆
+- 嚾
+- 囥
+- 墘
+- 忤
+- 怛
+- 扲
+- 扴
+- 攥
+- 昉
+- 栃
+- 栲
+- 栳
+- 榫
+- 欑
+- 氘
+- 烰
+- 牁
+- 牂
+- 瑀
+- 甪
+- 畷
+- 疕
+- 疶
+- 筅
+- 緁
+- 緹
+- 羶
+- 肟
+- 蝰
+- 袓
+- 誑
+- 诲
+- 谪
+- 迌
+- 郴
+- 郾
+- 铽
+- 镉
+- 镱
+- 颞
+- 髡
+- 鵠
+- 鼷
+- Ù
+- ၤ
+- ≡
+- ≪
+- 尕
+- 敧
+- 敨
+- 桅
+- 檨
+- 狛
+- 痀
+- 睑
+- 鑢
+- 閻
+- 骯
+- Ɔ
+- Ӌ
+- Ӷ
+- Ầ
+- ∨
+- Ⱅ
+- ゔ
+- 㓁
+- 䖙
+- 毵
+- 낋
+- 녘
+- 벘
+- 봣
+- 졉
+- "\x8D"
+- "\x92"
+- Ï
+- Ơ
+- ǃ
+- ǐ
+- ɑ
+- ʔ
+- ʹ
+- ̐
+- ̟
+- ̠
+- ̨
+- ̪
+- ̲
+- Ѐ
+- Ђ
+- Ѓ
+- ַ
+- ٓ
+- ٖ
+- ٴ
+- ۘ
+- ऌ
+- ऎ
+- ऩ
+- ॆ
+- ઼
+- ଊ
+- ୄ
+- ୨
+- ୪
+- ୫
+- ୬
+- ୭
+- ୮
+- ୯
+- ఋ
+- ་
+- ན
+- ሓ
+- ቓ
+- ቚ
+- ኜ
+- ኻ
+- ጇ
+- ፗ
+- ፦
+- ឦ
+- ឫ
+- ឳ
+- ៎
+- ḡ
+- ṟ
+- Ẩ
+- Ố
+- Ὑ
+- ₽
+- ∞
+- ∼
+- ≫
+- ⋅
+- ◦
+- ◯
+- ♂
+- ✅
+- ✔
+- ⟨
+- ⟩
+- Ⱎ
+- 㧎
+- 䒐
+- 䴘
+- 䴙
+- 仄
+- 仝
+- 仱
+- 伉
+- 侖
+- 侷
+- 俍
+- 倣
+- 倻
+- 偌
+- 傭
+- 僫
+- 儆
+- 儑
+- 儕
+- 儬
+- 冑
+- 刍
+- 匾
+- 呎
+- 唌
+- 唎
+- 唏
+- 唖
+- 啋
+- 啪
+- 啮
+- 喾
+- 嗙
+- 噂
+- 嚓
+- 嚡
+- 囤
+- 囪
+- 坜
+- 坵
+- 埈
+- 埞
+- 埲
+- 埽
+- 堑
+- 墀
+- 壅
+- 壙
+- 奣
+- 妫
+- 娆
+- 嬴
+- 嬷
+- 尣
+- 尨
+- 屄
+- 屢
+- 岂
+- 岈
+- 岫
+- 峄
+- 峇
+- 嵖
+- 嵗
+- 嵯
+- 巯
+- 幃
+- 幄
+- 弔
+- 弢
+- 彊
+- 彥
+- 怙
+- 恂
+- 恣
+- 悌
+- 惚
+- 惬
+- 愷
+- 憍
+- 懍
+- 扂
+- 扽
+- 抌
+- 抦
+- 拑
+- 挹
+- 捅
+- 掻
+- 掾
+- 揆
+- 搐
+- 搡
+- 搤
+- 搩
+- 摃
+- 摑
+- 摖
+- 摵
+- 撨
+- 擛
+- 攕
+- 攘
+- 斂
+- 旆
+- 旮
+- 旯
+- 晷
+- 杙
+- 桄
+- 梆
+- 棨
+- 椏
+- 楗
+- 槱
+- 樫
+- 樾
+- 橈
+- 殮
+- 氈
+- 氖
+- 泅
+- 泠
+- 泯
+- 泷
+- 洘
+- 浉
+- 涓
+- 涝
+- 湫
+- 溅
+- 溋
+- 漉
+- 澈
+- 瀼
+- 灸
+- 烺
+- 焐
+- 熥
+- 熾
+- 牕
+- 犅
+- 玆
+- 玠
+- 玹
+- 珞
+- 珺
+- 璵
+- 瓘
+- 瓛
+- 瓮
+- 疡
+- 痟
+- 痼
+- 盱
+- 眙
+- 睞
+- 睢
+- 瞽
+- 矸
+- 矽
+- 砧
+- 磕
+- 磲
+- 秅
+- 穢
+- 窈
+- 窍
+- 窕
+- 窘
+- 笆
+- 筏
+- 筮
+- 筵
+- 箸
+- 簸
+- 籴
+- 粕
+- 粧
+- 粱
+- 絮
+- 綖
+- 繃
+- 繆
+- 纶
+- 纻
+- 绚
+- 缎
+- 缭
+- 缵
+- 罒
+- 罓
+- 翕
+- 翥
+- 聾
+- 腻
+- 膑
+- 膵
+- 苡
+- 茭
+- 荖
+- 莨
+- 菖
+- 菪
+- 萁
+- 葺
+- 蒟
+- 蓍
+- 蔑
+- 蕲
+- 薰
+- 藔
+- 藳
+- 虯
+- 蛉
+- 蛲
+- 蟆
+- 蟯
+- 衮
+- 袒
+- 袴
+- 覇
+- 觅
+- 觴
+- 詆
+- 詛
+- 諡
+- 謔
+- 讐
+- 讥
+- 讪
+- 讫
+- 诔
+- 诙
+- 诟
+- 诧
+- 诩
+- 诹
+- 谌
+- 谠
+- 赘
+- 赟
+- 赡
+- 赭
+- 趒
+- 跙
+- 蹆
+- 蹶
+- 軾
+- 輊
+- 輒
+- 輷
+- 轆
+- 轶
+- 轼
+- 辊
+- 辶
+- 迍
+- 邛
+- 邬
+- 酃
+- 醴
+- 鋐
+- 鎊
+- 鎧
+- 钲
+- 铖
+- 镛
+- 镠
+- 镡
+- 镤
+- 镳
+- 闱
+- 闷
+- 闾
+- 阉
+- 隗
+- 雺
+- 韬
+- 頤
+- 飧
+- 饯
+- 騃
+- 驛
+- 驺
+- 骄
+- 骘
+- 骷
+- 髅
+- 鮓
+- 鮨
+- 鰆
+- 鱉
+- 鲎
+- 鲟
+- 鲴
+- 鷲
+- 鸚
+- 鸳
+- 麂
+- 鼢
+- Ꝑ
+- 갗
+- 걈
+- 걧
+- 겻
+- 괍
+- 괏
+- 귯
+- 깞
+- 깸
+- 꺅
+- 꺤
+- 껨
+- 꼇
+- 꽐
+- 꽜
+- 꽨
+- 꽹
+- 꾿
+- 꿑
+- 꿘
+- 꿜
+- 뀜
+- 뀰
+- 낢
+- 냣
+- 뉼
+- 늗
+- 닜
+- 닪
+- 닯
+- 닸
+- 댇
+- 뎡
+- 돚
+- 돠
+- 됀
+- 됑
+- 됫
+- 딿
+- 떌
+- 떱
+- 뗌
+- 뗏
+- 똡
+- 뙈
+- 뛌
+- 뛘
+- 뜷
+- 뜽
+- 띈
+- 띌
+- 랒
+- 랟
+- 랮
+- 럐
+- 럤
+- 렿
+- 뢴
+- 룃
+- 뤤
+- 릈
+- 맫
+- 맽
+- 먁
+- 먄
+- 먕
+- 먙
+- 뫠
+- 뫼
+- 묀
+- 묜
+- 뭑
+- 뭬
+- 믁
+- 믕
+- 밷
+- 뱍
+- 뱡
+- 벋
+- 벛
+- 벰
+- 벱
+- 벵
+- 볌
+- 볏
+- 볻
+- 봈
+- 봘
+- 봠
+- 봡
+- 봥
+- 붇
+- 뷘
+- 뷜
+- 뷸
+- 빱
+- 빶
+- 빾
+- 뺘
+- 뺸
+- 뼉
+- 뼜
+- 뽝
+- 뿝
+- 쁩
+- 삯
+- 샆
+- 셕
+- 셥
+- 셸
+- 솝
+- 쉪
+- 슦
+- 싰
+- 싷
+- 썃
+- 쎠
+- 쎤
+- 쎼
+- 쏱
+- 쐴
+- 쒀
+- 쒸
+- 쒹
+- 씃
+- 앏
+- 앜
+- 앟
+- 앺
+- 얖
+- 얠
+- 얬
+- 얾
+- 엏
+- 옂
+- 옉
+- 옘
+- 옭
+- 옼
+- 왐
+- 왝
+- 왬
+- 왯
+- 왱
+- 욌
+- 욧
+- 웝
+- 웟
+- 웡
+- 읓
+- 읗
+- 읭
+- 읺
+- 잍
+- 잛
+- 쟝
+- 쟬
+- 젔
+- 젭
+- 젰
+- 젹
+- 졀
+- 졔
+- 좇
+- 죈
+- 죔
+- 죤
+- 쥣
+- 쥰
+- 쩄
+- 쩟
+- 쩬
+- Ứ
+- ኽ
+- 놧
+- 뇰
+- 맇
+- 묙
+- 쁸
+- 쌕
+- ¤
+- ՚
+- ః
+- ☆
+- 傈
+- 僳
+- 噤
+- 崑
+- 懣
+- 摠
+- 曽
+- 穂
+- 酺
+- 鉀
+- 麶
+- 黗
+- 됭
+- 싧
+- 쌧
+- ฃ
+- 傱
+- 匍
+- 唁
+- 噱
+- 堃
+- 堯
+- 壠
+- 奅
+- 嫖
+- 峥
+- 崞
+- 嵬
+- 巳
+- 忝
+- 恁
+- 愴
+- 揤
+- 攣
+- 晦
+- 暸
+- 杻
+- 楮
+- 樨
+- 橂
+- 欹
+- 殒
+- 殡
+- 沬
+- 涤
+- 湟
+- 漯
+- 澶
+- 熁
+- 狯
+- 獴
+- 玑
+- 玜
+- 玳
+- 瓠
+- 痠
+- 痫
+- 癫
+- 砦
+- 硨
+- 簁
+- 糶
+- 緯
+- 缜
+- 胛
+- 脣
+- 腑
+- 芫
+- 茌
+- 荦
+- 蚩
+- 蜚
+- 蜞
+- 螋
+- 蠋
+- 裇
+- 裎
+- 褧
+- 覕
+- 詣
+- 誣
+- 酂
+- 钫
+- 钿
+- 铌
+- 阏
+- 餉
+- 馀
+- 駈
+- 髭
+- 鼯
+- 䴕
+- 伫
+- 侗
+- 俹
+- 偲
+- 勲
+- 呕
+- 嚏
+- 圉
+- 垠
+- 埤
+- 岜
+- 崮
+- 廆
+- 愃
+- 慄
+- 掴
+- 攒
+- 昵
+- 曦
+- 栀
+- 桁
+- 槌
+- 槛
+- 殚
+- 浞
+- 淯
+- 潷
+- 濮
+- 瀍
+- 燹
+- 猝
+- 獰
+- 瑮
+- 疱
+- 瞿
+- 筈
+- 羡
+- 耻
+- 胄
+- 胳
+- 蓀
+- 裬
+- 觑
+- 諲
+- 謖
+- 讹
+- 诣
+- 赓
+- 輘
+- 辕
+- 辫
+- 逡
+- 遑
+- 釁
+- 鈞
+- 鈾
+- 铟
+- 铮
+- 锔
+- 镎
+- 镗
+- 闕
+- 闼
+- 阑
+- 霰
+- 靛
+- 颔
+- 驀
+- 驪
+- 鰓
+- 鶇
+- 떻
+- 娣
+- 暎
+- 滦
+- 醮
+- 俵
+- 剽
+- 圹
+- 奘
+- 孛
+- 揷
+- 搅
+- 搥
+- 潢
+- 焮
+- 燐
+- 獭
+- 笕
+- 絹
+- 翽
+- 臯
+- 蛊
+- 贄
+- 醯
+- 饴
+- 惺
+- 桝
+- 淅
+- 蝋
+- 馊
+- 뭔
+- 膾
+- 钪
+- 陛
+- ํ
+- 蔦
+- <sos>
+- <eos>
+- <sop>
+init: null
+input_size: null
+ctc_conf:
+    dropout_rate: 0.0
+    ctc_type: builtin
+    reduce: true
+    ignore_nan_grad: null
+    zero_infinity: true
+    brctc_risk_strategy: exp
+    brctc_group_strategy: end
+    brctc_risk_factor: 0.0
+use_preprocessor: true
+token_type: bpe
+bpemodel: data/token_list/bpe_unigram50000/bpe.model
+non_linguistic_symbols: null
+cleaner: null
+g2p: null
+speech_volume_normalize: null
+rir_scp: null
+rir_apply_prob: 1.0
+noise_scp: null
+noise_apply_prob: 1.0
+noise_db_range: '13_15'
+short_noise_thres: 0.5
+frontend: default
+frontend_conf:
+    n_fft: 512
+    win_length: 400
+    hop_length: 160
+    fs: 16k
+specaug: specaug
+specaug_conf:
+    apply_time_warp: false
+    time_warp_window: 5
+    time_warp_mode: bicubic
+    apply_freq_mask: true
+    freq_mask_width_range:
+    - 0
+    - 27
+    num_freq_mask: 2
+    apply_time_mask: true
+    time_mask_width_ratio_range:
+    - 0.0
+    - 0.05
+    num_time_mask: 4
+normalize: global_mvn
+normalize_conf:
+    stats_file: exp/s2t_stats_raw_bpe50000/train/feats_stats.npz
+model: espnet
+model_conf:
+    ctc_weight: 0.3
+    lsm_weight: 0.1
+    length_normalized_loss: false
+    sym_na: <na>
+preencoder: null
+preencoder_conf: {}
+encoder: e_branchformer
+encoder_conf:
+    output_size: 768
+    attention_heads: 12
+    attention_layer_type: selfattn
+    pos_enc_layer_type: abs_pos
+    rel_pos_type: latest
+    attention_qk_norm: false
+    use_flash_attn: true
+    cgmlp_linear_units: 3072
+    cgmlp_conv_kernel: 31
+    use_linear_after_conv: false
+    gate_activation: identity
+    num_blocks: 9
+    dropout_rate: 0.1
+    positional_dropout_rate: 0.1
+    attention_dropout_rate: 0.1
+    input_layer: conv2d
+    layer_drop_rate: 0.0
+    linear_units: 3072
+    positionwise_layer_type: linear
+    use_ffn: true
+    macaron_ffn: true
+    merge_conv_kernel: 31
+postencoder: null
+postencoder_conf: {}
+decoder: transformer
+decoder_conf:
+    attention_heads: 12
+    linear_units: 3072
+    num_blocks: 9
+    dropout_rate: 0.1
+    positional_dropout_rate: 0.1
+    self_attention_dropout_rate: 0.1
+    src_attention_dropout_rate: 0.1
+    self_attention_qk_norm: false
+    src_attention_qk_norm: false
+    self_attention_use_flash_attn: true
+    src_attention_use_flash_attn: true
+preprocessor: s2t
+preprocessor_conf:
+    text_prev_name: text_prev
+    text_ctc_name: text_ctc
+    fs: 16000
+    na_symbol: <na>
+    speech_length: 30
+    speech_resolution: 0.02
+    speech_init_silence: 30
+    text_prev_apply_prob: 0.5
+    time_apply_prob: 0.5
+    notime_symbol: <notimestamps>
+    first_time_symbol: <0.00>
+    last_time_symbol: <30.00>
+required:
+- output_dir
+- token_list
+version: '202310'
+distributed: true
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/acc.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/acc.png
new file mode 100644
index 0000000000000000000000000000000000000000..8853c5f3ff99a599ed85ea77fbeabef96e34a551
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/acc.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/backward_time.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/backward_time.png
new file mode 100644
index 0000000000000000000000000000000000000000..e37bd66df95d0642062db825c4361a51d9b3b046
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/backward_time.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/cer.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/cer.png
new file mode 100644
index 0000000000000000000000000000000000000000..b6b20af54573edce4fde3315e348c723226c82c9
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/cer.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/cer_ctc.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/cer_ctc.png
new file mode 100644
index 0000000000000000000000000000000000000000..5feb6b1f23f7925f3fb2cad32a9dc0cc44f230a2
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/cer_ctc.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/clip.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/clip.png
new file mode 100644
index 0000000000000000000000000000000000000000..a9b24f6cc646381cca9c7c86f6e9caa93d392363
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/clip.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/forward_time.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/forward_time.png
new file mode 100644
index 0000000000000000000000000000000000000000..aee3ce0141d56c8625e4a5a487417641b75f36ba
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/forward_time.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/gpu_max_cached_mem_GB.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/gpu_max_cached_mem_GB.png
new file mode 100644
index 0000000000000000000000000000000000000000..45a9711b1672380b1543bf0202f7ea8714549e4a
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/gpu_max_cached_mem_GB.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/grad_norm.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/grad_norm.png
new file mode 100644
index 0000000000000000000000000000000000000000..00924f4379f744f4519768bc04aad1f374ca5c0f
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/grad_norm.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/iter_time.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/iter_time.png
new file mode 100644
index 0000000000000000000000000000000000000000..ce713e68019e5dbf547d774cba411b466fa9f3af
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/iter_time.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss.png
new file mode 100644
index 0000000000000000000000000000000000000000..4aee7ca7fa6f29d69e6ff02f735d8105769a82bd
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss_att.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss_att.png
new file mode 100644
index 0000000000000000000000000000000000000000..42c14c5dc55d0d92845280b87a22849d6569a746
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss_att.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss_ctc.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss_ctc.png
new file mode 100644
index 0000000000000000000000000000000000000000..90e758a9b13dfb42158e79b3e5f37affd83e14aa
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss_ctc.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss_scale.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss_scale.png
new file mode 100644
index 0000000000000000000000000000000000000000..2d48607292409d98942f790b4ca148f4136f3210
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/loss_scale.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/optim0_lr0.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/optim0_lr0.png
new file mode 100644
index 0000000000000000000000000000000000000000..69fb67fa2fc9a8188ac4291daf9b5ec1511d48c3
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/optim0_lr0.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/optim_step_time.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/optim_step_time.png
new file mode 100644
index 0000000000000000000000000000000000000000..bdbe8df71deb27b5ad9d3e3af6a1fd0625784aa0
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/optim_step_time.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/train_time.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/train_time.png
new file mode 100644
index 0000000000000000000000000000000000000000..f933bbd07194bb16c54fc478f67c9468bf784e0d
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/train_time.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/wer.png b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/wer.png
new file mode 100644
index 0000000000000000000000000000000000000000..6cc79bc697417d3b0321e85def9640a11fe9d98c
Binary files /dev/null and b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/images/wer.png differ
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/train.2.log b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/train.2.log
new file mode 100644
index 0000000000000000000000000000000000000000..8f59a2d8e4a10e29cc5bc824a5528c23bbb9dab4
--- /dev/null
+++ b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/train.2.log
@@ -0,0 +1,1322 @@
+# Running on ip-10-0-194-127
+# Started at Mon Feb 26 17:43:20 UTC 2024
+# SLURMD_NODENAME=ip-10-0-194-127
+# SLURM_CLUSTER_NAME=external
+# SLURM_CONF=/opt/slurm/etc/slurm.conf
+# SLURM_CPUS_ON_NODE=80
+# SLURM_EXPORT_ENV=PATH
+# SLURM_GET_USER_ENV=1
+# SLURM_GPUS_ON_NODE=8
+# SLURM_GPUS_PER_NODE=8
+# SLURM_GTIDS=0
+# SLURM_JOBID=81044
+# SLURM_JOB_ACCOUNT=wavlab
+# SLURM_JOB_CPUS_PER_NODE='80(x2)'
+# SLURM_JOB_END_TIME=1709574196
+# SLURM_JOB_GID=1879800513
+# SLURM_JOB_GPUS=0,1,2,3,4,5,6,7
+# SLURM_JOB_ID=81044
+# SLURM_JOB_NAME=exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/train.log
+# SLURM_JOB_NODELIST=ip-10-0-194-127,ip-10-0-216-33
+# SLURM_JOB_NUM_NODES=2
+# SLURM_JOB_PARTITION=a40x
+# SLURM_JOB_QOS=normal
+# SLURM_JOB_START_TIME=1708969395
+# SLURM_JOB_UID=1879802134
+# SLURM_JOB_USER=pengyf
+# SLURM_LOCALID=0
+# SLURM_MEM_PER_CPU=10000
+# SLURM_NNODES=2
+# SLURM_NODEID=0
+# SLURM_NODELIST=ip-10-0-194-127,ip-10-0-216-33
+# SLURM_NODE_ALIASES='(null)'
+# SLURM_OPEN_MODE=a
+# SLURM_PRIO_PROCESS=0
+# SLURM_PROCID=0
+# SLURM_SCRIPT_CONTEXT=prolog_task
+# SLURM_SUBMIT_DIR=/weka/home-pengyf/espnet-owsm-train/egs2/owsm_v3.1_licensefree_nosa/s2t1
+# SLURM_SUBMIT_HOST=ip-172-17-22-214
+# SLURM_TASKS_PER_NODE='80(x2)'
+# SLURM_TASK_PID=32642
+# SLURM_TOPOLOGY_ADDR=ip-10-0-194-127
+# SLURM_TOPOLOGY_ADDR_PATTERN=node
+# SLURM_WORKING_CLUSTER=external:ip-172-17-55-206:6817:9984:109
+# srun --export=ALL python3 -m espnet2.bin.s2t_train --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev/wav.scp,speech,kaldi_ark --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/speech_shape --resume true --fold_length 80000 --output_dir exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000 --config conf/train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp/s2t_stats_raw_bpe50000/train/feats_stats.npz --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/wav.scp,speech,kaldi_ark --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/speech_shape --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.prev,text_prev,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_prev_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.ctc,text_ctc,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_ctc_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text,text,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_shape.bpe --multiple_iterator true --valid_data_path_and_name_and_type dump/raw/dev/text.prev,text_prev,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_prev_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text.ctc,text_ctc,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_ctc_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text,text,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_shape.bpe --ngpu 8 --multiprocessing_distributed true --dist_launcher slurm --dist_init_method file:///weka/home-pengyf/espnet-owsm-train/egs2/owsm_v3.1_licensefree_nosa/s2t1/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/.dist_init_6161be84-6c49-455e-b1b4-8fb570936836 
+/weka/home-pengyf/espnet-owsm-train/tools/venv/bin/python3 /weka/home-pengyf/espnet-owsm-train/espnet2/bin/s2t_train.py --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev/wav.scp,speech,kaldi_ark --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/speech_shape --resume true --fold_length 80000 --output_dir exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000 --config conf/train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp/s2t_stats_raw_bpe50000/train/feats_stats.npz --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/wav.scp,speech,kaldi_ark --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/speech_shape --fold_length 150 --train_data_path_and_/weka/home-pengyf/espnet-owsm-train/tools/venv/bin/python3 /weka/home-pengyf/espnet-owsm-train/espnet2/bin/s2t_train.py --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev/wav.scp,speech,kaldi_ark --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/speech_shape --resume true --fold_length 80000 --output_dir exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000 --config conf/train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp/s2t_stats_raw_bpe50000/train/feats_stats.npz --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/wav.scp,speech,kaldi_ark --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/speech_shape --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.prev,text_prev,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_prev_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.ctc,text_ctc,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_ctc_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text,text,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_shape.bpe --multiple_iterator true --valid_data_path_and_name_and_type dump/raw/dev/text.prev,text_prev,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_prev_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text.ctc,text_ctc,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_ctc_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text,text,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_shape.bpe --ngpu 8 --multiprocessing_distributed true --dist_launcher slurm --dist_init_method file:///weka/home-pengyf/espnet-owsm-train/egs2/owsm_v3.1_licensefree_nosa/s2t1/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/.dist_init_6161be84-6c49-455e-b1b4-8fb570936836
+name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.prev,text_prev,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_prev_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.ctc,text_ctc,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_ctc_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text,text,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_shape.bpe --multiple_iterator true --valid_data_path_and_name_and_type dump/raw/dev/text.prev,text_prev,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_prev_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text.ctc,text_ctc,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_ctc_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text,text,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_shape.bpe --ngpu 8 --multiprocessing_distributed true --dist_launcher slurm --dist_init_method file:///weka/home-pengyf/espnet-owsm-train/egs2/owsm_v3.1_licensefree_nosa/s2t1/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/.dist_init_6161be84-6c49-455e-b1b4-8fb570936836
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:34,876 (distributed_c10d:319) INFO: Added key: store_based_barrier_key:1 to store for rank: 0
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:35,274 (distributed_c10d:353) INFO: Rank 0: Completed store-based barrier for key:store_based_barrier_key:1 with 16 nodes.
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:35,305 (s2t:464) INFO: Vocabulary size: 50002
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:40,465 (abs_task:1271) INFO: pytorch.version=1.13.1+cu117, cuda.available=True, cudnn.version=8500, cudnn.benchmark=False, cudnn.deterministic=True
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:40,473 (abs_task:1272) INFO: Model structure:
+ESPnetS2TModel(
+  (frontend): DefaultFrontend(
+    (stft): Stft(n_fft=512, win_length=400, hop_length=160, center=True, normalized=False, onesided=True)
+    (frontend): Frontend()
+    (logmel): LogMel(sr=16000, n_fft=512, n_mels=80, fmin=0, fmax=8000.0, htk=False)
+  )
+  (specaug): SpecAug(
+    (freq_mask): MaskAlongAxis(mask_width_range=[0, 27], num_mask=2, axis=freq)
+    (time_mask): MaskAlongAxisVariableMaxWidth(mask_width_ratio_range=[0.0, 0.05], num_mask=4, axis=time)
+  )
+  (normalize): GlobalMVN(stats_file=exp/s2t_stats_raw_bpe50000/train/feats_stats.npz, norm_means=True, norm_vars=True)
+  (encoder): EBranchformerEncoder(
+    (embed): Conv2dSubsampling(
+      (conv): Sequential(
+        (0): Conv2d(1, 768, kernel_size=(3, 3), stride=(2, 2))
+        (1): ReLU()
+        (2): Conv2d(768, 768, kernel_size=(3, 3), stride=(2, 2))
+        (3): ReLU()
+      )
+      (out): Sequential(
+        (0): Linear(in_features=14592, out_features=768, bias=True)
+        (1): PositionalEncoding(
+          (dropout): Dropout(p=0.1, inplace=False)
+        )
+      )
+    )
+    (encoders): MultiSequential(
+      (0): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (1): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (2): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (3): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (4): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (5): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (6): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (7): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (8): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+    )
+    (after_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+  )
+  (decoder): TransformerDecoder(
+    (embed): Sequential(
+      (0): Embedding(50002, 768)
+      (1): PositionalEncoding(
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+    )
+    (after_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+    (output_layer): Linear(in_features=768, out_features=50002, bias=True)
+    (decoders): MultiSequential(
+      (0): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (1): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (2): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (3): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (4): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (5): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (6): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (7): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (8): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+    )
+  )
+  (criterion_att): LabelSmoothingLoss(
+    (criterion): KLDivLoss()
+  )
+  (ctc): CTC(
+    (ctc_lo): Linear(in_features=768, out_features=50002, bias=True)
+    (ctc_loss): CTCLoss()
+  )
+)
+
+Model summary:
+    Class Name: ESPnetS2TModel
+    Total Number of model parameters: 366.65 M
+    Number of trainable parameters: 366.65 M (100.0%)
+    Size: 1.47 GB
+    Type: torch.float32
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:40,473 (abs_task:1275) INFO: Optimizer:
+AdamW (
+Parameter Group 0
+    amsgrad: False
+    betas: [0.9, 0.98]
+    capturable: False
+    eps: 1e-06
+    foreach: None
+    initial_lr: 0.0005
+    lr: 1.6666666666666667e-09
+    maximize: False
+    weight_decay: 0.0
+)
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:40,473 (abs_task:1276) INFO: Scheduler: PiecewiseLinearWarmupLR(warmup_steps_list=[0, 30000, 60000], warmup_lr_list=[0.0, 5e-05, 0.0005])
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:40,476 (abs_task:1285) INFO: Saving the configuration in exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/config.yaml
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:44,171 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:44,716 (abs_task:1663) INFO: [valid] dataset:
+ESPnetDataset(
+  speech: {"path": "dump/raw/dev/wav.scp", "type": "kaldi_ark"}
+  text_prev: {"path": "dump/raw/dev/text.prev", "type": "text"}
+  text_ctc: {"path": "dump/raw/dev/text.ctc", "type": "text"}
+  text: {"path": "dump/raw/dev/text", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7fe3f376e0a0>)
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:44,716 (abs_task:1664) INFO: [valid] Batch sampler: UnsortedBatchSampler(N-batch=2133, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/valid/speech_shape, 
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:44,717 (abs_task:1665) INFO: [valid] mini-batch sizes summary: N-batch=2133, mean=256.1, min=256, max=257
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Tracking run with wandb version 0.16.3
+wandb: Run data is saved locally in exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/wandb/run-20240226_174347-zw9g4dhv
+wandb: Run `wandb offline` to turn off syncing.
+wandb: Syncing run _weka_home-pengyf_espnet-owsm-train_egs2_owsm_v3.1_licensefree_nosa_s2t1
+wandb: ⭐️ View project at https://stability.wandb.io/pengyf/ESPnet_S2TTask
+wandb: 🚀 View run at https://stability.wandb.io/pengyf/ESPnet_S2TTask/runs/zw9g4dhv
+wandb: WARNING Serializing object of type list that is 400072 bytes
+wandb: WARNING Serializing object of type list that is 406488 bytes
+ip-10-0-194-127:32770:32770 [0] NCCL INFO Bootstrap : Using ens32:10.0.194.127<0>
+ip-10-0-194-127:32770:32770 [0] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-194-127:32770:32770 [0] NCCL INFO cudaDriverVersion 12000
+NCCL version 2.14.3+cuda11.7
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:51,253 (trainer:302) INFO: 1/45epoch started
+[ip-10-0-194-127:0/16] 2024-02-26 17:43:51,295 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-194-127:0/16] 2024-02-26 17:44:27,554 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-194-127:0/16] 2024-02-26 17:44:35,530 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7fe2fefab8b0>)
+[ip-10-0-194-127:0/16] 2024-02-26 17:44:35,530 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-194-127:0/16] 2024-02-26 17:44:35,535 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+ip-10-0-216-33:179936:179936 [2] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:179936:179936 [2] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:179936:179936 [2] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:179936:180196 [2] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:179936:180196 [2] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:179936:180196 [2] NCCL INFO Using network Socket
+ip-10-0-216-33:179936:180196 [2] NCCL INFO Setting affinity for GPU 2 to 0f,ffff0000,000fffff
+ip-10-0-216-33:179936:180196 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9
+ip-10-0-216-33:179936:180196 [2] NCCL INFO Channel 00/0 : 10[201c0] -> 9[101d0] via P2P/IPC/read
+ip-10-0-216-33:179936:180196 [2] NCCL INFO Channel 01/0 : 10[201c0] -> 9[101d0] via P2P/IPC/read
+ip-10-0-216-33:179936:180196 [2] NCCL INFO Connected all rings
+ip-10-0-216-33:179936:180196 [2] NCCL INFO Channel 00/0 : 10[201c0] -> 11[201d0] via P2P/IPC/read
+ip-10-0-216-33:179936:180196 [2] NCCL INFO Channel 01/0 : 10[201c0] -> 11[201d0] via P2P/IPC/read
+ip-10-0-216-33:179936:180196 [2] NCCL INFO Connected all trees
+ip-10-0-216-33:179936:180196 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:179936:180196 [2] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:179936:180196 [2] NCCL INFO comm 0x4dcec360 rank 10 nranks 16 cudaDev 2 busId 201c0 - Init COMPLETE
+ip-10-0-216-33:179937:179937 [3] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:179937:179937 [3] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:179937:179937 [3] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:179937:180193 [3] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:179937:180193 [3] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:179937:180193 [3] NCCL INFO Using network Socket
+ip-10-0-216-33:179937:180193 [3] NCCL INFO Setting affinity for GPU 3 to 0f,ffff0000,000fffff
+ip-10-0-216-33:179937:180193 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10
+ip-10-0-216-33:179937:180193 [3] NCCL INFO Channel 00/0 : 11[201d0] -> 10[201c0] via P2P/IPC/read
+ip-10-0-216-33:179937:180193 [3] NCCL INFO Channel 01/0 : 11[201d0] -> 10[201c0] via P2P/IPC/read
+ip-10-0-216-33:179937:180193 [3] NCCL INFO Connected all rings
+ip-10-0-216-33:179937:180193 [3] NCCL INFO Channel 00/0 : 11[201d0] -> 12[901c0] via P2P/IPC/read
+ip-10-0-216-33:179937:180193 [3] NCCL INFO Channel 01/0 : 11[201d0] -> 12[901c0] via P2P/IPC/read
+ip-10-0-216-33:179937:180193 [3] NCCL INFO Connected all trees
+ip-10-0-216-33:179937:180193 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:179937:180193 [3] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:179937:180193 [3] NCCL INFO comm 0x50f3b9d0 rank 11 nranks 16 cudaDev 3 busId 201d0 - Init COMPLETE
+ip-10-0-216-33:179935:179935 [1] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:179935:179935 [1] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:179935:179935 [1] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:179935:180195 [1] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:179935:180195 [1] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Using network Socket
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Setting affinity for GPU 1 to 0f,ffff0000,000fffff
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->8
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Channel 00/0 : 9[101d0] -> 0[101c0] [send] via NET/Socket/0
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Channel 01/0 : 9[101d0] -> 0[101c0] [send] via NET/Socket/0
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Connected all rings
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Channel 00/0 : 9[101d0] -> 10[201c0] via P2P/IPC/read
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Channel 01/0 : 9[101d0] -> 10[201c0] via P2P/IPC/read
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Channel 00/0 : 9[101d0] -> 8[101c0] via P2P/IPC/read
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Channel 01/0 : 9[101d0] -> 8[101c0] via P2P/IPC/read
+ip-10-0-216-33:179935:180195 [1] NCCL INFO Connected all trees
+ip-10-0-216-33:179935:180195 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:179935:180195 [1] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:179935:180195 [1] NCCL INFO comm 0x12fbe1c0 rank 9 nranks 16 cudaDev 1 busId 101d0 - Init COMPLETE
+ip-10-0-194-127:32774:32774 [4] NCCL INFO cudaDriverVersion 12000
+ip-10-0-194-127:32774:32774 [4] NCCL INFO Bootstrap : Using ens32:10.0.194.127<0>
+ip-10-0-194-127:32774:32774 [4] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-194-127:32774:33200 [4] NCCL INFO NET/IB : No device found.
+ip-10-0-194-127:32774:33200 [4] NCCL INFO NET/Socket : Using [0]ens32:10.0.194.127<0>
+ip-10-0-194-127:32774:33200 [4] NCCL INFO Using network Socket
+ip-10-0-194-127:32774:33200 [4] NCCL INFO Setting affinity for GPU 4 to 0fffff00,00000fff,ff000000
+ip-10-0-194-127:32774:33200 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3
+ip-10-0-194-127:32774:33200 [4] NCCL INFO Channel 00/0 : 4[901c0] -> 3[201d0] via P2P/IPC/read
+ip-10-0-194-127:32774:33200 [4] NCCL INFO Channel 01/0 : 4[901c0] -> 3[201d0] via P2P/IPC/read
+ip-10-0-194-127:32774:33200 [4] NCCL INFO Connected all rings
+ip-10-0-194-127:32774:33200 [4] NCCL INFO Channel 00/0 : 4[901c0] -> 5[901d0] via P2P/IPC/read
+ip-10-0-216-33:179941:179941 [7] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:179941:179941 [7] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:179941:179941 [7] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:179941:180191 [7] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:179941:180191 [7] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:179941:180191 [7] NCCL INFO Using network Socket
+ip-10-0-216-33:179941:180191 [7] NCCL INFO Setting affinity for GPU 7 to 0fffff00,00000fff,ff000000
+ip-10-0-216-33:179941:180191 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14
+ip-10-0-216-33:179941:180191 [7] NCCL INFO Channel 00/0 : 15[a01d0] -> 14[a01c0] via P2P/IPC/read
+ip-10-0-216-33:179941:180191 [7] NCCL INFO Channel 01/0 : 15[a01d0] -> 14[a01c0] via P2P/IPC/read
+ip-10-0-216-33:179941:180191 [7] NCCL INFO Connected all rings
+ip-10-0-216-33:179941:180191 [7] NCCL INFO Connected all trees
+ip-10-0-194-127:32774:33200 [4] NCCL INFO Channel 01/0 : 4[901c0] -> 5[901d0] via P2P/IPC/read
+ip-10-0-194-127:32774:33200 [4] NCCL INFO Connected all trees
+ip-10-0-194-127:32774:33200 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-194-127:32774:33200 [4] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-194-127:32774:33200 [4] NCCL INFO comm 0x1233b9a0 rank 4 nranks 16 cudaDev 4 busId 901c0 - Init COMPLETE
+ip-10-0-216-33:179941:180191 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:179941:180191 [7] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:179941:180191 [7] NCCL INFO comm 0x873b790 rank 15 nranks 16 cudaDev 7 busId a01d0 - Init COMPLETE
+ip-10-0-194-127:32771:32771 [1] NCCL INFO cudaDriverVersion 12000
+ip-10-0-194-127:32771:32771 [1] NCCL INFO Bootstrap : Using ens32:10.0.194.127<0>
+ip-10-0-194-127:32771:32771 [1] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-194-127:32771:33199 [1] NCCL INFO NET/IB : No device found.
+ip-10-0-194-127:32771:33199 [1] NCCL INFO NET/Socket : Using [0]ens32:10.0.194.127<0>
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Using network Socket
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Setting affinity for GPU 1 to 0f,ffff0000,000fffff
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Channel 00/0 : 1[101d0] -> 8[101c0] [send] via NET/Socket/0
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Channel 01/0 : 1[101d0] -> 8[101c0] [send] via NET/Socket/0
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Connected all rings
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Channel 00/0 : 1[101d0] -> 2[201c0] via P2P/IPC/read
+ip-10-0-216-33:179934:179934 [0] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:179934:179934 [0] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:179934:179934 [0] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:179934:180190 [0] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:179934:180190 [0] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Using network Socket
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Setting affinity for GPU 0 to 0f,ffff0000,000fffff
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Trees [0] 9/-1/-1->8->0 [1] 9/0/-1->8->-1
+ip-10-0-216-33:179934:180205 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Channel 00/0 : 1[101d0] -> 8[101c0] [receive] via NET/Socket/0
+ip-10-0-216-33:179934:180205 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Channel 01/0 : 1[101d0] -> 2[201c0] via P2P/IPC/read
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Channel 00/0 : 1[101d0] -> 0[101c0] via P2P/IPC/read
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Channel 01/0 : 1[101d0] -> 0[101c0] via P2P/IPC/read
+ip-10-0-194-127:32771:33199 [1] NCCL INFO Connected all trees
+ip-10-0-194-127:32771:33199 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-194-127:32771:33199 [1] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-194-127:32771:33199 [1] NCCL INFO comm 0x5084f1f0 rank 1 nranks 16 cudaDev 1 busId 101d0 - Init COMPLETE
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Channel 01/0 : 1[101d0] -> 8[101c0] [receive] via NET/Socket/0
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Channel 00/0 : 8[101c0] -> 15[a01d0] via P2P/IPC/read
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Channel 01/0 : 8[101c0] -> 15[a01d0] via P2P/IPC/read
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Connected all rings
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Channel 00/0 : 8[101c0] -> 9[101d0] via P2P/IPC/read
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Channel 01/0 : 8[101c0] -> 9[101d0] via P2P/IPC/read
+ip-10-0-216-33:179934:180205 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Channel 00/0 : 0[101c0] -> 8[101c0] [receive] via NET/Socket/0
+ip-10-0-216-33:179934:180205 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Channel 01/0 : 0[101c0] -> 8[101c0] [receive] via NET/Socket/0
+ip-10-0-194-127:32773:32773 [3] NCCL INFO cudaDriverVersion 12000
+ip-10-0-194-127:32773:32773 [3] NCCL INFO Bootstrap : Using ens32:10.0.194.127<0>
+ip-10-0-194-127:32773:32773 [3] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-194-127:32773:33195 [3] NCCL INFO NET/IB : No device found.
+ip-10-0-194-127:32773:33195 [3] NCCL INFO NET/Socket : Using [0]ens32:10.0.194.127<0>
+ip-10-0-194-127:32773:33195 [3] NCCL INFO Using network Socket
+ip-10-0-194-127:32773:33195 [3] NCCL INFO Setting affinity for GPU 3 to 0f,ffff0000,000fffff
+ip-10-0-194-127:32773:33195 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2
+ip-10-0-194-127:32773:33195 [3] NCCL INFO Channel 00/0 : 3[201d0] -> 2[201c0] via P2P/IPC/read
+ip-10-0-194-127:32773:33195 [3] NCCL INFO Channel 01/0 : 3[201d0] -> 2[201c0] via P2P/IPC/read
+ip-10-0-194-127:32773:33195 [3] NCCL INFO Connected all rings
+ip-10-0-194-127:32773:33195 [3] NCCL INFO Channel 00/0 : 3[201d0] -> 4[901c0] via P2P/IPC/read
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Channel 00/0 : 8[101c0] -> 0[101c0] [send] via NET/Socket/0
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Channel 01/0 : 8[101c0] -> 0[101c0] [send] via NET/Socket/0
+ip-10-0-216-33:179934:180190 [0] NCCL INFO Connected all trees
+ip-10-0-216-33:179934:180190 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:179934:180190 [0] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:179934:180190 [0] NCCL INFO comm 0x9464c40 rank 8 nranks 16 cudaDev 0 busId 101c0 - Init COMPLETE
+ip-10-0-194-127:32773:33195 [3] NCCL INFO Channel 01/0 : 3[201d0] -> 4[901c0] via P2P/IPC/read
+ip-10-0-194-127:32773:33195 [3] NCCL INFO Connected all trees
+ip-10-0-194-127:32773:33195 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-194-127:32773:33195 [3] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-194-127:32773:33195 [3] NCCL INFO comm 0x52a8a440 rank 3 nranks 16 cudaDev 3 busId 201d0 - Init COMPLETE
+ip-10-0-216-33:179938:179938 [4] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:179938:179938 [4] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:179938:179938 [4] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:179938:180192 [4] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:179938:180192 [4] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:179938:180192 [4] NCCL INFO Using network Socket
+ip-10-0-216-33:179938:180192 [4] NCCL INFO Setting affinity for GPU 4 to 0fffff00,00000fff,ff000000
+ip-10-0-216-33:179938:180192 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11
+ip-10-0-216-33:179938:180192 [4] NCCL INFO Channel 00/0 : 12[901c0] -> 11[201d0] via P2P/IPC/read
+ip-10-0-216-33:179938:180192 [4] NCCL INFO Channel 01/0 : 12[901c0] -> 11[201d0] via P2P/IPC/read
+ip-10-0-216-33:179938:180192 [4] NCCL INFO Connected all rings
+ip-10-0-194-127:32770:33193 [0] NCCL INFO NET/IB : No device found.
+ip-10-0-194-127:32770:33193 [0] NCCL INFO NET/Socket : Using [0]ens32:10.0.194.127<0>
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Using network Socket
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Setting affinity for GPU 0 to 0f,ffff0000,000fffff
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 00/02 :    0   7   6   5   4   3   2   1   8  15  14  13  12  11  10   9
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 01/02 :    0   7   6   5   4   3   2   1   8  15  14  13  12  11  10   9
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Trees [0] 1/8/-1->0->-1 [1] 1/-1/-1->0->8
+ip-10-0-194-127:32770:33203 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 00/0 : 9[101d0] -> 0[101c0] [receive] via NET/Socket/0
+ip-10-0-194-127:32770:33203 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-216-33:179938:180192 [4] NCCL INFO Channel 00/0 : 12[901c0] -> 13[901d0] via P2P/IPC/read
+ip-10-0-216-33:179938:180192 [4] NCCL INFO Channel 01/0 : 12[901c0] -> 13[901d0] via P2P/IPC/read
+ip-10-0-216-33:179938:180192 [4] NCCL INFO Connected all trees
+ip-10-0-216-33:179938:180192 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:179938:180192 [4] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:179938:180192 [4] NCCL INFO comm 0x50ef4030 rank 12 nranks 16 cudaDev 4 busId 901c0 - Init COMPLETE
+ip-10-0-216-33:179940:179940 [6] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:179940:179940 [6] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:179940:179940 [6] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:179940:180194 [6] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:179940:180194 [6] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 01/0 : 9[101d0] -> 0[101c0] [receive] via NET/Socket/0
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 00/0 : 0[101c0] -> 7[a01d0] via P2P/IPC/read
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 01/0 : 0[101c0] -> 7[a01d0] via P2P/IPC/read
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Connected all rings
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 00/0 : 0[101c0] -> 1[101d0] via P2P/IPC/read
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 01/0 : 0[101c0] -> 1[101d0] via P2P/IPC/read
+ip-10-0-194-127:32770:33203 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 00/0 : 8[101c0] -> 0[101c0] [receive] via NET/Socket/0
+ip-10-0-194-127:32770:33203 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 01/0 : 8[101c0] -> 0[101c0] [receive] via NET/Socket/0
+ip-10-0-216-33:179940:180194 [6] NCCL INFO Using network Socket
+ip-10-0-216-33:179940:180194 [6] NCCL INFO Setting affinity for GPU 6 to 0fffff00,00000fff,ff000000
+ip-10-0-216-33:179940:180194 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13
+ip-10-0-216-33:179940:180194 [6] NCCL INFO Channel 00/0 : 14[a01c0] -> 13[901d0] via P2P/IPC/read
+ip-10-0-216-33:179940:180194 [6] NCCL INFO Channel 01/0 : 14[a01c0] -> 13[901d0] via P2P/IPC/read
+ip-10-0-216-33:179940:180194 [6] NCCL INFO Connected all rings
+ip-10-0-216-33:179940:180194 [6] NCCL INFO Channel 00/0 : 14[a01c0] -> 15[a01d0] via P2P/IPC/read
+ip-10-0-216-33:179940:180194 [6] NCCL INFO Channel 01/0 : 14[a01c0] -> 15[a01d0] via P2P/IPC/read
+ip-10-0-216-33:179940:180194 [6] NCCL INFO Connected all trees
+ip-10-0-216-33:179940:180194 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:179940:180194 [6] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 00/0 : 0[101c0] -> 8[101c0] [send] via NET/Socket/0
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Channel 01/0 : 0[101c0] -> 8[101c0] [send] via NET/Socket/0
+ip-10-0-194-127:32770:33193 [0] NCCL INFO Connected all trees
+ip-10-0-194-127:32770:33193 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-194-127:32770:33193 [0] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-194-127:32770:33193 [0] NCCL INFO comm 0x134bb460 rank 0 nranks 16 cudaDev 0 busId 101c0 - Init COMPLETE
+ip-10-0-216-33:179940:180194 [6] NCCL INFO comm 0x13382de0 rank 14 nranks 16 cudaDev 6 busId a01c0 - Init COMPLETE
+ip-10-0-194-127:32775:32775 [5] NCCL INFO cudaDriverVersion 12000
+ip-10-0-194-127:32775:32775 [5] NCCL INFO Bootstrap : Using ens32:10.0.194.127<0>
+ip-10-0-194-127:32775:32775 [5] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-194-127:32775:33197 [5] NCCL INFO NET/IB : No device found.
+ip-10-0-194-127:32775:33197 [5] NCCL INFO NET/Socket : Using [0]ens32:10.0.194.127<0>
+ip-10-0-194-127:32775:33197 [5] NCCL INFO Using network Socket
+ip-10-0-194-127:32775:33197 [5] NCCL INFO Setting affinity for GPU 5 to 0fffff00,00000fff,ff000000
+ip-10-0-194-127:32775:33197 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4
+ip-10-0-194-127:32775:33197 [5] NCCL INFO Channel 00/0 : 5[901d0] -> 4[901c0] via P2P/IPC/read
+ip-10-0-194-127:32775:33197 [5] NCCL INFO Channel 01/0 : 5[901d0] -> 4[901c0] via P2P/IPC/read
+ip-10-0-194-127:32775:33197 [5] NCCL INFO Connected all rings
+ip-10-0-194-127:32775:33197 [5] NCCL INFO Channel 00/0 : 5[901d0] -> 6[a01c0] via P2P/IPC/read
+ip-10-0-216-33:179939:179939 [5] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:179939:179939 [5] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:179939:179939 [5] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:179939:180197 [5] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:179939:180197 [5] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:179939:180197 [5] NCCL INFO Using network Socket
+ip-10-0-216-33:179939:180197 [5] NCCL INFO Setting affinity for GPU 5 to 0fffff00,00000fff,ff000000
+ip-10-0-216-33:179939:180197 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12
+ip-10-0-216-33:179939:180197 [5] NCCL INFO Channel 00/0 : 13[901d0] -> 12[901c0] via P2P/IPC/read
+ip-10-0-216-33:179939:180197 [5] NCCL INFO Channel 01/0 : 13[901d0] -> 12[901c0] via P2P/IPC/read
+ip-10-0-216-33:179939:180197 [5] NCCL INFO Connected all rings
+ip-10-0-194-127:32775:33197 [5] NCCL INFO Channel 01/0 : 5[901d0] -> 6[a01c0] via P2P/IPC/read
+ip-10-0-194-127:32775:33197 [5] NCCL INFO Connected all trees
+ip-10-0-194-127:32775:33197 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-194-127:32775:33197 [5] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-194-127:32775:33197 [5] NCCL INFO comm 0x50b28b60 rank 5 nranks 16 cudaDev 5 busId 901d0 - Init COMPLETE
+ip-10-0-216-33:179939:180197 [5] NCCL INFO Channel 00/0 : 13[901d0] -> 14[a01c0] via P2P/IPC/read
+ip-10-0-216-33:179939:180197 [5] NCCL INFO Channel 01/0 : 13[901d0] -> 14[a01c0] via P2P/IPC/read
+ip-10-0-216-33:179939:180197 [5] NCCL INFO Connected all trees
+ip-10-0-216-33:179939:180197 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:179939:180197 [5] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:179939:180197 [5] NCCL INFO comm 0x4fde0b90 rank 13 nranks 16 cudaDev 5 busId 901d0 - Init COMPLETE
+ip-10-0-194-127:32776:32776 [6] NCCL INFO cudaDriverVersion 12000
+ip-10-0-194-127:32776:32776 [6] NCCL INFO Bootstrap : Using ens32:10.0.194.127<0>
+ip-10-0-194-127:32776:32776 [6] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-194-127:32776:33196 [6] NCCL INFO NET/IB : No device found.
+ip-10-0-194-127:32776:33196 [6] NCCL INFO NET/Socket : Using [0]ens32:10.0.194.127<0>
+ip-10-0-194-127:32776:33196 [6] NCCL INFO Using network Socket
+ip-10-0-194-127:32776:33196 [6] NCCL INFO Setting affinity for GPU 6 to 0fffff00,00000fff,ff000000
+ip-10-0-194-127:32776:33196 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5
+ip-10-0-194-127:32776:33196 [6] NCCL INFO Channel 00/0 : 6[a01c0] -> 5[901d0] via P2P/IPC/read
+ip-10-0-194-127:32776:33196 [6] NCCL INFO Channel 01/0 : 6[a01c0] -> 5[901d0] via P2P/IPC/read
+ip-10-0-194-127:32776:33196 [6] NCCL INFO Connected all rings
+ip-10-0-194-127:32776:33196 [6] NCCL INFO Channel 00/0 : 6[a01c0] -> 7[a01d0] via P2P/IPC/read
+ip-10-0-194-127:32776:33196 [6] NCCL INFO Channel 01/0 : 6[a01c0] -> 7[a01d0] via P2P/IPC/read
+ip-10-0-194-127:32776:33196 [6] NCCL INFO Connected all trees
+ip-10-0-194-127:32776:33196 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-194-127:32776:33196 [6] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-194-127:32776:33196 [6] NCCL INFO comm 0xa2f8460 rank 6 nranks 16 cudaDev 6 busId a01c0 - Init COMPLETE
+ip-10-0-194-127:32777:32777 [7] NCCL INFO cudaDriverVersion 12000
+ip-10-0-194-127:32777:32777 [7] NCCL INFO Bootstrap : Using ens32:10.0.194.127<0>
+ip-10-0-194-127:32777:32777 [7] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-194-127:32777:33194 [7] NCCL INFO NET/IB : No device found.
+ip-10-0-194-127:32777:33194 [7] NCCL INFO NET/Socket : Using [0]ens32:10.0.194.127<0>
+ip-10-0-194-127:32777:33194 [7] NCCL INFO Using network Socket
+ip-10-0-194-127:32777:33194 [7] NCCL INFO Setting affinity for GPU 7 to 0fffff00,00000fff,ff000000
+ip-10-0-194-127:32777:33194 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6
+ip-10-0-194-127:32777:33194 [7] NCCL INFO Channel 00/0 : 7[a01d0] -> 6[a01c0] via P2P/IPC/read
+ip-10-0-194-127:32777:33194 [7] NCCL INFO Channel 01/0 : 7[a01d0] -> 6[a01c0] via P2P/IPC/read
+ip-10-0-194-127:32777:33194 [7] NCCL INFO Connected all rings
+ip-10-0-194-127:32777:33194 [7] NCCL INFO Connected all trees
+ip-10-0-194-127:32777:33194 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-194-127:32777:33194 [7] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-194-127:32777:33194 [7] NCCL INFO comm 0x12908900 rank 7 nranks 16 cudaDev 7 busId a01d0 - Init COMPLETE
+ip-10-0-194-127:32772:32772 [2] NCCL INFO cudaDriverVersion 12000
+ip-10-0-194-127:32772:32772 [2] NCCL INFO Bootstrap : Using ens32:10.0.194.127<0>
+ip-10-0-194-127:32772:32772 [2] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-194-127:32772:33198 [2] NCCL INFO NET/IB : No device found.
+ip-10-0-194-127:32772:33198 [2] NCCL INFO NET/Socket : Using [0]ens32:10.0.194.127<0>
+ip-10-0-194-127:32772:33198 [2] NCCL INFO Using network Socket
+ip-10-0-194-127:32772:33198 [2] NCCL INFO Setting affinity for GPU 2 to 0f,ffff0000,000fffff
+ip-10-0-194-127:32772:33198 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1
+ip-10-0-194-127:32772:33198 [2] NCCL INFO Channel 00/0 : 2[201c0] -> 1[101d0] via P2P/IPC/read
+ip-10-0-194-127:32772:33198 [2] NCCL INFO Channel 01/0 : 2[201c0] -> 1[101d0] via P2P/IPC/read
+ip-10-0-194-127:32772:33198 [2] NCCL INFO Connected all rings
+ip-10-0-194-127:32772:33198 [2] NCCL INFO Channel 00/0 : 2[201c0] -> 3[201d0] via P2P/IPC/read
+ip-10-0-194-127:32772:33198 [2] NCCL INFO Channel 01/0 : 2[201c0] -> 3[201d0] via P2P/IPC/read
+ip-10-0-194-127:32772:33198 [2] NCCL INFO Connected all trees
+ip-10-0-194-127:32772:33198 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-194-127:32772:33198 [2] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-194-127:32772:33198 [2] NCCL INFO comm 0x4f0df330 rank 2 nranks 16 cudaDev 2 busId 201c0 - Init COMPLETE
+[ip-10-0-194-127:0/16] 2024-02-26 17:49:13,954 (distributed:1027) INFO: Reducer buckets have been rebuilt in this iteration.
+[ip-10-0-194-127:0/16] 2024-02-26 17:50:42,168 (trainer:762) INFO: 1epoch:train:1-100batch: iter_time=2.901, forward_time=0.228, loss_ctc=7.111e+03, loss_att=378.837, acc=1.735e-05, loss=2.398e+03, backward_time=0.226, grad_norm=1.351e+04, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.047, optim0_lr0=8.583e-08, train_time=4.108
+[ip-10-0-194-127:0/16] 2024-02-26 17:52:09,125 (trainer:762) INFO: 1epoch:train:101-200batch: iter_time=1.287e-04, forward_time=0.211, loss_ctc=3.710e+03, loss_att=386.323, acc=7.526e-06, loss=1.384e+03, backward_time=0.237, grad_norm=1.736e+04, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.047, optim0_lr0=2.525e-07, train_time=0.870
+[ip-10-0-194-127:0/16] 2024-02-26 17:53:36,961 (trainer:762) INFO: 1epoch:train:201-300batch: iter_time=1.209e-04, forward_time=0.208, loss_ctc=586.929, loss_att=393.938, acc=1.931e-05, loss=451.835, backward_time=0.228, grad_norm=2.394e+03, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=4.192e-07, train_time=0.878
+[ip-10-0-194-127:0/16] 2024-02-26 17:55:05,579 (trainer:762) INFO: 1epoch:train:301-400batch: iter_time=1.247e-04, forward_time=0.208, loss_ctc=382.183, loss_att=374.263, acc=1.407e-05, loss=376.639, backward_time=0.227, grad_norm=239.076, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=5.858e-07, train_time=0.886
+[ip-10-0-194-127:0/16] 2024-02-26 17:56:33,765 (trainer:762) INFO: 1epoch:train:401-500batch: iter_time=1.331e-04, forward_time=0.208, loss_ctc=398.318, loss_att=396.527, acc=1.945e-05, loss=397.064, backward_time=0.227, grad_norm=223.174, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=7.525e-07, train_time=0.882
+[ip-10-0-194-127:0/16] 2024-02-26 17:58:01,106 (trainer:762) INFO: 1epoch:train:501-600batch: iter_time=1.193e-04, forward_time=0.208, loss_ctc=369.990, loss_att=375.082, acc=4.130e-05, loss=373.555, backward_time=0.228, grad_norm=185.619, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=9.192e-07, train_time=0.873
+[ip-10-0-194-127:0/16] 2024-02-26 17:59:28,153 (trainer:762) INFO: 1epoch:train:601-700batch: iter_time=1.278e-04, forward_time=0.212, loss_ctc=397.123, loss_att=404.329, acc=3.762e-05, loss=402.167, backward_time=0.232, grad_norm=184.082, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=1.086e-06, train_time=0.870
+[ip-10-0-194-127:0/16] 2024-02-26 18:00:54,078 (trainer:762) INFO: 1epoch:train:701-800batch: iter_time=1.385e-04, forward_time=0.206, loss_ctc=363.640, loss_att=374.369, acc=2.057e-04, loss=371.150, backward_time=0.225, grad_norm=156.166, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=1.252e-06, train_time=0.859
+[ip-10-0-194-127:0/16] 2024-02-26 18:02:19,837 (trainer:762) INFO: 1epoch:train:801-900batch: iter_time=1.234e-04, forward_time=0.205, loss_ctc=331.091, loss_att=346.754, acc=0.003, loss=342.055, backward_time=0.225, grad_norm=119.672, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=1.419e-06, train_time=0.857
+[ip-10-0-194-127:0/16] 2024-02-26 18:03:46,623 (trainer:762) INFO: 1epoch:train:901-1000batch: iter_time=1.240e-04, forward_time=0.206, loss_ctc=351.395, loss_att=365.098, acc=0.012, loss=360.987, backward_time=0.227, grad_norm=164.205, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=1.586e-06, train_time=0.868
+[ip-10-0-194-127:0/16] 2024-02-26 18:05:14,082 (trainer:762) INFO: 1epoch:train:1001-1100batch: iter_time=1.227e-04, forward_time=0.206, loss_ctc=332.785, loss_att=348.034, acc=0.033, loss=343.459, backward_time=0.223, grad_norm=149.036, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=1.752e-06, train_time=0.874
+[ip-10-0-194-127:0/16] 2024-02-26 18:06:41,517 (trainer:762) INFO: 1epoch:train:1101-1200batch: iter_time=1.259e-04, forward_time=0.208, loss_ctc=357.640, loss_att=373.473, acc=0.057, loss=368.723, backward_time=0.224, grad_norm=154.917, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=1.919e-06, train_time=0.874
+[ip-10-0-194-127:0/16] 2024-02-26 18:08:08,555 (trainer:762) INFO: 1epoch:train:1201-1300batch: iter_time=1.268e-04, forward_time=0.208, loss_ctc=415.967, loss_att=427.118, acc=0.062, loss=423.773, backward_time=0.224, grad_norm=184.073, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=2.086e-06, train_time=0.870
+[ip-10-0-194-127:0/16] 2024-02-26 18:09:35,413 (trainer:762) INFO: 1epoch:train:1301-1400batch: iter_time=1.230e-04, forward_time=0.211, loss_ctc=359.397, loss_att=372.169, acc=0.072, loss=368.338, backward_time=0.224, grad_norm=152.314, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=2.253e-06, train_time=0.868
+[ip-10-0-194-127:0/16] 2024-02-26 18:11:01,145 (trainer:762) INFO: 1epoch:train:1401-1500batch: iter_time=1.312e-04, forward_time=0.206, loss_ctc=330.382, loss_att=341.968, acc=0.076, loss=338.492, backward_time=0.227, grad_norm=111.303, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=2.419e-06, train_time=0.857
+[ip-10-0-194-127:0/16] 2024-02-26 18:12:27,341 (trainer:762) INFO: 1epoch:train:1501-1600batch: iter_time=1.321e-04, forward_time=0.205, loss_ctc=324.784, loss_att=334.818, acc=0.077, loss=331.808, backward_time=0.233, grad_norm=124.252, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=2.586e-06, train_time=0.862
+[ip-10-0-194-127:0/16] 2024-02-26 18:13:54,065 (trainer:762) INFO: 1epoch:train:1601-1700batch: iter_time=1.206e-04, forward_time=0.207, loss_ctc=354.480, loss_att=358.645, acc=0.073, loss=357.395, backward_time=0.227, grad_norm=136.015, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=2.752e-06, train_time=0.867
+[ip-10-0-194-127:0/16] 2024-02-26 18:15:21,439 (trainer:762) INFO: 1epoch:train:1701-1800batch: iter_time=1.213e-04, forward_time=0.207, loss_ctc=359.639, loss_att=359.558, acc=0.072, loss=359.582, backward_time=0.226, grad_norm=130.160, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=2.919e-06, train_time=0.874
+[ip-10-0-194-127:0/16] 2024-02-26 18:16:48,534 (trainer:762) INFO: 1epoch:train:1801-1900batch: iter_time=1.276e-04, forward_time=0.207, loss_ctc=367.130, loss_att=361.095, acc=0.071, loss=362.906, backward_time=0.227, grad_norm=134.048, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=3.086e-06, train_time=0.871
+[ip-10-0-194-127:0/16] 2024-02-26 18:18:15,342 (trainer:762) INFO: 1epoch:train:1901-2000batch: iter_time=1.238e-04, forward_time=0.206, loss_ctc=337.273, loss_att=331.043, acc=0.078, loss=332.912, backward_time=0.227, grad_norm=106.037, clip=100.000, loss_scale=6.554e+04, optim_step_time=0.046, optim0_lr0=3.253e-06, train_time=0.868
+[ip-10-0-194-127:0/16] 2024-02-26 18:19:42,745 (trainer:762) INFO: 1epoch:train:2001-2100batch: iter_time=1.203e-04, forward_time=0.210, loss_ctc=297.283, loss_att=292.412, acc=0.082, loss=293.874, backward_time=0.224, grad_norm=90.566, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.047, optim0_lr0=3.419e-06, train_time=0.874
+[ip-10-0-194-127:0/16] 2024-02-26 18:21:11,014 (trainer:762) INFO: 1epoch:train:2101-2200batch: iter_time=1.266e-04, forward_time=0.206, loss_ctc=325.531, loss_att=315.570, acc=0.077, loss=318.559, backward_time=0.229, grad_norm=117.815, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=3.586e-06, train_time=0.882
+[ip-10-0-194-127:0/16] 2024-02-26 18:22:37,790 (trainer:762) INFO: 1epoch:train:2201-2300batch: iter_time=1.205e-04, forward_time=0.206, loss_ctc=305.955, loss_att=296.228, acc=0.082, loss=299.146, backward_time=0.228, grad_norm=95.359, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=3.753e-06, train_time=0.868
+[ip-10-0-194-127:0/16] 2024-02-26 18:24:05,639 (trainer:762) INFO: 1epoch:train:2301-2400batch: iter_time=1.239e-04, forward_time=0.206, loss_ctc=332.544, loss_att=319.134, acc=0.074, loss=323.157, backward_time=0.235, grad_norm=113.508, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=3.919e-06, train_time=0.878
+[ip-10-0-194-127:0/16] 2024-02-26 18:25:34,120 (trainer:762) INFO: 1epoch:train:2401-2500batch: iter_time=1.273e-04, forward_time=0.205, loss_ctc=296.780, loss_att=286.950, acc=0.082, loss=289.899, backward_time=0.234, grad_norm=93.266, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.047, optim0_lr0=4.086e-06, train_time=0.885
+[ip-10-0-194-127:0/16] 2024-02-26 18:27:01,815 (trainer:762) INFO: 1epoch:train:2501-2600batch: iter_time=1.256e-04, forward_time=0.205, loss_ctc=264.585, loss_att=256.608, acc=0.091, loss=259.001, backward_time=0.240, grad_norm=83.685, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.047, optim0_lr0=4.253e-06, train_time=0.877
+[ip-10-0-194-127:0/16] 2024-02-26 18:28:28,842 (trainer:762) INFO: 1epoch:train:2601-2700batch: iter_time=1.339e-04, forward_time=0.206, loss_ctc=324.382, loss_att=309.150, acc=0.081, loss=313.720, backward_time=0.227, grad_norm=94.710, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=4.419e-06, train_time=0.870
+[ip-10-0-194-127:0/16] 2024-02-26 18:29:56,912 (trainer:762) INFO: 1epoch:train:2701-2800batch: iter_time=1.178e-04, forward_time=0.211, loss_ctc=319.564, loss_att=303.544, acc=0.084, loss=308.350, backward_time=0.232, grad_norm=95.382, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=4.586e-06, train_time=0.880
+[ip-10-0-194-127:0/16] 2024-02-26 18:31:25,676 (trainer:762) INFO: 1epoch:train:2801-2900batch: iter_time=1.259e-04, forward_time=0.206, loss_ctc=297.661, loss_att=283.667, acc=0.094, loss=287.865, backward_time=0.225, grad_norm=82.929, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=4.752e-06, train_time=0.887
+[ip-10-0-194-127:0/16] 2024-02-26 18:32:54,602 (trainer:762) INFO: 1epoch:train:2901-3000batch: iter_time=1.191e-04, forward_time=0.205, loss_ctc=308.907, loss_att=292.756, acc=0.095, loss=297.602, backward_time=0.227, grad_norm=84.556, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=4.919e-06, train_time=0.889
+[ip-10-0-194-127:0/16] 2024-02-26 18:34:22,742 (trainer:762) INFO: 1epoch:train:3001-3100batch: iter_time=1.262e-04, forward_time=0.204, loss_ctc=309.807, loss_att=293.250, acc=0.100, loss=298.217, backward_time=0.228, grad_norm=88.738, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=5.086e-06, train_time=0.881
+[ip-10-0-194-127:0/16] 2024-02-26 18:35:50,469 (trainer:762) INFO: 1epoch:train:3101-3200batch: iter_time=1.191e-04, forward_time=0.207, loss_ctc=290.061, loss_att=275.076, acc=0.111, loss=279.572, backward_time=0.228, grad_norm=78.017, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=5.252e-06, train_time=0.877
+[ip-10-0-194-127:0/16] 2024-02-26 18:37:16,944 (trainer:762) INFO: 1epoch:train:3201-3300batch: iter_time=1.259e-04, forward_time=0.207, loss_ctc=312.624, loss_att=294.140, acc=0.110, loss=299.686, backward_time=0.235, grad_norm=84.888, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.047, optim0_lr0=5.419e-06, train_time=0.865
+[ip-10-0-194-127:0/16] 2024-02-26 18:38:43,514 (trainer:762) INFO: 1epoch:train:3301-3400batch: iter_time=1.228e-04, forward_time=0.207, loss_ctc=319.263, loss_att=299.785, acc=0.108, loss=305.629, backward_time=0.230, grad_norm=94.943, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=5.586e-06, train_time=0.865
+[ip-10-0-194-127:0/16] 2024-02-26 18:40:10,639 (trainer:762) INFO: 1epoch:train:3401-3500batch: iter_time=1.229e-04, forward_time=0.210, loss_ctc=299.034, loss_att=281.820, acc=0.111, loss=286.984, backward_time=0.233, grad_norm=84.916, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=5.753e-06, train_time=0.871
+[ip-10-0-194-127:0/16] 2024-02-26 18:41:37,678 (trainer:762) INFO: 1epoch:train:3501-3600batch: iter_time=1.193e-04, forward_time=0.206, loss_ctc=279.221, loss_att=264.745, acc=0.119, loss=269.088, backward_time=0.233, grad_norm=67.783, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=5.919e-06, train_time=0.870
+[ip-10-0-194-127:0/16] 2024-02-26 18:43:04,840 (trainer:762) INFO: 1epoch:train:3601-3700batch: iter_time=1.193e-04, forward_time=0.205, loss_ctc=268.629, loss_att=254.368, acc=0.118, loss=258.646, backward_time=0.227, grad_norm=72.575, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=6.086e-06, train_time=0.871
+[ip-10-0-194-127:0/16] 2024-02-26 18:44:32,750 (trainer:762) INFO: 1epoch:train:3701-3800batch: iter_time=1.209e-04, forward_time=0.205, loss_ctc=311.056, loss_att=290.889, acc=0.109, loss=296.939, backward_time=0.229, grad_norm=93.179, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=6.253e-06, train_time=0.879
+[ip-10-0-194-127:0/16] 2024-02-26 18:46:01,515 (trainer:762) INFO: 1epoch:train:3801-3900batch: iter_time=1.213e-04, forward_time=0.204, loss_ctc=267.265, loss_att=252.796, acc=0.118, loss=257.137, backward_time=0.226, grad_norm=70.675, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=6.419e-06, train_time=0.887
+[ip-10-0-194-127:0/16] 2024-02-26 18:47:30,577 (trainer:762) INFO: 1epoch:train:3901-4000batch: iter_time=1.223e-04, forward_time=0.204, loss_ctc=273.172, loss_att=256.213, acc=0.119, loss=261.301, backward_time=0.231, grad_norm=109.252, clip=100.000, loss_scale=1.311e+05, optim_step_time=0.046, optim0_lr0=6.586e-06, train_time=0.890
+[ip-10-0-194-127:0/16] 2024-02-26 18:48:58,271 (trainer:762) INFO: 1epoch:train:4001-4100batch: iter_time=1.195e-04, forward_time=0.206, loss_ctc=278.346, loss_att=261.453, acc=0.119, loss=266.521, backward_time=0.232, grad_norm=95.216, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.046, optim0_lr0=6.752e-06, train_time=0.877
+[ip-10-0-194-127:0/16] 2024-02-26 18:50:25,700 (trainer:762) INFO: 1epoch:train:4101-4200batch: iter_time=1.285e-04, forward_time=0.204, loss_ctc=242.842, loss_att=230.135, acc=0.126, loss=233.947, backward_time=0.227, grad_norm=70.129, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.046, optim0_lr0=6.919e-06, train_time=0.874
+[ip-10-0-194-127:0/16] 2024-02-26 18:51:52,896 (trainer:762) INFO: 1epoch:train:4201-4300batch: iter_time=1.219e-04, forward_time=0.211, loss_ctc=294.754, loss_att=274.524, acc=0.115, loss=280.593, backward_time=0.229, grad_norm=89.119, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.046, optim0_lr0=7.086e-06, train_time=0.872
+[ip-10-0-194-127:0/16] 2024-02-26 18:53:19,807 (trainer:762) INFO: 1epoch:train:4301-4400batch: iter_time=1.294e-04, forward_time=0.206, loss_ctc=264.005, loss_att=246.970, acc=0.122, loss=252.080, backward_time=0.233, grad_norm=86.415, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.046, optim0_lr0=7.253e-06, train_time=0.869
+[ip-10-0-194-127:0/16] 2024-02-26 18:54:47,139 (trainer:762) INFO: 1epoch:train:4401-4500batch: iter_time=1.276e-04, forward_time=0.206, loss_ctc=282.706, loss_att=264.177, acc=0.125, loss=269.736, backward_time=0.230, grad_norm=80.949, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.046, optim0_lr0=7.419e-06, train_time=0.873
+[ip-10-0-194-127:0/16] 2024-02-26 18:56:16,184 (trainer:762) INFO: 1epoch:train:4501-4600batch: iter_time=1.305e-04, forward_time=0.205, loss_ctc=280.986, loss_att=261.920, acc=0.121, loss=267.640, backward_time=0.225, grad_norm=75.555, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.046, optim0_lr0=7.586e-06, train_time=0.890
+[ip-10-0-194-127:0/16] 2024-02-26 18:57:47,096 (trainer:762) INFO: 1epoch:train:4601-4700batch: iter_time=1.290e-04, forward_time=0.204, loss_ctc=269.278, loss_att=249.948, acc=0.129, loss=255.747, backward_time=0.222, grad_norm=86.582, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.046, optim0_lr0=7.753e-06, train_time=0.909
+[ip-10-0-194-127:0/16] 2024-02-26 18:59:16,373 (trainer:762) INFO: 1epoch:train:4701-4800batch: iter_time=1.262e-04, forward_time=0.206, loss_ctc=301.689, loss_att=278.379, acc=0.120, loss=285.372, backward_time=0.230, grad_norm=88.262, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.046, optim0_lr0=7.919e-06, train_time=0.893
+[ip-10-0-194-127:0/16] 2024-02-26 19:00:44,796 (trainer:762) INFO: 1epoch:train:4801-4900batch: iter_time=1.229e-04, forward_time=0.205, loss_ctc=301.211, loss_att=277.859, acc=0.116, loss=284.864, backward_time=0.227, grad_norm=85.124, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.046, optim0_lr0=8.086e-06, train_time=0.884
+[ip-10-0-194-127:0/16] 2024-02-26 19:02:12,615 (trainer:762) INFO: 1epoch:train:4901-5000batch: iter_time=1.187e-04, forward_time=0.205, loss_ctc=286.504, loss_att=264.270, acc=0.120, loss=270.940, backward_time=0.224, grad_norm=98.913, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.046, optim0_lr0=8.252e-06, train_time=0.878
+[ip-10-0-194-127:0/16] 2024-02-26 19:02:16,370 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-194-127:0/16] 2024-02-26 19:02:53,542 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-194-127:0/16] 2024-02-26 19:03:02,248 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7fdf5f6e2e80>)
+[ip-10-0-194-127:0/16] 2024-02-26 19:03:02,248 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-194-127:0/16] 2024-02-26 19:03:02,252 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-194-127:0/16] 2024-02-26 19:09:11,112 (trainer:762) INFO: 1epoch:train:5001-5100batch: iter_time=3.167, forward_time=0.208, loss_ctc=277.451, loss_att=256.593, acc=0.123, loss=262.850, backward_time=0.225, grad_norm=76.591, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.047, optim0_lr0=8.419e-06, train_time=4.185
+[ip-10-0-194-127:0/16] 2024-02-26 19:10:39,833 (trainer:762) INFO: 1epoch:train:5101-5200batch: iter_time=1.317e-04, forward_time=0.207, loss_ctc=276.279, loss_att=253.885, acc=0.128, loss=260.603, backward_time=0.227, grad_norm=73.545, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.047, optim0_lr0=8.586e-06, train_time=0.887
+[ip-10-0-194-127:0/16] 2024-02-26 19:12:08,672 (trainer:762) INFO: 1epoch:train:5201-5300batch: iter_time=1.207e-04, forward_time=0.205, loss_ctc=297.919, loss_att=272.415, acc=0.117, loss=280.067, backward_time=0.229, grad_norm=135.636, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.047, optim0_lr0=8.753e-06, train_time=0.888
+[ip-10-0-194-127:0/16] 2024-02-26 19:13:36,929 (trainer:762) INFO: 1epoch:train:5301-5400batch: iter_time=1.296e-04, forward_time=0.206, loss_ctc=277.984, loss_att=254.871, acc=0.127, loss=261.805, backward_time=0.226, grad_norm=97.575, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.047, optim0_lr0=8.919e-06, train_time=0.882
+[ip-10-0-194-127:0/16] 2024-02-26 19:15:05,821 (trainer:762) INFO: 1epoch:train:5401-5500batch: iter_time=1.266e-04, forward_time=0.211, loss_ctc=279.957, loss_att=256.448, acc=0.125, loss=263.501, backward_time=0.228, grad_norm=69.107, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.047, optim0_lr0=9.086e-06, train_time=0.889
+[ip-10-0-194-127:0/16] 2024-02-26 19:16:33,731 (trainer:762) INFO: 1epoch:train:5501-5600batch: iter_time=1.264e-04, forward_time=0.206, loss_ctc=272.546, loss_att=248.638, acc=0.129, loss=255.810, backward_time=0.237, grad_norm=93.766, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.047, optim0_lr0=9.252e-06, train_time=0.879
+[ip-10-0-194-127:0/16] 2024-02-26 19:18:00,284 (trainer:762) INFO: 1epoch:train:5601-5700batch: iter_time=1.214e-04, forward_time=0.208, loss_ctc=293.381, loss_att=267.041, acc=0.125, loss=274.943, backward_time=0.227, grad_norm=75.392, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.047, optim0_lr0=9.419e-06, train_time=0.865
+[ip-10-0-194-127:0/16] 2024-02-26 19:19:27,773 (trainer:762) INFO: 1epoch:train:5701-5800batch: iter_time=1.215e-04, forward_time=0.207, loss_ctc=285.605, loss_att=259.623, acc=0.130, loss=267.418, backward_time=0.229, grad_norm=94.575, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.047, optim0_lr0=9.586e-06, train_time=0.875
+[ip-10-0-194-127:0/16] 2024-02-26 19:20:53,985 (trainer:762) INFO: 1epoch:train:5801-5900batch: iter_time=1.238e-04, forward_time=0.206, loss_ctc=248.278, loss_att=226.111, acc=0.141, loss=232.761, backward_time=0.227, grad_norm=78.659, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.047, optim0_lr0=9.753e-06, train_time=0.862
+[ip-10-0-194-127:0/16] 2024-02-26 19:22:20,153 (trainer:762) INFO: 1epoch:train:5901-6000batch: iter_time=1.221e-04, forward_time=0.206, loss_ctc=277.947, loss_att=251.078, acc=0.131, loss=259.139, backward_time=0.230, grad_norm=106.103, clip=100.000, loss_scale=2.621e+05, optim_step_time=0.047, optim0_lr0=9.919e-06, train_time=0.861
+[ip-10-0-194-127:0/16] 2024-02-26 19:23:46,531 (trainer:762) INFO: 1epoch:train:6001-6100batch: iter_time=1.218e-04, forward_time=0.206, loss_ctc=261.811, loss_att=237.940, acc=0.135, loss=245.101, backward_time=0.235, grad_norm=95.022, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.009e-05, train_time=0.864
+[ip-10-0-194-127:0/16] 2024-02-26 19:25:12,885 (trainer:762) INFO: 1epoch:train:6101-6200batch: iter_time=1.249e-04, forward_time=0.208, loss_ctc=272.295, loss_att=246.242, acc=0.134, loss=254.058, backward_time=0.231, grad_norm=64.117, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.025e-05, train_time=0.863
+[ip-10-0-194-127:0/16] 2024-02-26 19:26:41,467 (trainer:762) INFO: 1epoch:train:6201-6300batch: iter_time=1.198e-04, forward_time=0.213, loss_ctc=334.577, loss_att=298.030, acc=0.125, loss=308.994, backward_time=0.234, grad_norm=92.297, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.042e-05, train_time=0.886
+[ip-10-0-194-127:0/16] 2024-02-26 19:28:09,431 (trainer:762) INFO: 1epoch:train:6301-6400batch: iter_time=1.258e-04, forward_time=0.208, loss_ctc=287.127, loss_att=257.654, acc=0.133, loss=266.496, backward_time=0.231, grad_norm=88.839, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.059e-05, train_time=0.879
+[ip-10-0-194-127:0/16] 2024-02-26 19:29:36,002 (trainer:762) INFO: 1epoch:train:6401-6500batch: iter_time=1.205e-04, forward_time=0.207, loss_ctc=273.679, loss_att=244.497, acc=0.135, loss=253.251, backward_time=0.225, grad_norm=98.685, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.075e-05, train_time=0.865
+[ip-10-0-194-127:0/16] 2024-02-26 19:31:05,591 (trainer:762) INFO: 1epoch:train:6501-6600batch: iter_time=1.226e-04, forward_time=0.206, loss_ctc=263.604, loss_att=236.343, acc=0.142, loss=244.521, backward_time=0.228, grad_norm=100.158, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.092e-05, train_time=0.896
+[ip-10-0-194-127:0/16] 2024-02-26 19:32:35,318 (trainer:762) INFO: 1epoch:train:6601-6700batch: iter_time=1.196e-04, forward_time=0.207, loss_ctc=291.770, loss_att=258.547, acc=0.137, loss=268.514, backward_time=0.228, grad_norm=104.222, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.109e-05, train_time=0.897
+[ip-10-0-194-127:0/16] 2024-02-26 19:34:02,613 (trainer:762) INFO: 1epoch:train:6701-6800batch: iter_time=1.188e-04, forward_time=0.208, loss_ctc=303.046, loss_att=268.186, acc=0.139, loss=278.644, backward_time=0.230, grad_norm=77.458, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.125e-05, train_time=0.873
+[ip-10-0-194-127:0/16] 2024-02-26 19:35:29,671 (trainer:762) INFO: 1epoch:train:6801-6900batch: iter_time=1.213e-04, forward_time=0.208, loss_ctc=317.919, loss_att=278.917, acc=0.134, loss=290.617, backward_time=0.230, grad_norm=94.846, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.142e-05, train_time=0.870
+[ip-10-0-194-127:0/16] 2024-02-26 19:36:58,003 (trainer:762) INFO: 1epoch:train:6901-7000batch: iter_time=1.214e-04, forward_time=0.207, loss_ctc=296.070, loss_att=259.075, acc=0.143, loss=270.173, backward_time=0.227, grad_norm=97.500, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.159e-05, train_time=0.883
+[ip-10-0-194-127:0/16] 2024-02-26 19:38:25,431 (trainer:762) INFO: 1epoch:train:7001-7100batch: iter_time=1.193e-04, forward_time=0.212, loss_ctc=260.655, loss_att=231.750, acc=0.147, loss=240.422, backward_time=0.227, grad_norm=92.821, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.175e-05, train_time=0.874
+[ip-10-0-194-127:0/16] 2024-02-26 19:39:52,531 (trainer:762) INFO: 1epoch:train:7101-7200batch: iter_time=1.297e-04, forward_time=0.208, loss_ctc=289.502, loss_att=253.895, acc=0.142, loss=264.577, backward_time=0.229, grad_norm=78.297, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.192e-05, train_time=0.871
+[ip-10-0-194-127:0/16] 2024-02-26 19:41:19,293 (trainer:762) INFO: 1epoch:train:7201-7300batch: iter_time=1.193e-04, forward_time=0.208, loss_ctc=275.926, loss_att=242.022, acc=0.144, loss=252.193, backward_time=0.227, grad_norm=73.335, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.209e-05, train_time=0.867
+[ip-10-0-194-127:0/16] 2024-02-26 19:42:46,166 (trainer:762) INFO: 1epoch:train:7301-7400batch: iter_time=1.229e-04, forward_time=0.208, loss_ctc=299.598, loss_att=260.355, acc=0.141, loss=272.128, backward_time=0.231, grad_norm=87.701, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.225e-05, train_time=0.868
+[ip-10-0-194-127:0/16] 2024-02-26 19:44:13,950 (trainer:762) INFO: 1epoch:train:7401-7500batch: iter_time=1.237e-04, forward_time=0.207, loss_ctc=270.860, loss_att=238.578, acc=0.144, loss=248.263, backward_time=0.230, grad_norm=65.354, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.242e-05, train_time=0.878
+[ip-10-0-194-127:0/16] 2024-02-26 19:45:40,538 (trainer:762) INFO: 1epoch:train:7501-7600batch: iter_time=1.185e-04, forward_time=0.207, loss_ctc=243.727, loss_att=215.062, acc=0.154, loss=223.662, backward_time=0.226, grad_norm=74.115, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.259e-05, train_time=0.866
+[ip-10-0-194-127:0/16] 2024-02-26 19:47:08,284 (trainer:762) INFO: 1epoch:train:7601-7700batch: iter_time=1.236e-04, forward_time=0.207, loss_ctc=304.529, loss_att=263.530, acc=0.144, loss=275.829, backward_time=0.228, grad_norm=78.166, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.275e-05, train_time=0.877
+[ip-10-0-194-127:0/16] 2024-02-26 19:48:37,260 (trainer:762) INFO: 1epoch:train:7701-7800batch: iter_time=1.195e-04, forward_time=0.205, loss_ctc=305.792, loss_att=262.582, acc=0.148, loss=275.545, backward_time=0.230, grad_norm=88.186, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.292e-05, train_time=0.890
+[ip-10-0-194-127:0/16] 2024-02-26 19:50:05,805 (trainer:762) INFO: 1epoch:train:7801-7900batch: iter_time=1.264e-04, forward_time=0.205, loss_ctc=281.026, loss_att=243.983, acc=0.148, loss=255.096, backward_time=0.226, grad_norm=83.355, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.309e-05, train_time=0.885
+[ip-10-0-194-127:0/16] 2024-02-26 19:51:34,021 (trainer:762) INFO: 1epoch:train:7901-8000batch: iter_time=1.310e-04, forward_time=0.212, loss_ctc=295.082, loss_att=254.133, acc=0.145, loss=266.417, backward_time=0.235, grad_norm=92.345, clip=100.000, loss_scale=5.243e+05, optim_step_time=0.047, optim0_lr0=1.325e-05, train_time=0.882
+[ip-10-0-194-127:0/16] 2024-02-26 19:53:02,160 (trainer:762) INFO: 1epoch:train:8001-8100batch: iter_time=1.232e-04, forward_time=0.207, loss_ctc=294.867, loss_att=253.557, acc=0.148, loss=265.950, backward_time=0.227, grad_norm=73.774, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.342e-05, train_time=0.881
+[ip-10-0-194-127:0/16] 2024-02-26 19:54:30,100 (trainer:762) INFO: 1epoch:train:8101-8200batch: iter_time=1.216e-04, forward_time=0.207, loss_ctc=278.344, loss_att=239.802, acc=0.154, loss=251.365, backward_time=0.228, grad_norm=63.387, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.359e-05, train_time=0.879
+[ip-10-0-194-127:0/16] 2024-02-26 19:55:59,580 (trainer:762) INFO: 1epoch:train:8201-8300batch: iter_time=1.207e-04, forward_time=0.208, loss_ctc=304.304, loss_att=258.932, acc=0.151, loss=272.544, backward_time=0.226, grad_norm=90.322, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.375e-05, train_time=0.895
+[ip-10-0-194-127:0/16] 2024-02-26 19:57:29,470 (trainer:762) INFO: 1epoch:train:8301-8400batch: iter_time=1.222e-04, forward_time=0.207, loss_ctc=310.901, loss_att=264.509, acc=0.148, loss=278.426, backward_time=0.226, grad_norm=103.182, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.392e-05, train_time=0.899
+[ip-10-0-194-127:0/16] 2024-02-26 19:58:57,890 (trainer:762) INFO: 1epoch:train:8401-8500batch: iter_time=1.213e-04, forward_time=0.207, loss_ctc=292.893, loss_att=249.679, acc=0.152, loss=262.644, backward_time=0.227, grad_norm=90.881, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.409e-05, train_time=0.884
+[ip-10-0-194-127:0/16] 2024-02-26 20:00:26,010 (trainer:762) INFO: 1epoch:train:8501-8600batch: iter_time=1.221e-04, forward_time=0.207, loss_ctc=269.743, loss_att=232.757, acc=0.158, loss=243.853, backward_time=0.230, grad_norm=60.230, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.425e-05, train_time=0.881
+[ip-10-0-194-127:0/16] 2024-02-26 20:01:52,263 (trainer:762) INFO: 1epoch:train:8601-8700batch: iter_time=1.225e-04, forward_time=0.207, loss_ctc=262.547, loss_att=225.659, acc=0.163, loss=236.725, backward_time=0.224, grad_norm=64.720, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.442e-05, train_time=0.862
+[ip-10-0-194-127:0/16] 2024-02-26 20:03:20,399 (trainer:762) INFO: 1epoch:train:8701-8800batch: iter_time=1.311e-04, forward_time=0.208, loss_ctc=305.190, loss_att=260.249, acc=0.148, loss=273.731, backward_time=0.228, grad_norm=66.523, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.459e-05, train_time=0.881
+[ip-10-0-194-127:0/16] 2024-02-26 20:04:47,007 (trainer:762) INFO: 1epoch:train:8801-8900batch: iter_time=1.203e-04, forward_time=0.212, loss_ctc=259.471, loss_att=223.594, acc=0.159, loss=234.357, backward_time=0.227, grad_norm=73.127, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.475e-05, train_time=0.866
+[ip-10-0-194-127:0/16] 2024-02-26 20:06:13,643 (trainer:762) INFO: 1epoch:train:8901-9000batch: iter_time=1.261e-04, forward_time=0.207, loss_ctc=264.430, loss_att=228.027, acc=0.162, loss=238.948, backward_time=0.234, grad_norm=59.111, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.492e-05, train_time=0.866
+[ip-10-0-194-127:0/16] 2024-02-26 20:07:39,912 (trainer:762) INFO: 1epoch:train:9001-9100batch: iter_time=1.262e-04, forward_time=0.208, loss_ctc=269.638, loss_att=232.835, acc=0.160, loss=243.876, backward_time=0.231, grad_norm=61.304, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.509e-05, train_time=0.862
+[ip-10-0-194-127:0/16] 2024-02-26 20:09:07,842 (trainer:762) INFO: 1epoch:train:9101-9200batch: iter_time=1.263e-04, forward_time=0.206, loss_ctc=237.598, loss_att=206.452, acc=0.166, loss=215.796, backward_time=0.232, grad_norm=61.439, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.525e-05, train_time=0.879
+[ip-10-0-194-127:0/16] 2024-02-26 20:10:34,379 (trainer:762) INFO: 1epoch:train:9201-9300batch: iter_time=1.203e-04, forward_time=0.207, loss_ctc=289.052, loss_att=246.029, acc=0.158, loss=258.936, backward_time=0.227, grad_norm=77.423, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.542e-05, train_time=0.865
+[ip-10-0-194-127:0/16] 2024-02-26 20:12:00,684 (trainer:762) INFO: 1epoch:train:9301-9400batch: iter_time=1.206e-04, forward_time=0.207, loss_ctc=257.369, loss_att=220.702, acc=0.163, loss=231.702, backward_time=0.225, grad_norm=83.790, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.559e-05, train_time=0.863
+[ip-10-0-194-127:0/16] 2024-02-26 20:13:28,746 (trainer:762) INFO: 1epoch:train:9401-9500batch: iter_time=1.240e-04, forward_time=0.207, loss_ctc=276.029, loss_att=237.682, acc=0.164, loss=249.186, backward_time=0.229, grad_norm=64.503, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.575e-05, train_time=0.880
+[ip-10-0-194-127:0/16] 2024-02-26 20:14:55,384 (trainer:762) INFO: 1epoch:train:9501-9600batch: iter_time=1.239e-04, forward_time=0.208, loss_ctc=273.993, loss_att=235.115, acc=0.160, loss=246.778, backward_time=0.227, grad_norm=71.500, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.592e-05, train_time=0.866
+[ip-10-0-194-127:0/16] 2024-02-26 20:16:21,901 (trainer:762) INFO: 1epoch:train:9601-9700batch: iter_time=1.219e-04, forward_time=0.207, loss_ctc=264.628, loss_att=226.467, acc=0.166, loss=237.916, backward_time=0.227, grad_norm=71.739, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.609e-05, train_time=0.865
+[ip-10-0-194-127:0/16] 2024-02-26 20:17:49,563 (trainer:762) INFO: 1epoch:train:9701-9800batch: iter_time=1.264e-04, forward_time=0.214, loss_ctc=295.024, loss_att=250.532, acc=0.161, loss=263.880, backward_time=0.235, grad_norm=69.373, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.625e-05, train_time=0.876
+[ip-10-0-194-127:0/16] 2024-02-26 20:19:16,139 (trainer:762) INFO: 1epoch:train:9801-9900batch: iter_time=1.245e-04, forward_time=0.208, loss_ctc=297.650, loss_att=251.969, acc=0.152, loss=265.674, backward_time=0.227, grad_norm=82.459, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.642e-05, train_time=0.866
+[ip-10-0-194-127:0/16] 2024-02-26 20:20:45,152 (trainer:762) INFO: 1epoch:train:9901-10000batch: iter_time=1.248e-04, forward_time=0.206, loss_ctc=280.965, loss_att=238.298, acc=0.161, loss=251.098, backward_time=0.226, grad_norm=69.757, clip=100.000, loss_scale=1.049e+06, optim_step_time=0.047, optim0_lr0=1.659e-05, train_time=0.890
+[ip-10-0-194-127:0/16] 2024-02-26 20:20:48,363 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-194-127:0/16] 2024-02-26 20:21:25,427 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-194-127:0/16] 2024-02-26 20:21:33,440 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7fe3b6689df0>)
+[ip-10-0-194-127:0/16] 2024-02-26 20:21:33,441 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-194-127:0/16] 2024-02-26 20:21:33,445 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-194-127:0/16] 2024-02-26 20:27:35,339 (trainer:762) INFO: 1epoch:train:10001-10100batch: iter_time=3.168, forward_time=0.208, loss_ctc=274.206, loss_att=233.204, acc=0.159, loss=245.505, backward_time=0.226, grad_norm=78.377, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.675e-05, train_time=4.102
+[ip-10-0-194-127:0/16] 2024-02-26 20:29:02,786 (trainer:762) INFO: 1epoch:train:10101-10200batch: iter_time=1.240e-04, forward_time=0.213, loss_ctc=270.821, loss_att=229.759, acc=0.167, loss=242.078, backward_time=0.230, grad_norm=70.798, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.692e-05, train_time=0.874
+[ip-10-0-194-127:0/16] 2024-02-26 20:30:30,312 (trainer:762) INFO: 1epoch:train:10201-10300batch: iter_time=1.189e-04, forward_time=0.208, loss_ctc=294.889, loss_att=248.095, acc=0.153, loss=262.133, backward_time=0.237, grad_norm=82.573, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.709e-05, train_time=0.875
+[ip-10-0-194-127:0/16] 2024-02-26 20:32:00,062 (trainer:762) INFO: 1epoch:train:10301-10400batch: iter_time=1.249e-04, forward_time=0.206, loss_ctc=271.551, loss_att=231.102, acc=0.162, loss=243.237, backward_time=0.233, grad_norm=64.014, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.725e-05, train_time=0.897
+[ip-10-0-194-127:0/16] 2024-02-26 20:33:27,147 (trainer:762) INFO: 1epoch:train:10401-10500batch: iter_time=1.262e-04, forward_time=0.208, loss_ctc=273.743, loss_att=234.023, acc=0.162, loss=245.939, backward_time=0.226, grad_norm=62.096, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.742e-05, train_time=0.871
+[ip-10-0-194-127:0/16] 2024-02-26 20:34:54,694 (trainer:762) INFO: 1epoch:train:10501-10600batch: iter_time=1.212e-04, forward_time=0.208, loss_ctc=265.311, loss_att=226.444, acc=0.166, loss=238.104, backward_time=0.230, grad_norm=72.303, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.759e-05, train_time=0.875
+[ip-10-0-194-127:0/16] 2024-02-26 20:36:22,632 (trainer:762) INFO: 1epoch:train:10601-10700batch: iter_time=1.242e-04, forward_time=0.208, loss_ctc=287.062, loss_att=244.508, acc=0.159, loss=257.274, backward_time=0.231, grad_norm=66.010, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.775e-05, train_time=0.879
+[ip-10-0-194-127:0/16] 2024-02-26 20:37:51,353 (trainer:762) INFO: 1epoch:train:10701-10800batch: iter_time=1.278e-04, forward_time=0.207, loss_ctc=279.507, loss_att=234.790, acc=0.165, loss=248.205, backward_time=0.230, grad_norm=69.539, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.792e-05, train_time=0.887
+[ip-10-0-194-127:0/16] 2024-02-26 20:39:19,738 (trainer:762) INFO: 1epoch:train:10801-10900batch: iter_time=1.256e-04, forward_time=0.206, loss_ctc=242.369, loss_att=206.944, acc=0.177, loss=217.572, backward_time=0.227, grad_norm=52.606, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.809e-05, train_time=0.884
+[ip-10-0-194-127:0/16] 2024-02-26 20:40:48,838 (trainer:762) INFO: 1epoch:train:10901-11000batch: iter_time=1.228e-04, forward_time=0.206, loss_ctc=269.985, loss_att=229.851, acc=0.166, loss=241.891, backward_time=0.226, grad_norm=76.759, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.825e-05, train_time=0.891
+[ip-10-0-194-127:0/16] 2024-02-26 20:42:16,178 (trainer:762) INFO: 1epoch:train:11001-11100batch: iter_time=1.253e-04, forward_time=0.206, loss_ctc=258.516, loss_att=219.082, acc=0.165, loss=230.912, backward_time=0.224, grad_norm=73.145, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.842e-05, train_time=0.873
+[ip-10-0-194-127:0/16] 2024-02-26 20:43:45,206 (trainer:762) INFO: 1epoch:train:11101-11200batch: iter_time=1.218e-04, forward_time=0.212, loss_ctc=265.786, loss_att=227.889, acc=0.162, loss=239.258, backward_time=0.227, grad_norm=53.494, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.859e-05, train_time=0.890
+[ip-10-0-194-127:0/16] 2024-02-26 20:45:13,786 (trainer:762) INFO: 1epoch:train:11201-11300batch: iter_time=1.222e-04, forward_time=0.209, loss_ctc=328.097, loss_att=271.344, acc=0.153, loss=288.370, backward_time=0.228, grad_norm=71.839, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.875e-05, train_time=0.886
+[ip-10-0-194-127:0/16] 2024-02-26 20:46:42,127 (trainer:762) INFO: 1epoch:train:11301-11400batch: iter_time=1.214e-04, forward_time=0.206, loss_ctc=281.339, loss_att=236.820, acc=0.157, loss=250.176, backward_time=0.225, grad_norm=63.493, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.892e-05, train_time=0.883
+[ip-10-0-194-127:0/16] 2024-02-26 20:48:10,670 (trainer:762) INFO: 1epoch:train:11401-11500batch: iter_time=1.208e-04, forward_time=0.207, loss_ctc=266.062, loss_att=224.188, acc=0.163, loss=236.750, backward_time=0.225, grad_norm=74.829, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.909e-05, train_time=0.885
+[ip-10-0-194-127:0/16] 2024-02-26 20:49:39,328 (trainer:762) INFO: 1epoch:train:11501-11600batch: iter_time=1.218e-04, forward_time=0.208, loss_ctc=257.506, loss_att=220.133, acc=0.168, loss=231.345, backward_time=0.227, grad_norm=67.180, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.925e-05, train_time=0.886
+[ip-10-0-194-127:0/16] 2024-02-26 20:51:09,208 (trainer:762) INFO: 1epoch:train:11601-11700batch: iter_time=1.204e-04, forward_time=0.205, loss_ctc=286.241, loss_att=241.440, acc=0.161, loss=254.880, backward_time=0.229, grad_norm=85.449, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.942e-05, train_time=0.899
+[ip-10-0-194-127:0/16] 2024-02-26 20:52:38,609 (trainer:762) INFO: 1epoch:train:11701-11800batch: iter_time=1.221e-04, forward_time=0.207, loss_ctc=296.663, loss_att=249.582, acc=0.167, loss=263.707, backward_time=0.229, grad_norm=62.021, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.959e-05, train_time=0.894
+[ip-10-0-194-127:0/16] 2024-02-26 20:54:08,239 (trainer:762) INFO: 1epoch:train:11801-11900batch: iter_time=1.213e-04, forward_time=0.206, loss_ctc=310.740, loss_att=258.270, acc=0.158, loss=274.011, backward_time=0.234, grad_norm=71.097, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.975e-05, train_time=0.896
+[ip-10-0-194-127:0/16] 2024-02-26 20:55:37,687 (trainer:762) INFO: 1epoch:train:11901-12000batch: iter_time=1.229e-04, forward_time=0.205, loss_ctc=287.200, loss_att=239.229, acc=0.167, loss=253.621, backward_time=0.226, grad_norm=76.108, clip=100.000, loss_scale=2.097e+06, optim_step_time=0.047, optim0_lr0=1.992e-05, train_time=0.894
+[ip-10-0-194-127:0/16] 2024-02-26 20:57:06,610 (trainer:762) INFO: 1epoch:train:12001-12100batch: iter_time=1.202e-04, forward_time=0.210, loss_ctc=255.147, loss_att=215.622, acc=0.170, loss=227.479, backward_time=0.226, grad_norm=56.142, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.009e-05, train_time=0.889
+[ip-10-0-194-127:0/16] 2024-02-26 20:58:35,163 (trainer:762) INFO: 1epoch:train:12101-12200batch: iter_time=1.200e-04, forward_time=0.207, loss_ctc=280.781, loss_att=235.336, acc=0.164, loss=248.969, backward_time=0.225, grad_norm=59.149, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.025e-05, train_time=0.885
+[ip-10-0-194-127:0/16] 2024-02-26 21:00:04,869 (trainer:762) INFO: 1epoch:train:12201-12300batch: iter_time=1.227e-04, forward_time=0.206, loss_ctc=270.286, loss_att=225.498, acc=0.166, loss=238.934, backward_time=0.237, grad_norm=73.088, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.042e-05, train_time=0.897
+[ip-10-0-194-127:0/16] 2024-02-26 21:01:34,534 (trainer:762) INFO: 1epoch:train:12301-12400batch: iter_time=1.308e-04, forward_time=0.206, loss_ctc=293.157, loss_att=244.544, acc=0.161, loss=259.128, backward_time=0.231, grad_norm=73.996, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.059e-05, train_time=0.896
+[ip-10-0-194-127:0/16] 2024-02-26 21:03:05,237 (trainer:762) INFO: 1epoch:train:12401-12500batch: iter_time=1.234e-04, forward_time=0.204, loss_ctc=265.070, loss_att=222.643, acc=0.162, loss=235.371, backward_time=0.231, grad_norm=65.333, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.075e-05, train_time=0.907
+[ip-10-0-194-127:0/16] 2024-02-26 21:04:35,638 (trainer:762) INFO: 1epoch:train:12501-12600batch: iter_time=1.200e-04, forward_time=0.203, loss_ctc=236.632, loss_att=199.922, acc=0.174, loss=210.935, backward_time=0.228, grad_norm=62.036, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.092e-05, train_time=0.904
+[ip-10-0-194-127:0/16] 2024-02-26 21:06:05,135 (trainer:762) INFO: 1epoch:train:12601-12700batch: iter_time=1.306e-04, forward_time=0.207, loss_ctc=297.039, loss_att=246.556, acc=0.163, loss=261.701, backward_time=0.229, grad_norm=74.163, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.109e-05, train_time=0.895
+[ip-10-0-194-127:0/16] 2024-02-26 21:07:34,700 (trainer:762) INFO: 1epoch:train:12701-12800batch: iter_time=1.216e-04, forward_time=0.206, loss_ctc=296.309, loss_att=245.351, acc=0.165, loss=260.638, backward_time=0.226, grad_norm=75.679, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.125e-05, train_time=0.895
+[ip-10-0-194-127:0/16] 2024-02-26 21:09:03,857 (trainer:762) INFO: 1epoch:train:12801-12900batch: iter_time=1.219e-04, forward_time=0.207, loss_ctc=271.767, loss_att=228.559, acc=0.166, loss=241.522, backward_time=0.227, grad_norm=79.418, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.142e-05, train_time=0.891
+[ip-10-0-194-127:0/16] 2024-02-26 21:10:33,239 (trainer:762) INFO: 1epoch:train:12901-13000batch: iter_time=1.241e-04, forward_time=0.207, loss_ctc=285.176, loss_att=238.270, acc=0.161, loss=252.342, backward_time=0.229, grad_norm=66.794, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.159e-05, train_time=0.894
+[ip-10-0-194-127:0/16] 2024-02-26 21:12:02,035 (trainer:762) INFO: 1epoch:train:13001-13100batch: iter_time=1.220e-04, forward_time=0.213, loss_ctc=286.943, loss_att=239.227, acc=0.166, loss=253.542, backward_time=0.233, grad_norm=71.261, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.175e-05, train_time=0.888
+[ip-10-0-194-127:0/16] 2024-02-26 21:13:29,414 (trainer:762) INFO: 1epoch:train:13101-13200batch: iter_time=1.234e-04, forward_time=0.208, loss_ctc=270.367, loss_att=225.939, acc=0.175, loss=239.267, backward_time=0.235, grad_norm=63.029, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.192e-05, train_time=0.874
+[ip-10-0-194-127:0/16] 2024-02-26 21:14:56,551 (trainer:762) INFO: 1epoch:train:13201-13300batch: iter_time=1.234e-04, forward_time=0.209, loss_ctc=294.938, loss_att=243.522, acc=0.167, loss=258.947, backward_time=0.232, grad_norm=73.530, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.209e-05, train_time=0.871
+[ip-10-0-194-127:0/16] 2024-02-26 21:16:23,770 (trainer:762) INFO: 1epoch:train:13301-13400batch: iter_time=1.288e-04, forward_time=0.208, loss_ctc=301.262, loss_att=249.999, acc=0.164, loss=265.378, backward_time=0.230, grad_norm=83.931, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.225e-05, train_time=0.872
+[ip-10-0-194-127:0/16] 2024-02-26 21:17:53,861 (trainer:762) INFO: 1epoch:train:13401-13500batch: iter_time=1.288e-04, forward_time=0.206, loss_ctc=284.965, loss_att=234.533, acc=0.167, loss=249.663, backward_time=0.230, grad_norm=70.857, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.242e-05, train_time=0.901
+[ip-10-0-194-127:0/16] 2024-02-26 21:19:23,429 (trainer:762) INFO: 1epoch:train:13501-13600batch: iter_time=1.323e-04, forward_time=0.205, loss_ctc=262.577, loss_att=221.048, acc=0.173, loss=233.507, backward_time=0.227, grad_norm=65.231, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.259e-05, train_time=0.895
+[ip-10-0-194-127:0/16] 2024-02-26 21:20:52,697 (trainer:762) INFO: 1epoch:train:13601-13700batch: iter_time=1.345e-04, forward_time=0.205, loss_ctc=254.896, loss_att=211.852, acc=0.178, loss=224.766, backward_time=0.228, grad_norm=57.879, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.275e-05, train_time=0.892
+[ip-10-0-194-127:0/16] 2024-02-26 21:22:22,695 (trainer:762) INFO: 1epoch:train:13701-13800batch: iter_time=1.226e-04, forward_time=0.206, loss_ctc=294.940, loss_att=244.213, acc=0.161, loss=259.431, backward_time=0.228, grad_norm=60.283, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.292e-05, train_time=0.900
+[ip-10-0-194-127:0/16] 2024-02-26 21:23:51,163 (trainer:762) INFO: 1epoch:train:13801-13900batch: iter_time=1.217e-04, forward_time=0.206, loss_ctc=250.450, loss_att=210.920, acc=0.175, loss=222.779, backward_time=0.226, grad_norm=70.338, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.309e-05, train_time=0.884
+[ip-10-0-194-127:0/16] 2024-02-26 21:25:18,807 (trainer:762) INFO: 1epoch:train:13901-14000batch: iter_time=1.204e-04, forward_time=0.207, loss_ctc=256.129, loss_att=214.181, acc=0.176, loss=226.766, backward_time=0.231, grad_norm=58.162, clip=100.000, loss_scale=4.194e+06, optim_step_time=0.047, optim0_lr0=2.325e-05, train_time=0.876
+[ip-10-0-194-127:0/16] 2024-02-26 21:26:46,376 (trainer:762) INFO: 1epoch:train:14001-14100batch: iter_time=1.153e-04, forward_time=0.213, loss_ctc=260.161, loss_att=220.731, acc=0.175, loss=232.560, backward_time=0.232, grad_norm=56.751, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.047, optim0_lr0=2.342e-05, train_time=0.875
+[ip-10-0-194-127:0/16] 2024-02-26 21:28:12,398 (trainer:762) INFO: 1epoch:train:14101-14200batch: iter_time=1.262e-04, forward_time=0.206, loss_ctc=230.295, loss_att=194.488, acc=0.179, loss=205.230, backward_time=0.226, grad_norm=69.696, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.047, optim0_lr0=2.359e-05, train_time=0.860
+[ip-10-0-194-127:0/16] 2024-02-26 21:29:39,642 (trainer:762) INFO: 1epoch:train:14201-14300batch: iter_time=1.311e-04, forward_time=0.208, loss_ctc=279.281, loss_att=233.178, acc=0.171, loss=247.009, backward_time=0.231, grad_norm=78.201, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.048, optim0_lr0=2.375e-05, train_time=0.872
+[ip-10-0-194-127:0/16] 2024-02-26 21:31:07,058 (trainer:762) INFO: 1epoch:train:14301-14400batch: iter_time=1.209e-04, forward_time=0.208, loss_ctc=248.757, loss_att=207.664, acc=0.177, loss=219.992, backward_time=0.235, grad_norm=77.868, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.047, optim0_lr0=2.392e-05, train_time=0.874
+[ip-10-0-194-127:0/16] 2024-02-26 21:32:37,153 (trainer:762) INFO: 1epoch:train:14401-14500batch: iter_time=1.303e-04, forward_time=0.206, loss_ctc=264.503, loss_att=225.138, acc=0.178, loss=236.947, backward_time=0.231, grad_norm=59.921, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.047, optim0_lr0=2.409e-05, train_time=0.901
+[ip-10-0-194-127:0/16] 2024-02-26 21:34:07,078 (trainer:762) INFO: 1epoch:train:14501-14600batch: iter_time=1.272e-04, forward_time=0.206, loss_ctc=265.304, loss_att=222.492, acc=0.173, loss=235.335, backward_time=0.228, grad_norm=57.736, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.047, optim0_lr0=2.425e-05, train_time=0.899
+[ip-10-0-194-127:0/16] 2024-02-26 21:35:35,682 (trainer:762) INFO: 1epoch:train:14601-14700batch: iter_time=1.268e-04, forward_time=0.205, loss_ctc=254.661, loss_att=213.470, acc=0.181, loss=225.828, backward_time=0.230, grad_norm=67.707, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.047, optim0_lr0=2.442e-05, train_time=0.886
+[ip-10-0-194-127:0/16] 2024-02-26 21:37:03,796 (trainer:762) INFO: 1epoch:train:14701-14800batch: iter_time=1.240e-04, forward_time=0.208, loss_ctc=284.194, loss_att=237.421, acc=0.172, loss=251.453, backward_time=0.229, grad_norm=60.057, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.047, optim0_lr0=2.459e-05, train_time=0.881
+[ip-10-0-194-127:0/16] 2024-02-26 21:38:32,415 (trainer:762) INFO: 1epoch:train:14801-14900batch: iter_time=1.207e-04, forward_time=0.207, loss_ctc=286.827, loss_att=237.332, acc=0.165, loss=252.180, backward_time=0.228, grad_norm=63.796, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.048, optim0_lr0=2.475e-05, train_time=0.886
+[ip-10-0-194-127:0/16] 2024-02-26 21:39:59,409 (trainer:762) INFO: 1epoch:train:14901-15000batch: iter_time=1.238e-04, forward_time=0.208, loss_ctc=269.988, loss_att=224.008, acc=0.176, loss=237.802, backward_time=0.227, grad_norm=72.390, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.047, optim0_lr0=2.492e-05, train_time=0.870
+[ip-10-0-194-127:0/16] 2024-02-26 21:54:20,516 (trainer:361) INFO: 1epoch results: [train] iter_time=0.062, forward_time=0.207, loss_ctc=361.118, loss_att=264.223, acc=0.130, loss=293.291, backward_time=0.229, grad_norm=307.142, clip=100.000, loss_scale=1.669e+06, optim_step_time=0.047, optim0_lr0=1.250e-05, train_time=0.944, time=3 hours, 56 minutes and 21.96 seconds, total_count=15000, gpu_max_cached_mem_GB=35.193, [valid] loss_ctc=214.855, cer_ctc=0.963, loss_att=189.728, acc=0.145, cer=0.728, wer=1.000, loss=197.266, time=14 minutes and 6.91 seconds, total_count=2133, gpu_max_cached_mem_GB=35.193
+[ip-10-0-194-127:0/16] 2024-02-26 21:54:22,703 (font_manager:1547) INFO: generated new fontManager
+[ip-10-0-194-127:0/16] 2024-02-26 21:54:32,529 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-194-127:0/16] 2024-02-26 21:54:32,530 (trainer:290) INFO: 2/45epoch started. Estimated time to finish: 1 week, 15 hours and 50 minutes
+[ip-10-0-194-127:0/16] 2024-02-26 21:54:32,542 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-194-127:0/16] 2024-02-26 21:55:08,460 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-194-127:0/16] 2024-02-26 21:55:16,256 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7fdfc5bbfd00>)
+[ip-10-0-194-127:0/16] 2024-02-26 21:55:16,257 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-194-127:0/16] 2024-02-26 21:55:16,261 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+srun: Job step aborted: Waiting up to 32 seconds for job step to finish.
+slurmstepd: error: *** REASON: burst_buffer/lua: Stage-out in progress ***
+slurmstepd: error: *** STEP 81044.0 ON ip-10-0-194-127 CANCELLED AT 2024-02-26T21:58:53 ***
+slurmstepd: error: *** REASON: burst_buffer/lua: Stage-out in progress ***
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/train.log b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..70d17cf01f6b3e65afd5b3441b0ad4bb1f2d8c01
--- /dev/null
+++ b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/train.log
@@ -0,0 +1,4364 @@
+# Running on ip-10-0-216-33
+# Started at Mon Feb 26 22:03:32 UTC 2024
+# SLURMD_NODENAME=ip-10-0-216-33
+# SLURM_CLUSTER_NAME=external
+# SLURM_CONF=/opt/slurm/etc/slurm.conf
+# SLURM_CPUS_ON_NODE=80
+# SLURM_EXPORT_ENV=PATH
+# SLURM_GET_USER_ENV=1
+# SLURM_GPUS_ON_NODE=8
+# SLURM_GPUS_PER_NODE=8
+# SLURM_GTIDS=0
+# SLURM_JOBID=81480
+# SLURM_JOB_ACCOUNT=wavlab
+# SLURM_JOB_CPUS_PER_NODE='80(x2)'
+# SLURM_JOB_END_TIME=1709762608
+# SLURM_JOB_GID=1879800513
+# SLURM_JOB_GPUS=0,1,2,3,4,5,6,7
+# SLURM_JOB_ID=81480
+# SLURM_JOB_NAME=exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/train.log
+# SLURM_JOB_NODELIST=ip-10-0-216-33,ip-10-0-224-187
+# SLURM_JOB_NUM_NODES=2
+# SLURM_JOB_PARTITION=a40x
+# SLURM_JOB_QOS=normal
+# SLURM_JOB_START_TIME=1708985007
+# SLURM_JOB_UID=1879802134
+# SLURM_JOB_USER=pengyf
+# SLURM_LOCALID=0
+# SLURM_MEM_PER_CPU=10000
+# SLURM_NNODES=2
+# SLURM_NODEID=0
+# SLURM_NODELIST=ip-10-0-216-33,ip-10-0-224-187
+# SLURM_NODE_ALIASES='(null)'
+# SLURM_OPEN_MODE=a
+# SLURM_PRIO_PROCESS=0
+# SLURM_PROCID=0
+# SLURM_SCRIPT_CONTEXT=prolog_task
+# SLURM_SUBMIT_DIR=/weka/home-pengyf/espnet-owsm-train/egs2/owsm_v3.1_licensefree_nosa/s2t1
+# SLURM_SUBMIT_HOST=ip-172-17-22-214
+# SLURM_TASKS_PER_NODE='80(x2)'
+# SLURM_TASK_PID=311799
+# SLURM_TOPOLOGY_ADDR=ip-10-0-216-33
+# SLURM_TOPOLOGY_ADDR_PATTERN=node
+# SLURM_WORKING_CLUSTER=external:ip-172-17-55-206:6817:9984:109
+# srun --export=ALL python3 -m espnet2.bin.s2t_train --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev/wav.scp,speech,kaldi_ark --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/speech_shape --resume true --fold_length 80000 --output_dir exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000 --config conf/train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp/s2t_stats_raw_bpe50000/train/feats_stats.npz --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/wav.scp,speech,kaldi_ark --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/speech_shape --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.prev,text_prev,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_prev_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.ctc,text_ctc,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_ctc_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text,text,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_shape.bpe --multiple_iterator true --valid_data_path_and_name_and_type dump/raw/dev/text.prev,text_prev,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_prev_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text.ctc,text_ctc,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_ctc_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text,text,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_shape.bpe --ngpu 8 --multiprocessing_distributed true --dist_launcher slurm --dist_init_method file:///weka/home-pengyf/espnet-owsm-train/egs2/owsm_v3.1_licensefree_nosa/s2t1/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/.dist_init_d681a8ba-15d8-4a99-88f9-39beba562b1a 
+/weka/home-pengyf/espnet-owsm-train/tools/venv/bin/python3 /weka/home-pengyf/espnet-owsm-train/espnet2/bin/s2t_train.py --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev/wav.scp,speech,kaldi_ark --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/speech_shape --resume true --fold_length 80000 --output_dir exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000 --config conf/train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp/s2t_stats_raw_bpe50000/train/feats_stats.npz --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/wav.scp,speech,kaldi_ark --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/speech_shape --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.prev,text_prev,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_prev_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.ctc,text_ctc,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_ctc_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text,text,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_shape.bpe --multiple_iterator true --valid_data_path_and_name_and_type dump/raw/dev/text.prev,text_prev,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_prev_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text.ctc,text_ctc,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_ctc_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text,text,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_shape.bpe --ngpu 8 --multiprocessing_distributed true --dist_launcher slurm --dist_init_method file:///weka/home-pengyf/espnet-owsm-train/egs2/owsm_v3.1_licensefree_nosa/s2t1/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/.dist_init_d681a8ba-15d8-4a99-88f9-39beba562b1a
+/weka/home-pengyf/espnet-owsm-train/tools/venv/bin/python3 /weka/home-pengyf/espnet-owsm-train/espnet2/bin/s2t_train.py --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev/wav.scp,speech,kaldi_ark --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/speech_shape --resume true --fold_length 80000 --output_dir exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000 --config conf/train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp/s2t_stats_raw_bpe50000/train/feats_stats.npz --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/wav.scp,speech,kaldi_ark --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/speech_shape --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.prev,text_prev,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_prev_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text.ctc,text_ctc,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_ctc_shape.bpe --fold_length 150 --train_data_path_and_name_and_type exp/s2t_stats_raw_bpe50000/splits3/text,text,text --train_shape_file exp/s2t_stats_raw_bpe50000/splits3/text_shape.bpe --multiple_iterator true --valid_data_path_and_name_and_type dump/raw/dev/text.prev,text_prev,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_prev_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text.ctc,text_ctc,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_ctc_shape.bpe --valid_data_path_and_name_and_type dump/raw/dev/text,text,text --valid_shape_file exp/s2t_stats_raw_bpe50000/valid/text_shape.bpe --ngpu 8 --multiprocessing_distributed true --dist_launcher slurm --dist_init_method file:///weka/home-pengyf/espnet-owsm-train/egs2/owsm_v3.1_licensefree_nosa/s2t1/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/.dist_init_d681a8ba-15d8-4a99-88f9-39beba562b1a
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:44,136 (distributed_c10d:319) INFO: Added key: store_based_barrier_key:1 to store for rank: 0
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:50,057 (distributed_c10d:353) INFO: Rank 0: Completed store-based barrier for key:store_based_barrier_key:1 with 16 nodes.
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:50,099 (s2t:464) INFO: Vocabulary size: 50002
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:55,322 (abs_task:1271) INFO: pytorch.version=1.13.1+cu117, cuda.available=True, cudnn.version=8500, cudnn.benchmark=False, cudnn.deterministic=True
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:55,329 (abs_task:1272) INFO: Model structure:
+ESPnetS2TModel(
+  (frontend): DefaultFrontend(
+    (stft): Stft(n_fft=512, win_length=400, hop_length=160, center=True, normalized=False, onesided=True)
+    (frontend): Frontend()
+    (logmel): LogMel(sr=16000, n_fft=512, n_mels=80, fmin=0, fmax=8000.0, htk=False)
+  )
+  (specaug): SpecAug(
+    (freq_mask): MaskAlongAxis(mask_width_range=[0, 27], num_mask=2, axis=freq)
+    (time_mask): MaskAlongAxisVariableMaxWidth(mask_width_ratio_range=[0.0, 0.05], num_mask=4, axis=time)
+  )
+  (normalize): GlobalMVN(stats_file=exp/s2t_stats_raw_bpe50000/train/feats_stats.npz, norm_means=True, norm_vars=True)
+  (encoder): EBranchformerEncoder(
+    (embed): Conv2dSubsampling(
+      (conv): Sequential(
+        (0): Conv2d(1, 768, kernel_size=(3, 3), stride=(2, 2))
+        (1): ReLU()
+        (2): Conv2d(768, 768, kernel_size=(3, 3), stride=(2, 2))
+        (3): ReLU()
+      )
+      (out): Sequential(
+        (0): Linear(in_features=14592, out_features=768, bias=True)
+        (1): PositionalEncoding(
+          (dropout): Dropout(p=0.1, inplace=False)
+        )
+      )
+    )
+    (encoders): MultiSequential(
+      (0): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (1): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (2): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (3): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (4): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (5): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (6): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (7): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+      (8): EBranchformerEncoderLayer(
+        (attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (cgmlp): ConvolutionalGatingMLP(
+          (channel_proj1): Sequential(
+            (0): Linear(in_features=768, out_features=3072, bias=True)
+            (1): GELU(approximate='none')
+          )
+          (csgu): ConvolutionalSpatialGatingUnit(
+            (norm): LayerNorm((1536,), eps=1e-12, elementwise_affine=True)
+            (conv): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+            (act): Identity()
+            (dropout): Dropout(p=0.1, inplace=False)
+          )
+          (channel_proj2): Linear(in_features=1536, out_features=768, bias=True)
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (feed_forward_macaron): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): Swish()
+        )
+        (norm_ff): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_ff_macaron): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mha): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_mlp): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm_final): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+        (depthwise_conv_fusion): Conv1d(1536, 1536, kernel_size=(31,), stride=(1,), padding=(15,), groups=1536)
+        (merge_proj): Linear(in_features=1536, out_features=768, bias=True)
+      )
+    )
+    (after_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+  )
+  (decoder): TransformerDecoder(
+    (embed): Sequential(
+      (0): Embedding(50002, 768)
+      (1): PositionalEncoding(
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+    )
+    (after_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+    (output_layer): Linear(in_features=768, out_features=50002, bias=True)
+    (decoders): MultiSequential(
+      (0): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (1): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (2): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (3): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (4): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (5): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (6): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (7): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+      (8): DecoderLayer(
+        (self_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (src_attn): MultiHeadedAttention(
+          (linear_q): Linear(in_features=768, out_features=768, bias=True)
+          (linear_k): Linear(in_features=768, out_features=768, bias=True)
+          (linear_v): Linear(in_features=768, out_features=768, bias=True)
+          (linear_out): Linear(in_features=768, out_features=768, bias=True)
+          (dropout): Identity()
+          (q_norm): Identity()
+          (k_norm): Identity()
+        )
+        (feed_forward): PositionwiseFeedForward(
+          (w_1): Linear(in_features=768, out_features=3072, bias=True)
+          (w_2): Linear(in_features=3072, out_features=768, bias=True)
+          (dropout): Dropout(p=0.1, inplace=False)
+          (activation): ReLU()
+        )
+        (norm1): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm2): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (norm3): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+        (dropout): Dropout(p=0.1, inplace=False)
+      )
+    )
+  )
+  (criterion_att): LabelSmoothingLoss(
+    (criterion): KLDivLoss()
+  )
+  (ctc): CTC(
+    (ctc_lo): Linear(in_features=768, out_features=50002, bias=True)
+    (ctc_loss): CTCLoss()
+  )
+)
+
+Model summary:
+    Class Name: ESPnetS2TModel
+    Total Number of model parameters: 366.65 M
+    Number of trainable parameters: 366.65 M (100.0%)
+    Size: 1.47 GB
+    Type: torch.float32
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:55,330 (abs_task:1275) INFO: Optimizer:
+AdamW (
+Parameter Group 0
+    amsgrad: False
+    betas: [0.9, 0.98]
+    capturable: False
+    eps: 1e-06
+    foreach: None
+    initial_lr: 0.0005
+    lr: 1.6666666666666667e-09
+    maximize: False
+    weight_decay: 0.0
+)
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:55,330 (abs_task:1276) INFO: Scheduler: PiecewiseLinearWarmupLR(warmup_steps_list=[0, 30000, 60000], warmup_lr_list=[0.0, 5e-05, 0.0005])
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:55,331 (abs_task:1285) INFO: Saving the configuration in exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/config.yaml
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:59,181 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:59,737 (abs_task:1663) INFO: [valid] dataset:
+ESPnetDataset(
+  speech: {"path": "dump/raw/dev/wav.scp", "type": "kaldi_ark"}
+  text_prev: {"path": "dump/raw/dev/text.prev", "type": "text"}
+  text_ctc: {"path": "dump/raw/dev/text.ctc", "type": "text"}
+  text: {"path": "dump/raw/dev/text", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f6670a0c490>)
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:59,737 (abs_task:1664) INFO: [valid] Batch sampler: UnsortedBatchSampler(N-batch=2133, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/valid/speech_shape, 
+[ip-10-0-216-33:0/16] 2024-02-26 22:03:59,738 (abs_task:1665) INFO: [valid] mini-batch sizes summary: N-batch=2133, mean=256.1, min=256, max=257
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: Currently logged in as: pengyf. Use `wandb login --relogin` to force relogin
+wandb: - Waiting for wandb.init()...
wandb: \ Waiting for wandb.init()...
wandb: | Waiting for wandb.init()...
wandb: / Waiting for wandb.init()...
wandb: WARNING Serializing object of type list that is 406488 bytes
+wandb: Tracking run with wandb version 0.16.3
+wandb: Run data is saved locally in exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/wandb/run-20240226_220402-zw9g4dhv
+wandb: Run `wandb offline` to turn off syncing.
+wandb: Resuming run _weka_home-pengyf_espnet-owsm-train_egs2_owsm_v3.1_licensefree_nosa_s2t1
+wandb: ⭐️ View project at https://stability.wandb.io/pengyf/ESPnet_S2TTask
+wandb: 🚀 View run at https://stability.wandb.io/pengyf/ESPnet_S2TTask/runs/zw9g4dhv
+wandb: WARNING Serializing object of type list that is 400072 bytes
+wandb: WARNING Serializing object of type list that is 406488 bytes
+[ip-10-0-216-33:0/16] 2024-02-26 22:04:09,855 (trainer:168) INFO: The training was resumed using exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/checkpoint.pth
+ip-10-0-216-33:311880:311880 [0] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:311880:311880 [0] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:311880:311880 [0] NCCL INFO cudaDriverVersion 12000
+NCCL version 2.14.3+cuda11.7
+[ip-10-0-216-33:0/16] 2024-02-26 22:04:12,263 (trainer:302) INFO: 2/45epoch started
+[ip-10-0-216-33:0/16] 2024-02-26 22:04:12,307 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-26 22:04:48,404 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-26 22:04:56,324 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f64bc0d92e0>)
+[ip-10-0-216-33:0/16] 2024-02-26 22:04:56,324 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-26 22:04:56,329 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+ip-10-0-216-33:311883:311883 [3] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:311883:311883 [3] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:311883:311883 [3] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:311883:312363 [3] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:311883:312363 [3] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:311883:312363 [3] NCCL INFO Using network Socket
+ip-10-0-216-33:311883:312363 [3] NCCL INFO Setting affinity for GPU 3 to 0f,ffff0000,000fffff
+ip-10-0-216-33:311883:312363 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2
+ip-10-0-216-33:311883:312363 [3] NCCL INFO Channel 00/0 : 3[201d0] -> 2[201c0] via P2P/IPC/read
+ip-10-0-216-33:311883:312363 [3] NCCL INFO Channel 01/0 : 3[201d0] -> 2[201c0] via P2P/IPC/read
+ip-10-0-216-33:311883:312363 [3] NCCL INFO Connected all rings
+ip-10-0-216-33:311883:312363 [3] NCCL INFO Channel 00/0 : 3[201d0] -> 4[901c0] via P2P/IPC/read
+ip-10-0-216-33:311883:312363 [3] NCCL INFO Channel 01/0 : 3[201d0] -> 4[901c0] via P2P/IPC/read
+ip-10-0-216-33:311883:312363 [3] NCCL INFO Connected all trees
+ip-10-0-216-33:311883:312363 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:311883:312363 [3] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:311883:312363 [3] NCCL INFO comm 0x6f267bc0 rank 3 nranks 16 cudaDev 3 busId 201d0 - Init COMPLETE
+ip-10-0-216-33:311884:311884 [4] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:311884:311884 [4] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:311884:311884 [4] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:311884:312360 [4] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:311884:312360 [4] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:311884:312360 [4] NCCL INFO Using network Socket
+ip-10-0-216-33:311884:312360 [4] NCCL INFO Setting affinity for GPU 4 to 0fffff00,00000fff,ff000000
+ip-10-0-216-33:311884:312360 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3
+ip-10-0-216-33:311884:312360 [4] NCCL INFO Channel 00/0 : 4[901c0] -> 3[201d0] via P2P/IPC/read
+ip-10-0-216-33:311884:312360 [4] NCCL INFO Channel 01/0 : 4[901c0] -> 3[201d0] via P2P/IPC/read
+ip-10-0-216-33:311884:312360 [4] NCCL INFO Connected all rings
+ip-10-0-216-33:311884:312360 [4] NCCL INFO Channel 00/0 : 4[901c0] -> 5[901d0] via P2P/IPC/read
+ip-10-0-216-33:311884:312360 [4] NCCL INFO Channel 01/0 : 4[901c0] -> 5[901d0] via P2P/IPC/read
+ip-10-0-216-33:311884:312360 [4] NCCL INFO Connected all trees
+ip-10-0-216-33:311884:312360 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:311884:312360 [4] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:311884:312360 [4] NCCL INFO comm 0x9d5b200 rank 4 nranks 16 cudaDev 4 busId 901c0 - Init COMPLETE
+ip-10-0-216-33:311881:311881 [1] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:311881:311881 [1] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:311881:311881 [1] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:311881:312361 [1] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:311881:312361 [1] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Using network Socket
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Setting affinity for GPU 1 to 0f,ffff0000,000fffff
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Channel 00/0 : 1[101d0] -> 8[101c0] [send] via NET/Socket/0
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Channel 01/0 : 1[101d0] -> 8[101c0] [send] via NET/Socket/0
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Connected all rings
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Channel 00/0 : 1[101d0] -> 2[201c0] via P2P/IPC/read
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Channel 01/0 : 1[101d0] -> 2[201c0] via P2P/IPC/read
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Channel 00/0 : 1[101d0] -> 0[101c0] via P2P/IPC/read
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Channel 01/0 : 1[101d0] -> 0[101c0] via P2P/IPC/read
+ip-10-0-216-33:311881:312361 [1] NCCL INFO Connected all trees
+ip-10-0-216-33:311881:312361 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:311881:312361 [1] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:311881:312361 [1] NCCL INFO comm 0x9a92260 rank 1 nranks 16 cudaDev 1 busId 101d0 - Init COMPLETE
+ip-10-0-216-33:311886:311886 [6] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:311886:311886 [6] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:311886:311886 [6] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:311886:312359 [6] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:311886:312359 [6] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:311886:312359 [6] NCCL INFO Using network Socket
+ip-10-0-216-33:311886:312359 [6] NCCL INFO Setting affinity for GPU 6 to 0fffff00,00000fff,ff000000
+ip-10-0-216-33:311886:312359 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5
+ip-10-0-216-33:311886:312359 [6] NCCL INFO Channel 00/0 : 6[a01c0] -> 5[901d0] via P2P/IPC/read
+ip-10-0-216-33:311886:312359 [6] NCCL INFO Channel 01/0 : 6[a01c0] -> 5[901d0] via P2P/IPC/read
+ip-10-0-216-33:311886:312359 [6] NCCL INFO Connected all rings
+ip-10-0-216-33:311886:312359 [6] NCCL INFO Channel 00/0 : 6[a01c0] -> 7[a01d0] via P2P/IPC/read
+ip-10-0-216-33:311886:312359 [6] NCCL INFO Channel 01/0 : 6[a01c0] -> 7[a01d0] via P2P/IPC/read
+ip-10-0-216-33:311886:312359 [6] NCCL INFO Connected all trees
+ip-10-0-216-33:311886:312359 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:311886:312359 [6] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:311886:312359 [6] NCCL INFO comm 0x42c0fbc0 rank 6 nranks 16 cudaDev 6 busId a01c0 - Init COMPLETE
+ip-10-0-216-33:311882:311882 [2] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:311882:311882 [2] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:311882:311882 [2] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:311882:312358 [2] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:311882:312358 [2] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:311882:312358 [2] NCCL INFO Using network Socket
+ip-10-0-216-33:311882:312358 [2] NCCL INFO Setting affinity for GPU 2 to 0f,ffff0000,000fffff
+ip-10-0-216-33:311882:312358 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1
+ip-10-0-216-33:311882:312358 [2] NCCL INFO Channel 00/0 : 2[201c0] -> 1[101d0] via P2P/IPC/read
+ip-10-0-216-33:311882:312358 [2] NCCL INFO Channel 01/0 : 2[201c0] -> 1[101d0] via P2P/IPC/read
+ip-10-0-216-33:311882:312358 [2] NCCL INFO Connected all rings
+ip-10-0-216-33:311882:312358 [2] NCCL INFO Channel 00/0 : 2[201c0] -> 3[201d0] via P2P/IPC/read
+ip-10-0-216-33:311882:312358 [2] NCCL INFO Channel 01/0 : 2[201c0] -> 3[201d0] via P2P/IPC/read
+ip-10-0-216-33:311882:312358 [2] NCCL INFO Connected all trees
+ip-10-0-216-33:311882:312358 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:311882:312358 [2] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:311882:312358 [2] NCCL INFO comm 0x511a1b90 rank 2 nranks 16 cudaDev 2 busId 201c0 - Init COMPLETE
+ip-10-0-224-187:3788824:3788824 [6] NCCL INFO cudaDriverVersion 12000
+ip-10-0-224-187:3788824:3788824 [6] NCCL INFO Bootstrap : Using ens32:10.0.224.187<0>
+ip-10-0-224-187:3788824:3788824 [6] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO NET/IB : No device found.
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO NET/Socket : Using [0]ens32:10.0.224.187<0>
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO Using network Socket
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO Setting affinity for GPU 6 to 0fffff00,00000fff,ff000000
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO Channel 00/0 : 14[a01c0] -> 13[901d0] via P2P/IPC/read
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO Channel 01/0 : 14[a01c0] -> 13[901d0] via P2P/IPC/read
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO Connected all rings
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO Channel 00/0 : 14[a01c0] -> 15[a01d0] via P2P/IPC/read
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO Channel 01/0 : 14[a01c0] -> 15[a01d0] via P2P/IPC/read
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO Connected all trees
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-224-187:3788824:3789267 [6] NCCL INFO comm 0x704da1d0 rank 14 nranks 16 cudaDev 6 busId a01c0 - Init COMPLETE
+ip-10-0-224-187:3788822:3788822 [4] NCCL INFO cudaDriverVersion 12000
+ip-10-0-224-187:3788822:3788822 [4] NCCL INFO Bootstrap : Using ens32:10.0.224.187<0>
+ip-10-0-224-187:3788822:3788822 [4] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO NET/IB : No device found.
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO NET/Socket : Using [0]ens32:10.0.224.187<0>
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO Using network Socket
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO Setting affinity for GPU 4 to 0fffff00,00000fff,ff000000
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO Channel 00/0 : 12[901c0] -> 11[201d0] via P2P/IPC/read
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO Channel 01/0 : 12[901c0] -> 11[201d0] via P2P/IPC/read
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO Connected all rings
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO Channel 00/0 : 12[901c0] -> 13[901d0] via P2P/IPC/read
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO Channel 01/0 : 12[901c0] -> 13[901d0] via P2P/IPC/read
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO Connected all trees
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-224-187:3788822:3789266 [4] NCCL INFO comm 0x51831900 rank 12 nranks 16 cudaDev 4 busId 901c0 - Init COMPLETE
+ip-10-0-224-187:3788826:3788826 [7] NCCL INFO cudaDriverVersion 12000
+ip-10-0-224-187:3788826:3788826 [7] NCCL INFO Bootstrap : Using ens32:10.0.224.187<0>
+ip-10-0-224-187:3788826:3788826 [7] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO NET/IB : No device found.
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO NET/Socket : Using [0]ens32:10.0.224.187<0>
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO Using network Socket
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO Setting affinity for GPU 7 to 0fffff00,00000fff,ff000000
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO Channel 00/0 : 15[a01d0] -> 14[a01c0] via P2P/IPC/read
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO Channel 01/0 : 15[a01d0] -> 14[a01c0] via P2P/IPC/read
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO Connected all rings
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO Connected all trees
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-224-187:3788826:3789273 [7] NCCL INFO comm 0x42154f70 rank 15 nranks 16 cudaDev 7 busId a01d0 - Init COMPLETE
+ip-10-0-224-187:3788818:3788818 [0] NCCL INFO cudaDriverVersion 12000
+ip-10-0-224-187:3788818:3788818 [0] NCCL INFO Bootstrap : Using ens32:10.0.224.187<0>
+ip-10-0-224-187:3788818:3788818 [0] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO NET/IB : No device found.
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO NET/Socket : Using [0]ens32:10.0.224.187<0>
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Using network Socket
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Setting affinity for GPU 0 to 0f,ffff0000,000fffff
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Trees [0] 9/-1/-1->8->0 [1] 9/0/-1->8->-1
+ip-10-0-224-187:3788818:3789278 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Channel 00/0 : 1[101d0] -> 8[101c0] [receive] via NET/Socket/0
+ip-10-0-224-187:3788818:3789278 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Channel 01/0 : 1[101d0] -> 8[101c0] [receive] via NET/Socket/0
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Channel 00/0 : 8[101c0] -> 15[a01d0] via P2P/IPC/read
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Channel 01/0 : 8[101c0] -> 15[a01d0] via P2P/IPC/read
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Connected all rings
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Channel 00/0 : 8[101c0] -> 9[101d0] via P2P/IPC/read
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Channel 01/0 : 8[101c0] -> 9[101d0] via P2P/IPC/read
+ip-10-0-224-187:3788818:3789278 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Channel 00/0 : 0[101c0] -> 8[101c0] [receive] via NET/Socket/0
+ip-10-0-224-187:3788818:3789278 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Channel 01/0 : 0[101c0] -> 8[101c0] [receive] via NET/Socket/0
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Channel 00/0 : 8[101c0] -> 0[101c0] [send] via NET/Socket/0
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Channel 01/0 : 8[101c0] -> 0[101c0] [send] via NET/Socket/0
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO Connected all trees
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-224-187:3788818:3789270 [0] NCCL INFO comm 0x804dad0 rank 8 nranks 16 cudaDev 0 busId 101c0 - Init COMPLETE
+ip-10-0-224-187:3788823:3788823 [5] NCCL INFO cudaDriverVersion 12000
+ip-10-0-224-187:3788823:3788823 [5] NCCL INFO Bootstrap : Using ens32:10.0.224.187<0>
+ip-10-0-224-187:3788823:3788823 [5] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO NET/IB : No device found.
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO NET/Socket : Using [0]ens32:10.0.224.187<0>
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO Using network Socket
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO Setting affinity for GPU 5 to 0fffff00,00000fff,ff000000
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO Channel 00/0 : 13[901d0] -> 12[901c0] via P2P/IPC/read
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO Channel 01/0 : 13[901d0] -> 12[901c0] via P2P/IPC/read
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO Connected all rings
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO Channel 00/0 : 13[901d0] -> 14[a01c0] via P2P/IPC/read
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO Channel 01/0 : 13[901d0] -> 14[a01c0] via P2P/IPC/read
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO Connected all trees
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-224-187:3788823:3789269 [5] NCCL INFO comm 0xa4dcd80 rank 13 nranks 16 cudaDev 5 busId 901d0 - Init COMPLETE
+ip-10-0-224-187:3788821:3788821 [3] NCCL INFO cudaDriverVersion 12000
+ip-10-0-224-187:3788821:3788821 [3] NCCL INFO Bootstrap : Using ens32:10.0.224.187<0>
+ip-10-0-224-187:3788821:3788821 [3] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO NET/IB : No device found.
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO NET/Socket : Using [0]ens32:10.0.224.187<0>
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO Using network Socket
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO Setting affinity for GPU 3 to 0f,ffff0000,000fffff
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO Channel 00/0 : 11[201d0] -> 10[201c0] via P2P/IPC/read
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO Channel 01/0 : 11[201d0] -> 10[201c0] via P2P/IPC/read
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO Connected all rings
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO Channel 00/0 : 11[201d0] -> 12[901c0] via P2P/IPC/read
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO Channel 01/0 : 11[201d0] -> 12[901c0] via P2P/IPC/read
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO Connected all trees
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-224-187:3788821:3789268 [3] NCCL INFO comm 0x729fe240 rank 11 nranks 16 cudaDev 3 busId 201d0 - Init COMPLETE
+ip-10-0-224-187:3788819:3788819 [1] NCCL INFO cudaDriverVersion 12000
+ip-10-0-224-187:3788819:3788819 [1] NCCL INFO Bootstrap : Using ens32:10.0.224.187<0>
+ip-10-0-224-187:3788819:3788819 [1] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO NET/IB : No device found.
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO NET/Socket : Using [0]ens32:10.0.224.187<0>
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Using network Socket
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Setting affinity for GPU 1 to 0f,ffff0000,000fffff
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->8
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Channel 00/0 : 9[101d0] -> 0[101c0] [send] via NET/Socket/0
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Channel 01/0 : 9[101d0] -> 0[101c0] [send] via NET/Socket/0
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Connected all rings
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Channel 00/0 : 9[101d0] -> 10[201c0] via P2P/IPC/read
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Channel 01/0 : 9[101d0] -> 10[201c0] via P2P/IPC/read
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Channel 00/0 : 9[101d0] -> 8[101c0] via P2P/IPC/read
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Channel 01/0 : 9[101d0] -> 8[101c0] via P2P/IPC/read
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO Connected all trees
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-224-187:3788819:3789272 [1] NCCL INFO comm 0x1236d920 rank 9 nranks 16 cudaDev 1 busId 101d0 - Init COMPLETE
+ip-10-0-216-33:311887:311887 [7] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:311887:311887 [7] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:311887:311887 [7] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:311887:312362 [7] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:311887:312362 [7] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:311887:312362 [7] NCCL INFO Using network Socket
+ip-10-0-216-33:311887:312362 [7] NCCL INFO Setting affinity for GPU 7 to 0fffff00,00000fff,ff000000
+ip-10-0-216-33:311887:312362 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6
+ip-10-0-216-33:311887:312362 [7] NCCL INFO Channel 00/0 : 7[a01d0] -> 6[a01c0] via P2P/IPC/read
+ip-10-0-216-33:311887:312362 [7] NCCL INFO Channel 01/0 : 7[a01d0] -> 6[a01c0] via P2P/IPC/read
+ip-10-0-216-33:311887:312362 [7] NCCL INFO Connected all rings
+ip-10-0-216-33:311887:312362 [7] NCCL INFO Connected all trees
+ip-10-0-216-33:311887:312362 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:311887:312362 [7] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:311887:312362 [7] NCCL INFO comm 0x6ee1f170 rank 7 nranks 16 cudaDev 7 busId a01d0 - Init COMPLETE
+ip-10-0-216-33:311885:311885 [5] NCCL INFO cudaDriverVersion 12000
+ip-10-0-216-33:311885:311885 [5] NCCL INFO Bootstrap : Using ens32:10.0.216.33<0>
+ip-10-0-216-33:311885:311885 [5] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-216-33:311885:312364 [5] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:311885:312364 [5] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:311885:312364 [5] NCCL INFO Using network Socket
+ip-10-0-216-33:311885:312364 [5] NCCL INFO Setting affinity for GPU 5 to 0fffff00,00000fff,ff000000
+ip-10-0-216-33:311885:312364 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4
+ip-10-0-216-33:311885:312364 [5] NCCL INFO Channel 00/0 : 5[901d0] -> 4[901c0] via P2P/IPC/read
+ip-10-0-216-33:311885:312364 [5] NCCL INFO Channel 01/0 : 5[901d0] -> 4[901c0] via P2P/IPC/read
+ip-10-0-216-33:311885:312364 [5] NCCL INFO Connected all rings
+ip-10-0-216-33:311885:312364 [5] NCCL INFO Channel 00/0 : 5[901d0] -> 6[a01c0] via P2P/IPC/read
+ip-10-0-216-33:311885:312364 [5] NCCL INFO Channel 01/0 : 5[901d0] -> 6[a01c0] via P2P/IPC/read
+ip-10-0-216-33:311885:312364 [5] NCCL INFO Connected all trees
+ip-10-0-216-33:311885:312364 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:311885:312364 [5] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:311885:312364 [5] NCCL INFO comm 0x45e56440 rank 5 nranks 16 cudaDev 5 busId 901d0 - Init COMPLETE
+ip-10-0-224-187:3788820:3788820 [2] NCCL INFO cudaDriverVersion 12000
+ip-10-0-224-187:3788820:3788820 [2] NCCL INFO Bootstrap : Using ens32:10.0.224.187<0>
+ip-10-0-224-187:3788820:3788820 [2] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO NET/IB : No device found.
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO NET/Socket : Using [0]ens32:10.0.224.187<0>
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO Using network Socket
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO Setting affinity for GPU 2 to 0f,ffff0000,000fffff
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO Channel 00/0 : 10[201c0] -> 9[101d0] via P2P/IPC/read
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO Channel 01/0 : 10[201c0] -> 9[101d0] via P2P/IPC/read
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO Connected all rings
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO Channel 00/0 : 10[201c0] -> 11[201d0] via P2P/IPC/read
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO Channel 01/0 : 10[201c0] -> 11[201d0] via P2P/IPC/read
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO Connected all trees
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-224-187:3788820:3789271 [2] NCCL INFO comm 0x965e2c0 rank 10 nranks 16 cudaDev 2 busId 201c0 - Init COMPLETE
+ip-10-0-216-33:311880:312357 [0] NCCL INFO NET/IB : No device found.
+ip-10-0-216-33:311880:312357 [0] NCCL INFO NET/Socket : Using [0]ens32:10.0.216.33<0>
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Using network Socket
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Setting affinity for GPU 0 to 0f,ffff0000,000fffff
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 00/02 :    0   7   6   5   4   3   2   1   8  15  14  13  12  11  10   9
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 01/02 :    0   7   6   5   4   3   2   1   8  15  14  13  12  11  10   9
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Trees [0] 1/8/-1->0->-1 [1] 1/-1/-1->0->8
+ip-10-0-216-33:311880:312455 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 00/0 : 9[101d0] -> 0[101c0] [receive] via NET/Socket/0
+ip-10-0-216-33:311880:312455 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 01/0 : 9[101d0] -> 0[101c0] [receive] via NET/Socket/0
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 00/0 : 0[101c0] -> 7[a01d0] via P2P/IPC/read
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 01/0 : 0[101c0] -> 7[a01d0] via P2P/IPC/read
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Connected all rings
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 00/0 : 0[101c0] -> 1[101d0] via P2P/IPC/read
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 01/0 : 0[101c0] -> 1[101d0] via P2P/IPC/read
+ip-10-0-216-33:311880:312455 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 00/0 : 8[101c0] -> 0[101c0] [receive] via NET/Socket/0
+ip-10-0-216-33:311880:312455 [0] NCCL INFO NET/Socket: Using 2 threads and 8 sockets per thread
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 01/0 : 8[101c0] -> 0[101c0] [receive] via NET/Socket/0
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 00/0 : 0[101c0] -> 8[101c0] [send] via NET/Socket/0
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Channel 01/0 : 0[101c0] -> 8[101c0] [send] via NET/Socket/0
+ip-10-0-216-33:311880:312357 [0] NCCL INFO Connected all trees
+ip-10-0-216-33:311880:312357 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512
+ip-10-0-216-33:311880:312357 [0] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
+ip-10-0-216-33:311880:312357 [0] NCCL INFO comm 0x90d2b80 rank 0 nranks 16 cudaDev 0 busId 101c0 - Init COMPLETE
+[ip-10-0-216-33:0/16] 2024-02-26 22:11:41,417 (distributed:1027) INFO: Reducer buckets have been rebuilt in this iteration.
+[ip-10-0-216-33:0/16] 2024-02-26 22:19:15,625 (trainer:762) INFO: 2epoch:train:1-500batch: iter_time=0.878, forward_time=0.211, loss_ctc=256.046, loss_att=215.117, acc=0.175, loss=227.395, backward_time=0.230, grad_norm=71.288, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.075, optim0_lr0=2.542e-05, train_time=1.806
+[ip-10-0-216-33:0/16] 2024-02-26 22:26:45,984 (trainer:762) INFO: 2epoch:train:501-1000batch: iter_time=1.560e-04, forward_time=0.205, loss_ctc=264.017, loss_att=222.907, acc=0.179, loss=235.240, backward_time=0.226, grad_norm=78.517, clip=100.000, loss_scale=8.389e+06, optim_step_time=0.074, optim0_lr0=2.625e-05, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-26 22:34:18,504 (trainer:762) INFO: 2epoch:train:1001-1500batch: iter_time=1.508e-04, forward_time=0.207, loss_ctc=252.138, loss_att=216.662, acc=0.180, loss=227.305, backward_time=0.227, grad_norm=74.385, clip=100.000, loss_scale=1.678e+07, optim_step_time=0.074, optim0_lr0=2.709e-05, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-26 22:41:52,449 (trainer:762) INFO: 2epoch:train:1501-2000batch: iter_time=1.466e-04, forward_time=0.206, loss_ctc=244.770, loss_att=214.569, acc=0.180, loss=223.629, backward_time=0.228, grad_norm=78.179, clip=100.000, loss_scale=1.678e+07, optim_step_time=0.074, optim0_lr0=2.792e-05, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-26 22:49:28,708 (trainer:762) INFO: 2epoch:train:2001-2500batch: iter_time=1.502e-04, forward_time=0.205, loss_ctc=248.980, loss_att=224.210, acc=0.179, loss=231.641, backward_time=0.225, grad_norm=76.967, clip=100.000, loss_scale=1.678e+07, optim_step_time=0.074, optim0_lr0=2.875e-05, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-26 22:56:57,090 (trainer:762) INFO: 2epoch:train:2501-3000batch: iter_time=1.455e-04, forward_time=0.206, loss_ctc=222.364, loss_att=203.249, acc=0.183, loss=208.983, backward_time=0.227, grad_norm=80.769, clip=100.000, loss_scale=1.678e+07, optim_step_time=0.074, optim0_lr0=2.959e-05, train_time=0.896
+[ip-10-0-216-33:0/16] 2024-02-26 23:04:26,477 (trainer:762) INFO: 2epoch:train:3001-3500batch: iter_time=1.456e-04, forward_time=0.206, loss_ctc=225.000, loss_att=210.333, acc=0.184, loss=214.733, backward_time=0.229, grad_norm=83.386, clip=100.000, loss_scale=3.355e+07, optim_step_time=0.074, optim0_lr0=3.042e-05, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-02-26 23:11:59,520 (trainer:762) INFO: 2epoch:train:3501-4000batch: iter_time=1.482e-04, forward_time=0.205, loss_ctc=229.381, loss_att=220.539, acc=0.181, loss=223.191, backward_time=0.229, grad_norm=86.785, clip=100.000, loss_scale=3.355e+07, optim_step_time=0.074, optim0_lr0=3.125e-05, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-26 23:19:31,753 (trainer:762) INFO: 2epoch:train:4001-4500batch: iter_time=1.463e-04, forward_time=0.206, loss_ctc=222.672, loss_att=220.824, acc=0.183, loss=221.379, backward_time=0.227, grad_norm=83.928, clip=100.000, loss_scale=3.355e+07, optim_step_time=0.073, optim0_lr0=3.209e-05, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-26 23:27:06,817 (trainer:762) INFO: 2epoch:train:4501-5000batch: iter_time=1.458e-04, forward_time=0.206, loss_ctc=210.652, loss_att=208.617, acc=0.190, loss=209.228, backward_time=0.226, grad_norm=82.040, clip=100.000, loss_scale=3.355e+07, optim_step_time=0.074, optim0_lr0=3.292e-05, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-26 23:27:13,412 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-26 23:27:50,474 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-26 23:27:58,432 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f64bc461a90>)
+[ip-10-0-216-33:0/16] 2024-02-26 23:27:58,433 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-26 23:27:58,437 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-26 23:42:10,656 (trainer:762) INFO: 2epoch:train:5001-5500batch: iter_time=0.876, forward_time=0.207, loss_ctc=200.892, loss_att=204.306, acc=0.189, loss=203.281, backward_time=0.230, grad_norm=81.583, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.073, optim0_lr0=3.375e-05, train_time=1.807
+[ip-10-0-216-33:0/16] 2024-02-26 23:49:42,183 (trainer:762) INFO: 2epoch:train:5501-6000batch: iter_time=1.485e-04, forward_time=0.208, loss_ctc=199.205, loss_att=212.214, acc=0.192, loss=208.312, backward_time=0.229, grad_norm=77.359, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.073, optim0_lr0=3.459e-05, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-26 23:57:15,796 (trainer:762) INFO: 2epoch:train:6001-6500batch: iter_time=1.494e-04, forward_time=0.206, loss_ctc=191.261, loss_att=206.538, acc=0.194, loss=201.955, backward_time=0.233, grad_norm=77.979, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.073, optim0_lr0=3.542e-05, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-27 00:04:42,410 (trainer:762) INFO: 2epoch:train:6501-7000batch: iter_time=1.455e-04, forward_time=0.208, loss_ctc=187.001, loss_att=204.608, acc=0.195, loss=199.326, backward_time=0.229, grad_norm=75.064, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.073, optim0_lr0=3.625e-05, train_time=0.893
+[ip-10-0-216-33:0/16] 2024-02-27 00:12:17,971 (trainer:762) INFO: 2epoch:train:7001-7500batch: iter_time=1.487e-04, forward_time=0.207, loss_ctc=186.698, loss_att=214.496, acc=0.193, loss=206.157, backward_time=0.230, grad_norm=73.954, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.073, optim0_lr0=3.709e-05, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-27 00:19:45,681 (trainer:762) INFO: 2epoch:train:7501-8000batch: iter_time=1.459e-04, forward_time=0.207, loss_ctc=171.050, loss_att=193.820, acc=0.197, loss=186.989, backward_time=0.231, grad_norm=66.564, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.073, optim0_lr0=3.792e-05, train_time=0.895
+[ip-10-0-216-33:0/16] 2024-02-27 00:27:16,194 (trainer:762) INFO: 2epoch:train:8001-8500batch: iter_time=1.467e-04, forward_time=0.208, loss_ctc=170.404, loss_att=200.856, acc=0.199, loss=191.720, backward_time=0.230, grad_norm=66.618, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.074, optim0_lr0=3.875e-05, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-27 00:34:48,445 (trainer:762) INFO: 2epoch:train:8501-9000batch: iter_time=1.544e-04, forward_time=0.207, loss_ctc=175.160, loss_att=210.792, acc=0.196, loss=200.102, backward_time=0.226, grad_norm=72.661, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.074, optim0_lr0=3.959e-05, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-27 00:42:14,488 (trainer:762) INFO: 2epoch:train:9001-9500batch: iter_time=1.511e-04, forward_time=0.209, loss_ctc=172.136, loss_att=212.137, acc=0.199, loss=200.137, backward_time=0.229, grad_norm=68.953, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.074, optim0_lr0=4.042e-05, train_time=0.892
+[ip-10-0-216-33:0/16] 2024-02-27 00:49:47,158 (trainer:762) INFO: 2epoch:train:9501-10000batch: iter_time=1.526e-04, forward_time=0.207, loss_ctc=164.211, loss_att=199.953, acc=0.206, loss=189.231, backward_time=0.230, grad_norm=64.642, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.073, optim0_lr0=4.125e-05, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-27 00:49:51,618 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 00:50:29,712 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 00:50:38,084 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ec7046d00>)
+[ip-10-0-216-33:0/16] 2024-02-27 00:50:38,084 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-27 00:50:38,088 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 01:04:55,686 (trainer:762) INFO: 2epoch:train:10001-10500batch: iter_time=0.901, forward_time=0.208, loss_ctc=158.463, loss_att=195.850, acc=0.204, loss=184.634, backward_time=0.226, grad_norm=64.366, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.074, optim0_lr0=4.209e-05, train_time=1.817
+[ip-10-0-216-33:0/16] 2024-02-27 01:12:23,475 (trainer:762) INFO: 2epoch:train:10501-11000batch: iter_time=1.477e-04, forward_time=0.209, loss_ctc=155.730, loss_att=203.552, acc=0.207, loss=189.205, backward_time=0.227, grad_norm=62.679, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.074, optim0_lr0=4.292e-05, train_time=0.895
+[ip-10-0-216-33:0/16] 2024-02-27 01:19:52,735 (trainer:762) INFO: 2epoch:train:11001-11500batch: iter_time=1.530e-04, forward_time=0.209, loss_ctc=151.626, loss_att=198.325, acc=0.210, loss=184.315, backward_time=0.228, grad_norm=62.646, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.074, optim0_lr0=4.375e-05, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-02-27 01:27:25,077 (trainer:762) INFO: 2epoch:train:11501-12000batch: iter_time=1.498e-04, forward_time=0.208, loss_ctc=151.375, loss_att=197.014, acc=0.210, loss=183.322, backward_time=0.229, grad_norm=61.172, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.074, optim0_lr0=4.459e-05, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-27 01:34:53,578 (trainer:762) INFO: 2epoch:train:12001-12500batch: iter_time=1.487e-04, forward_time=0.209, loss_ctc=149.224, loss_att=206.618, acc=0.209, loss=189.400, backward_time=0.228, grad_norm=59.147, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.074, optim0_lr0=4.542e-05, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-02-27 01:42:26,491 (trainer:762) INFO: 2epoch:train:12501-13000batch: iter_time=1.507e-04, forward_time=0.206, loss_ctc=139.965, loss_att=185.753, acc=0.215, loss=172.016, backward_time=0.229, grad_norm=56.759, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.074, optim0_lr0=4.625e-05, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-27 01:49:56,749 (trainer:762) INFO: 2epoch:train:13001-13500batch: iter_time=1.556e-04, forward_time=0.209, loss_ctc=138.346, loss_att=193.083, acc=0.217, loss=176.662, backward_time=0.232, grad_norm=56.796, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.074, optim0_lr0=4.709e-05, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-02-27 01:57:33,250 (trainer:762) INFO: 2epoch:train:13501-14000batch: iter_time=1.487e-04, forward_time=0.207, loss_ctc=142.910, loss_att=202.412, acc=0.215, loss=184.561, backward_time=0.228, grad_norm=58.482, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.074, optim0_lr0=4.792e-05, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-27 02:04:59,661 (trainer:762) INFO: 2epoch:train:14001-14500batch: iter_time=1.471e-04, forward_time=0.209, loss_ctc=141.863, loss_att=203.245, acc=0.219, loss=184.831, backward_time=0.226, grad_norm=59.454, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.074, optim0_lr0=4.875e-05, train_time=0.893
+[ip-10-0-216-33:0/16] 2024-02-27 02:12:30,672 (trainer:762) INFO: 2epoch:train:14501-15000batch: iter_time=1.487e-04, forward_time=0.208, loss_ctc=136.369, loss_att=192.084, acc=0.228, loss=175.369, backward_time=0.227, grad_norm=54.535, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.074, optim0_lr0=4.959e-05, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-02-27 02:26:25,064 (trainer:361) INFO: 2epoch results: [train] iter_time=0.089, forward_time=0.207, loss_ctc=188.664, loss_att=206.489, acc=0.197, loss=201.142, backward_time=0.228, grad_norm=70.589, clip=100.000, loss_scale=2.847e+08, optim_step_time=0.074, optim0_lr0=3.750e-05, train_time=0.993, time=4 hours, 8 minutes and 34.41 seconds, total_count=30000, gpu_max_cached_mem_GB=36.488, [valid] loss_ctc=157.611, cer_ctc=0.761, loss_att=155.798, acc=0.181, cer=0.695, wer=1.000, loss=156.342, time=13 minutes and 37.86 seconds, total_count=4266, gpu_max_cached_mem_GB=36.488
+[ip-10-0-216-33:0/16] 2024-02-27 02:26:35,611 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-27 02:26:35,613 (trainer:290) INFO: 3/45epoch started. Estimated time to finish: 1 week, 20 hours and 2 minutes
+[ip-10-0-216-33:0/16] 2024-02-27 02:26:35,620 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 02:27:11,863 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 02:27:20,162 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0401d90>)
+[ip-10-0-216-33:0/16] 2024-02-27 02:27:20,162 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-27 02:27:20,167 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 02:41:13,923 (trainer:762) INFO: 3epoch:train:1-500batch: iter_time=0.826, forward_time=0.209, loss_ctc=135.945, loss_att=193.788, acc=0.222, loss=176.435, backward_time=0.236, grad_norm=57.775, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.076, optim0_lr0=5.377e-05, train_time=1.756
+[ip-10-0-216-33:0/16] 2024-02-27 02:48:48,093 (trainer:762) INFO: 3epoch:train:501-1000batch: iter_time=1.616e-04, forward_time=0.207, loss_ctc=128.326, loss_att=188.941, acc=0.228, loss=170.757, backward_time=0.231, grad_norm=52.671, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.076, optim0_lr0=6.127e-05, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 02:56:18,565 (trainer:762) INFO: 3epoch:train:1001-1500batch: iter_time=1.595e-04, forward_time=0.209, loss_ctc=136.890, loss_att=192.633, acc=0.229, loss=175.910, backward_time=0.235, grad_norm=60.697, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.076, optim0_lr0=6.877e-05, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-27 03:03:44,536 (trainer:762) INFO: 3epoch:train:1501-2000batch: iter_time=1.589e-04, forward_time=0.209, loss_ctc=127.544, loss_att=187.574, acc=0.235, loss=169.565, backward_time=0.231, grad_norm=54.367, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.076, optim0_lr0=7.627e-05, train_time=0.892
+[ip-10-0-216-33:0/16] 2024-02-27 03:11:13,737 (trainer:762) INFO: 3epoch:train:2001-2500batch: iter_time=1.620e-04, forward_time=0.209, loss_ctc=132.618, loss_att=194.409, acc=0.238, loss=175.872, backward_time=0.233, grad_norm=56.408, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.077, optim0_lr0=8.377e-05, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-02-27 03:18:43,505 (trainer:762) INFO: 3epoch:train:2501-3000batch: iter_time=1.600e-04, forward_time=0.209, loss_ctc=118.924, loss_att=177.962, acc=0.247, loss=160.251, backward_time=0.234, grad_norm=52.048, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.076, optim0_lr0=9.127e-05, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-02-27 03:26:15,388 (trainer:762) INFO: 3epoch:train:3001-3500batch: iter_time=1.655e-04, forward_time=0.210, loss_ctc=125.405, loss_att=190.989, acc=0.255, loss=171.313, backward_time=0.230, grad_norm=56.934, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.076, optim0_lr0=9.877e-05, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-27 03:33:43,188 (trainer:762) INFO: 3epoch:train:3501-4000batch: iter_time=1.567e-04, forward_time=0.208, loss_ctc=121.248, loss_att=174.649, acc=0.270, loss=158.629, backward_time=0.233, grad_norm=56.527, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.076, optim0_lr0=1.063e-04, train_time=0.895
+[ip-10-0-216-33:0/16] 2024-02-27 03:41:14,821 (trainer:762) INFO: 3epoch:train:4001-4500batch: iter_time=1.580e-04, forward_time=0.209, loss_ctc=123.202, loss_att=171.310, acc=0.282, loss=156.878, backward_time=0.234, grad_norm=63.597, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.076, optim0_lr0=1.138e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-27 03:48:45,888 (trainer:762) INFO: 3epoch:train:4501-5000batch: iter_time=1.564e-04, forward_time=0.209, loss_ctc=124.991, loss_att=172.746, acc=0.298, loss=158.420, backward_time=0.233, grad_norm=67.897, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.076, optim0_lr0=1.213e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-02-27 03:48:51,600 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 03:49:28,552 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 03:49:36,441 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfc83280>)
+[ip-10-0-216-33:0/16] 2024-02-27 03:49:36,441 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-27 03:49:36,446 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 04:03:51,603 (trainer:762) INFO: 3epoch:train:5001-5500batch: iter_time=0.858, forward_time=0.208, loss_ctc=121.894, loss_att=160.912, acc=0.307, loss=149.207, backward_time=0.230, grad_norm=69.924, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.075, optim0_lr0=1.288e-04, train_time=1.811
+[ip-10-0-216-33:0/16] 2024-02-27 04:11:20,815 (trainer:762) INFO: 3epoch:train:5501-6000batch: iter_time=1.456e-04, forward_time=0.209, loss_ctc=113.492, loss_att=150.486, acc=0.330, loss=139.388, backward_time=0.232, grad_norm=64.511, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.075, optim0_lr0=1.363e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-02-27 04:18:56,625 (trainer:762) INFO: 3epoch:train:6001-6500batch: iter_time=1.478e-04, forward_time=0.208, loss_ctc=121.604, loss_att=148.927, acc=0.343, loss=140.730, backward_time=0.231, grad_norm=69.876, clip=100.000, loss_scale=1.718e+10, optim_step_time=0.075, optim0_lr0=1.438e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-27 04:26:33,295 (trainer:762) INFO: 3epoch:train:6501-7000batch: iter_time=1.517e-04, forward_time=0.209, loss_ctc=111.241, loss_att=136.597, acc=0.371, loss=128.990, backward_time=0.233, grad_norm=63.932, clip=100.000, loss_scale=1.718e+10, optim_step_time=0.075, optim0_lr0=1.513e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-27 04:34:08,281 (trainer:762) INFO: 3epoch:train:7001-7500batch: iter_time=1.503e-04, forward_time=0.209, loss_ctc=114.210, loss_att=137.943, acc=0.390, loss=130.823, backward_time=0.231, grad_norm=69.867, clip=100.000, loss_scale=1.718e+10, optim_step_time=0.075, optim0_lr0=1.588e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-27 04:41:40,310 (trainer:762) INFO: 3epoch:train:7501-8000batch: iter_time=1.490e-04, forward_time=0.208, loss_ctc=102.651, loss_att=120.730, acc=0.405, loss=115.306, backward_time=0.233, grad_norm=64.007, clip=100.000, loss_scale=1.718e+10, optim_step_time=0.075, optim0_lr0=1.663e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-27 04:49:09,708 (trainer:762) INFO: 3epoch:train:8001-8500batch: iter_time=1.509e-04, forward_time=0.210, loss_ctc=107.805, loss_att=124.129, acc=0.434, loss=119.232, backward_time=0.230, grad_norm=67.397, clip=100.000, loss_scale=3.436e+10, optim_step_time=0.075, optim0_lr0=1.738e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-02-27 04:56:38,541 (trainer:762) INFO: 3epoch:train:8501-9000batch: iter_time=1.493e-04, forward_time=0.208, loss_ctc=104.751, loss_att=114.477, acc=0.439, loss=111.559, backward_time=0.233, grad_norm=66.045, clip=100.000, loss_scale=3.436e+10, optim_step_time=0.075, optim0_lr0=1.813e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-02-27 05:04:10,795 (trainer:762) INFO: 3epoch:train:9001-9500batch: iter_time=1.486e-04, forward_time=0.210, loss_ctc=105.849, loss_att=111.652, acc=0.451, loss=109.911, backward_time=0.233, grad_norm=64.720, clip=100.000, loss_scale=3.436e+10, optim_step_time=0.075, optim0_lr0=1.888e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-27 05:11:43,972 (trainer:762) INFO: 3epoch:train:9501-10000batch: iter_time=1.495e-04, forward_time=0.209, loss_ctc=107.498, loss_att=113.025, acc=0.464, loss=111.367, backward_time=0.231, grad_norm=64.861, clip=100.000, loss_scale=3.436e+10, optim_step_time=0.075, optim0_lr0=1.963e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-27 05:11:48,750 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 05:12:26,339 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 05:12:34,270 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e56dc6460>)
+[ip-10-0-216-33:0/16] 2024-02-27 05:12:34,270 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-27 05:12:34,275 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 05:26:48,665 (trainer:762) INFO: 3epoch:train:10001-10500batch: iter_time=0.870, forward_time=0.209, loss_ctc=104.848, loss_att=105.161, acc=0.468, loss=105.067, backward_time=0.232, grad_norm=60.487, clip=100.000, loss_scale=6.872e+10, optim_step_time=0.075, optim0_lr0=2.038e-04, train_time=1.809
+[ip-10-0-216-33:0/16] 2024-02-27 05:34:17,580 (trainer:762) INFO: 3epoch:train:10501-11000batch: iter_time=1.533e-04, forward_time=0.210, loss_ctc=97.527, loss_att=95.828, acc=0.494, loss=96.337, backward_time=0.236, grad_norm=54.260, clip=100.000, loss_scale=6.872e+10, optim_step_time=0.075, optim0_lr0=2.113e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-02-27 05:41:46,444 (trainer:762) INFO: 3epoch:train:11001-11500batch: iter_time=1.555e-04, forward_time=0.207, loss_ctc=105.042, loss_att=99.354, acc=0.492, loss=101.061, backward_time=0.229, grad_norm=59.958, clip=100.000, loss_scale=6.872e+10, optim_step_time=0.075, optim0_lr0=2.188e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-02-27 05:49:14,719 (trainer:762) INFO: 3epoch:train:11501-12000batch: iter_time=1.496e-04, forward_time=0.210, loss_ctc=96.238, loss_att=90.789, acc=0.515, loss=92.424, backward_time=0.229, grad_norm=51.190, clip=100.000, loss_scale=6.872e+10, optim_step_time=0.075, optim0_lr0=2.263e-04, train_time=0.896
+[ip-10-0-216-33:0/16] 2024-02-27 05:56:40,991 (trainer:762) INFO: 3epoch:train:12001-12500batch: iter_time=1.573e-04, forward_time=0.209, loss_ctc=99.778, loss_att=94.010, acc=0.527, loss=95.740, backward_time=0.229, grad_norm=54.650, clip=100.000, loss_scale=1.374e+11, optim_step_time=0.075, optim0_lr0=2.338e-04, train_time=0.892
+[ip-10-0-216-33:0/16] 2024-02-27 06:04:14,310 (trainer:762) INFO: 3epoch:train:12501-13000batch: iter_time=1.460e-04, forward_time=0.209, loss_ctc=89.097, loss_att=81.066, acc=0.540, loss=83.475, backward_time=0.231, grad_norm=49.938, clip=100.000, loss_scale=1.374e+11, optim_step_time=0.075, optim0_lr0=2.413e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-27 06:11:48,631 (trainer:762) INFO: 3epoch:train:13001-13500batch: iter_time=1.531e-04, forward_time=0.208, loss_ctc=94.148, loss_att=85.374, acc=0.559, loss=88.006, backward_time=0.234, grad_norm=52.360, clip=100.000, loss_scale=1.374e+11, optim_step_time=0.075, optim0_lr0=2.488e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 06:19:23,950 (trainer:762) INFO: 3epoch:train:13501-14000batch: iter_time=1.515e-04, forward_time=0.209, loss_ctc=91.811, loss_att=80.539, acc=0.555, loss=83.921, backward_time=0.233, grad_norm=49.335, clip=100.000, loss_scale=1.374e+11, optim_step_time=0.075, optim0_lr0=2.563e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-27 06:26:56,849 (trainer:762) INFO: 3epoch:train:14001-14500batch: iter_time=1.593e-04, forward_time=0.209, loss_ctc=94.035, loss_att=81.132, acc=0.557, loss=85.003, backward_time=0.236, grad_norm=52.272, clip=100.000, loss_scale=2.749e+11, optim_step_time=0.075, optim0_lr0=2.638e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-27 06:34:32,508 (trainer:762) INFO: 3epoch:train:14501-15000batch: iter_time=1.549e-04, forward_time=0.209, loss_ctc=95.925, loss_att=82.365, acc=0.564, loss=86.433, backward_time=0.231, grad_norm=52.475, clip=100.000, loss_scale=2.749e+11, optim_step_time=0.075, optim0_lr0=2.713e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-27 06:48:34,517 (trainer:361) INFO: 3epoch results: [train] iter_time=0.085, forward_time=0.209, loss_ctc=111.818, loss_att=135.317, acc=0.390, loss=128.267, backward_time=0.232, grad_norm=59.366, clip=100.000, loss_scale=5.469e+10, optim_step_time=0.076, optim0_lr0=1.625e-04, train_time=0.992, time=4 hours, 8 minutes and 15.84 seconds, total_count=45000, gpu_max_cached_mem_GB=36.488, [valid] loss_ctc=122.762, cer_ctc=0.641, loss_att=94.393, acc=0.368, cer=0.564, wer=1.000, loss=102.904, time=13 minutes and 42.76 seconds, total_count=6399, gpu_max_cached_mem_GB=36.488
+[ip-10-0-216-33:0/16] 2024-02-27 06:48:43,958 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-27 06:48:43,964 (trainer:290) INFO: 4/45epoch started. Estimated time to finish: 1 week, 15 hours and 35 minutes
+[ip-10-0-216-33:0/16] 2024-02-27 06:48:43,971 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 06:49:19,869 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 06:49:28,426 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ece1ea940>)
+[ip-10-0-216-33:0/16] 2024-02-27 06:49:28,426 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-27 06:49:28,431 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 07:03:19,395 (trainer:762) INFO: 4epoch:train:1-500batch: iter_time=0.838, forward_time=0.209, loss_ctc=89.908, loss_att=76.108, acc=0.581, loss=80.248, backward_time=0.231, grad_norm=47.409, clip=100.000, loss_scale=2.749e+11, optim_step_time=0.076, optim0_lr0=2.788e-04, train_time=1.751
+[ip-10-0-216-33:0/16] 2024-02-27 07:10:54,532 (trainer:762) INFO: 4epoch:train:501-1000batch: iter_time=1.619e-04, forward_time=0.208, loss_ctc=91.105, loss_att=76.163, acc=0.583, loss=80.645, backward_time=0.230, grad_norm=47.963, clip=100.000, loss_scale=2.749e+11, optim_step_time=0.076, optim0_lr0=2.863e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-27 07:18:30,465 (trainer:762) INFO: 4epoch:train:1001-1500batch: iter_time=1.577e-04, forward_time=0.209, loss_ctc=87.124, loss_att=72.397, acc=0.562, loss=76.815, backward_time=0.232, grad_norm=47.813, clip=100.000, loss_scale=5.498e+11, optim_step_time=0.077, optim0_lr0=2.938e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-27 07:26:01,344 (trainer:762) INFO: 4epoch:train:1501-2000batch: iter_time=1.586e-04, forward_time=0.209, loss_ctc=92.127, loss_att=74.093, acc=0.588, loss=79.504, backward_time=0.236, grad_norm=49.163, clip=100.000, loss_scale=5.498e+11, optim_step_time=0.076, optim0_lr0=3.013e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-27 07:33:38,071 (trainer:762) INFO: 4epoch:train:2001-2500batch: iter_time=1.631e-04, forward_time=0.208, loss_ctc=83.559, loss_att=67.890, acc=0.586, loss=72.591, backward_time=0.235, grad_norm=44.398, clip=100.000, loss_scale=5.498e+11, optim_step_time=0.077, optim0_lr0=3.088e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-27 07:41:12,036 (trainer:762) INFO: 4epoch:train:2501-3000batch: iter_time=1.622e-04, forward_time=0.210, loss_ctc=92.337, loss_att=73.519, acc=0.589, loss=79.164, backward_time=0.232, grad_norm=48.594, clip=100.000, loss_scale=5.498e+11, optim_step_time=0.077, optim0_lr0=3.163e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 07:48:45,675 (trainer:762) INFO: 4epoch:train:3001-3500batch: iter_time=1.652e-04, forward_time=0.208, loss_ctc=90.005, loss_att=71.692, acc=0.592, loss=77.186, backward_time=0.236, grad_norm=49.674, clip=100.000, loss_scale=1.100e+12, optim_step_time=0.076, optim0_lr0=3.238e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-27 07:56:17,666 (trainer:762) INFO: 4epoch:train:3501-4000batch: iter_time=1.619e-04, forward_time=0.209, loss_ctc=85.248, loss_att=68.407, acc=0.605, loss=73.459, backward_time=0.231, grad_norm=45.809, clip=100.000, loss_scale=1.100e+12, optim_step_time=0.076, optim0_lr0=3.313e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-27 08:03:52,552 (trainer:762) INFO: 4epoch:train:4001-4500batch: iter_time=1.608e-04, forward_time=0.209, loss_ctc=88.440, loss_att=69.571, acc=0.609, loss=75.232, backward_time=0.233, grad_norm=44.831, clip=100.000, loss_scale=1.100e+12, optim_step_time=0.076, optim0_lr0=3.388e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-27 08:11:22,675 (trainer:762) INFO: 4epoch:train:4501-5000batch: iter_time=1.637e-04, forward_time=0.209, loss_ctc=89.398, loss_att=69.407, acc=0.614, loss=75.404, backward_time=0.230, grad_norm=46.979, clip=100.000, loss_scale=1.100e+12, optim_step_time=0.077, optim0_lr0=3.463e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-02-27 08:11:27,356 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 08:12:03,891 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 08:12:11,898 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfbe7040>)
+[ip-10-0-216-33:0/16] 2024-02-27 08:12:11,898 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-27 08:12:11,903 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 08:26:27,989 (trainer:762) INFO: 4epoch:train:5001-5500batch: iter_time=0.859, forward_time=0.209, loss_ctc=83.607, loss_att=65.366, acc=0.622, loss=70.838, backward_time=0.231, grad_norm=43.176, clip=100.000, loss_scale=2.199e+12, optim_step_time=0.075, optim0_lr0=3.538e-04, train_time=1.810
+[ip-10-0-216-33:0/16] 2024-02-27 08:33:56,866 (trainer:762) INFO: 4epoch:train:5501-6000batch: iter_time=1.521e-04, forward_time=0.210, loss_ctc=85.301, loss_att=65.728, acc=0.623, loss=71.600, backward_time=0.238, grad_norm=44.177, clip=100.000, loss_scale=2.199e+12, optim_step_time=0.075, optim0_lr0=3.613e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-02-27 08:41:24,628 (trainer:762) INFO: 4epoch:train:6001-6500batch: iter_time=1.516e-04, forward_time=0.208, loss_ctc=81.639, loss_att=62.222, acc=0.604, loss=68.047, backward_time=0.234, grad_norm=46.103, clip=100.000, loss_scale=2.199e+12, optim_step_time=0.075, optim0_lr0=3.688e-04, train_time=0.895
+[ip-10-0-216-33:0/16] 2024-02-27 08:48:51,164 (trainer:762) INFO: 4epoch:train:6501-7000batch: iter_time=1.513e-04, forward_time=0.210, loss_ctc=86.938, loss_att=65.196, acc=0.624, loss=71.718, backward_time=0.230, grad_norm=42.860, clip=100.000, loss_scale=2.199e+12, optim_step_time=0.075, optim0_lr0=3.763e-04, train_time=0.893
+[ip-10-0-216-33:0/16] 2024-02-27 08:56:22,696 (trainer:762) INFO: 4epoch:train:7001-7500batch: iter_time=1.504e-04, forward_time=0.208, loss_ctc=79.158, loss_att=60.141, acc=0.621, loss=65.846, backward_time=0.235, grad_norm=41.406, clip=100.000, loss_scale=4.398e+12, optim_step_time=0.075, optim0_lr0=3.838e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-27 09:03:58,953 (trainer:762) INFO: 4epoch:train:7501-8000batch: iter_time=1.487e-04, forward_time=0.209, loss_ctc=87.307, loss_att=65.402, acc=0.623, loss=71.973, backward_time=0.235, grad_norm=46.315, clip=100.000, loss_scale=4.398e+12, optim_step_time=0.075, optim0_lr0=3.913e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-27 09:11:26,596 (trainer:762) INFO: 4epoch:train:8001-8500batch: iter_time=1.504e-04, forward_time=0.209, loss_ctc=84.155, loss_att=63.665, acc=0.626, loss=69.812, backward_time=0.226, grad_norm=43.179, clip=100.000, loss_scale=4.398e+12, optim_step_time=0.075, optim0_lr0=3.988e-04, train_time=0.895
+[ip-10-0-216-33:0/16] 2024-02-27 09:19:00,692 (trainer:762) INFO: 4epoch:train:8501-9000batch: iter_time=1.504e-04, forward_time=0.208, loss_ctc=80.802, loss_att=61.303, acc=0.636, loss=67.152, backward_time=0.225, grad_norm=42.446, clip=100.000, loss_scale=4.398e+12, optim_step_time=0.075, optim0_lr0=4.063e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 09:26:30,820 (trainer:762) INFO: 4epoch:train:9001-9500batch: iter_time=1.545e-04, forward_time=0.210, loss_ctc=83.775, loss_att=62.770, acc=0.638, loss=69.072, backward_time=0.233, grad_norm=42.727, clip=100.000, loss_scale=8.796e+12, optim_step_time=0.075, optim0_lr0=4.138e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-02-27 09:34:08,536 (trainer:762) INFO: 4epoch:train:9501-10000batch: iter_time=1.500e-04, forward_time=0.208, loss_ctc=85.144, loss_att=62.825, acc=0.641, loss=69.520, backward_time=0.224, grad_norm=43.320, clip=100.000, loss_scale=8.796e+12, optim_step_time=0.075, optim0_lr0=4.213e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-27 09:34:13,301 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 09:34:50,461 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 09:34:58,431 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e6fd553a0>)
+[ip-10-0-216-33:0/16] 2024-02-27 09:34:58,432 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-27 09:34:58,437 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 09:49:42,174 (trainer:762) INFO: 4epoch:train:10001-10500batch: iter_time=0.870, forward_time=0.209, loss_ctc=79.965, loss_att=59.814, acc=0.647, loss=65.859, backward_time=0.231, grad_norm=40.840, clip=100.000, loss_scale=8.796e+12, optim_step_time=0.075, optim0_lr0=4.288e-04, train_time=1.867
+[ip-10-0-216-33:0/16] 2024-02-27 09:57:24,148 (trainer:762) INFO: 4epoch:train:10501-11000batch: iter_time=1.492e-04, forward_time=0.209, loss_ctc=81.199, loss_att=60.307, acc=0.647, loss=66.575, backward_time=0.236, grad_norm=41.587, clip=100.000, loss_scale=8.796e+12, optim_step_time=0.075, optim0_lr0=4.363e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-02-27 10:04:54,705 (trainer:762) INFO: 4epoch:train:11001-11500batch: iter_time=1.587e-04, forward_time=0.208, loss_ctc=78.102, loss_att=57.441, acc=0.627, loss=63.639, backward_time=0.234, grad_norm=41.880, clip=100.000, loss_scale=1.759e+13, optim_step_time=0.076, optim0_lr0=4.438e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-27 10:12:30,589 (trainer:762) INFO: 4epoch:train:11501-12000batch: iter_time=1.512e-04, forward_time=0.210, loss_ctc=83.247, loss_att=60.272, acc=0.645, loss=67.165, backward_time=0.231, grad_norm=41.062, clip=100.000, loss_scale=1.759e+13, optim_step_time=0.075, optim0_lr0=4.513e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-27 10:20:03,359 (trainer:762) INFO: 4epoch:train:12001-12500batch: iter_time=1.574e-04, forward_time=0.208, loss_ctc=75.921, loss_att=55.904, acc=0.641, loss=61.909, backward_time=0.232, grad_norm=39.086, clip=100.000, loss_scale=1.759e+13, optim_step_time=0.075, optim0_lr0=4.588e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-27 10:27:39,323 (trainer:762) INFO: 4epoch:train:12501-13000batch: iter_time=1.506e-04, forward_time=0.210, loss_ctc=83.684, loss_att=61.040, acc=0.641, loss=67.834, backward_time=0.242, grad_norm=43.179, clip=100.000, loss_scale=1.759e+13, optim_step_time=0.075, optim0_lr0=4.663e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-27 10:35:13,836 (trainer:762) INFO: 4epoch:train:13001-13500batch: iter_time=1.480e-04, forward_time=0.208, loss_ctc=81.368, loss_att=59.631, acc=0.644, loss=66.152, backward_time=0.235, grad_norm=42.340, clip=100.000, loss_scale=3.518e+13, optim_step_time=0.075, optim0_lr0=4.738e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-27 10:42:47,918 (trainer:762) INFO: 4epoch:train:13501-14000batch: iter_time=1.479e-04, forward_time=0.209, loss_ctc=77.680, loss_att=57.072, acc=0.655, loss=63.254, backward_time=0.235, grad_norm=42.128, clip=100.000, loss_scale=3.518e+13, optim_step_time=0.075, optim0_lr0=4.813e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 10:50:21,924 (trainer:762) INFO: 4epoch:train:14001-14500batch: iter_time=1.538e-04, forward_time=0.210, loss_ctc=80.874, loss_att=59.017, acc=0.655, loss=65.574, backward_time=0.238, grad_norm=39.740, clip=100.000, loss_scale=3.518e+13, optim_step_time=0.075, optim0_lr0=4.888e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 10:57:55,813 (trainer:762) INFO: 4epoch:train:14501-15000batch: iter_time=1.558e-04, forward_time=0.209, loss_ctc=82.264, loss_att=59.042, acc=0.658, loss=66.009, backward_time=0.234, grad_norm=41.081, clip=100.000, loss_scale=3.518e+13, optim_step_time=0.075, optim0_lr0=4.963e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 11:11:55,600 (trainer:361) INFO: 4epoch results: [train] iter_time=0.086, forward_time=0.209, loss_ctc=84.379, loss_att=64.780, acc=0.621, loss=70.660, backward_time=0.233, grad_norm=44.042, clip=100.000, loss_scale=9.328e+12, optim_step_time=0.076, optim0_lr0=3.875e-04, train_time=0.997, time=4 hours, 9 minutes and 29.05 seconds, total_count=60000, gpu_max_cached_mem_GB=36.488, [valid] loss_ctc=105.137, cer_ctc=0.547, loss_att=71.042, acc=0.482, cer=0.478, wer=1.000, loss=81.270, time=13 minutes and 42.29 seconds, total_count=8532, gpu_max_cached_mem_GB=36.488
+[ip-10-0-216-33:0/16] 2024-02-27 11:12:04,705 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-27 11:12:04,710 (trainer:290) INFO: 5/45epoch started. Estimated time to finish: 1 week, 11 hours and 27 minutes
+[ip-10-0-216-33:0/16] 2024-02-27 11:12:04,718 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 11:12:40,711 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 11:12:49,091 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed00764f0>)
+[ip-10-0-216-33:0/16] 2024-02-27 11:12:49,092 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-27 11:12:49,096 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 11:26:47,490 (trainer:762) INFO: 5epoch:train:1-500batch: iter_time=0.842, forward_time=0.209, loss_ctc=78.099, loss_att=57.223, acc=0.649, loss=63.486, backward_time=0.235, grad_norm=39.053, clip=100.000, loss_scale=7.037e+13, optim_step_time=0.076, optim0_lr0=4.990e-04, train_time=1.765
+[ip-10-0-216-33:0/16] 2024-02-27 11:34:24,900 (trainer:762) INFO: 5epoch:train:501-1000batch: iter_time=1.574e-04, forward_time=0.209, loss_ctc=81.470, loss_att=59.317, acc=0.653, loss=65.963, backward_time=0.237, grad_norm=40.972, clip=100.000, loss_scale=7.037e+13, optim_step_time=0.077, optim0_lr0=4.969e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-27 11:42:04,851 (trainer:762) INFO: 5epoch:train:1001-1500batch: iter_time=1.622e-04, forward_time=0.209, loss_ctc=75.352, loss_att=54.506, acc=0.665, loss=60.760, backward_time=0.236, grad_norm=36.750, clip=100.000, loss_scale=7.037e+13, optim_step_time=0.076, optim0_lr0=4.949e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-02-27 11:49:35,868 (trainer:762) INFO: 5epoch:train:1501-2000batch: iter_time=1.559e-04, forward_time=0.209, loss_ctc=76.724, loss_att=55.314, acc=0.664, loss=61.737, backward_time=0.234, grad_norm=37.156, clip=100.000, loss_scale=7.037e+13, optim_step_time=0.076, optim0_lr0=4.929e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-02-27 11:57:11,605 (trainer:762) INFO: 5epoch:train:2001-2500batch: iter_time=1.596e-04, forward_time=0.208, loss_ctc=76.989, loss_att=55.658, acc=0.659, loss=62.057, backward_time=0.232, grad_norm=38.117, clip=100.000, loss_scale=1.407e+14, optim_step_time=0.076, optim0_lr0=4.909e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-27 12:04:46,983 (trainer:762) INFO: 5epoch:train:2501-3000batch: iter_time=1.605e-04, forward_time=0.209, loss_ctc=78.420, loss_att=55.917, acc=0.657, loss=62.668, backward_time=0.234, grad_norm=39.851, clip=100.000, loss_scale=1.407e+14, optim_step_time=0.076, optim0_lr0=4.889e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-27 12:12:25,549 (trainer:762) INFO: 5epoch:train:3001-3500batch: iter_time=1.553e-04, forward_time=0.209, loss_ctc=78.645, loss_att=55.853, acc=0.674, loss=62.690, backward_time=0.231, grad_norm=37.075, clip=100.000, loss_scale=1.407e+14, optim_step_time=0.076, optim0_lr0=4.870e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-27 12:20:01,183 (trainer:762) INFO: 5epoch:train:3501-4000batch: iter_time=1.587e-04, forward_time=0.208, loss_ctc=76.079, loss_att=54.128, acc=0.661, loss=60.714, backward_time=0.235, grad_norm=36.804, clip=100.000, loss_scale=1.407e+14, optim_step_time=0.076, optim0_lr0=4.851e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-27 12:27:39,654 (trainer:762) INFO: 5epoch:train:4001-4500batch: iter_time=1.593e-04, forward_time=0.210, loss_ctc=77.331, loss_att=54.769, acc=0.669, loss=61.537, backward_time=0.235, grad_norm=37.711, clip=100.000, loss_scale=2.815e+14, optim_step_time=0.076, optim0_lr0=4.832e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-27 12:35:13,140 (trainer:762) INFO: 5epoch:train:4501-5000batch: iter_time=1.565e-04, forward_time=0.208, loss_ctc=78.513, loss_att=55.151, acc=0.659, loss=62.159, backward_time=0.236, grad_norm=37.515, clip=100.000, loss_scale=2.815e+14, optim_step_time=0.076, optim0_lr0=4.813e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-27 12:35:17,971 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 12:35:54,485 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 12:36:03,208 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecdbb7190>)
+[ip-10-0-216-33:0/16] 2024-02-27 12:36:03,209 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-27 12:36:03,213 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 12:50:19,737 (trainer:762) INFO: 5epoch:train:5001-5500batch: iter_time=0.862, forward_time=0.208, loss_ctc=71.945, loss_att=51.314, acc=0.677, loss=57.503, backward_time=0.231, grad_norm=34.374, clip=100.000, loss_scale=2.815e+14, optim_step_time=0.075, optim0_lr0=4.795e-04, train_time=1.813
+[ip-10-0-216-33:0/16] 2024-02-27 12:57:45,762 (trainer:762) INFO: 5epoch:train:5501-6000batch: iter_time=1.489e-04, forward_time=0.209, loss_ctc=74.887, loss_att=53.196, acc=0.681, loss=59.703, backward_time=0.234, grad_norm=35.571, clip=100.000, loss_scale=2.815e+14, optim_step_time=0.075, optim0_lr0=4.776e-04, train_time=0.892
+[ip-10-0-216-33:0/16] 2024-02-27 13:05:23,737 (trainer:762) INFO: 5epoch:train:6001-6500batch: iter_time=1.485e-04, forward_time=0.208, loss_ctc=69.972, loss_att=49.428, acc=0.690, loss=55.591, backward_time=0.236, grad_norm=33.388, clip=100.000, loss_scale=5.629e+14, optim_step_time=0.075, optim0_lr0=4.758e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-27 13:12:56,462 (trainer:762) INFO: 5epoch:train:6501-7000batch: iter_time=1.454e-04, forward_time=0.209, loss_ctc=71.738, loss_att=50.553, acc=0.689, loss=56.908, backward_time=0.236, grad_norm=33.246, clip=100.000, loss_scale=5.629e+14, optim_step_time=0.075, optim0_lr0=4.740e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-27 13:20:31,269 (trainer:762) INFO: 5epoch:train:7001-7500batch: iter_time=1.525e-04, forward_time=0.208, loss_ctc=71.918, loss_att=50.815, acc=0.685, loss=57.146, backward_time=0.237, grad_norm=34.711, clip=100.000, loss_scale=5.629e+14, optim_step_time=0.075, optim0_lr0=4.723e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-27 13:28:07,320 (trainer:762) INFO: 5epoch:train:7501-8000batch: iter_time=1.455e-04, forward_time=0.208, loss_ctc=73.038, loss_att=50.965, acc=0.682, loss=57.587, backward_time=0.237, grad_norm=33.418, clip=100.000, loss_scale=5.629e+14, optim_step_time=0.075, optim0_lr0=4.705e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-27 13:35:37,698 (trainer:762) INFO: 5epoch:train:8001-8500batch: iter_time=1.480e-04, forward_time=0.210, loss_ctc=73.700, loss_att=51.441, acc=0.696, loss=58.118, backward_time=0.237, grad_norm=33.600, clip=100.000, loss_scale=1.126e+15, optim_step_time=0.075, optim0_lr0=4.688e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-27 13:43:12,690 (trainer:762) INFO: 5epoch:train:8501-9000batch: iter_time=1.476e-04, forward_time=0.207, loss_ctc=70.588, loss_att=49.394, acc=0.685, loss=55.752, backward_time=0.237, grad_norm=33.395, clip=100.000, loss_scale=1.126e+15, optim_step_time=0.075, optim0_lr0=4.671e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-27 13:50:51,586 (trainer:762) INFO: 5epoch:train:9001-9500batch: iter_time=1.498e-04, forward_time=0.209, loss_ctc=72.179, loss_att=50.129, acc=0.691, loss=56.744, backward_time=0.237, grad_norm=34.855, clip=100.000, loss_scale=1.126e+15, optim_step_time=0.075, optim0_lr0=4.654e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-02-27 13:58:20,527 (trainer:762) INFO: 5epoch:train:9501-10000batch: iter_time=1.435e-04, forward_time=0.208, loss_ctc=73.841, loss_att=50.899, acc=0.681, loss=57.781, backward_time=0.232, grad_norm=34.180, clip=100.000, loss_scale=1.126e+15, optim_step_time=0.075, optim0_lr0=4.637e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-02-27 13:58:25,577 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 13:59:03,193 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 13:59:11,421 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f64bc0d93d0>)
+[ip-10-0-216-33:0/16] 2024-02-27 13:59:11,422 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-27 13:59:11,426 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 14:15:57,202 (trainer:762) INFO: 5epoch:train:10001-10500batch: iter_time=1.173, forward_time=0.209, loss_ctc=67.739, loss_att=47.632, acc=0.697, loss=53.664, backward_time=0.231, grad_norm=35.429, clip=100.000, loss_scale=2.252e+15, optim_step_time=0.075, optim0_lr0=4.621e-04, train_time=2.113
+[ip-10-0-216-33:0/16] 2024-02-27 14:23:31,557 (trainer:762) INFO: 5epoch:train:10501-11000batch: iter_time=1.448e-04, forward_time=0.210, loss_ctc=70.352, loss_att=49.517, acc=0.700, loss=55.767, backward_time=0.236, grad_norm=32.618, clip=100.000, loss_scale=2.252e+15, optim_step_time=0.075, optim0_lr0=4.604e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 14:30:57,929 (trainer:762) INFO: 5epoch:train:11001-11500batch: iter_time=1.461e-04, forward_time=0.209, loss_ctc=66.046, loss_att=46.009, acc=0.709, loss=52.020, backward_time=0.235, grad_norm=30.961, clip=100.000, loss_scale=2.252e+15, optim_step_time=0.075, optim0_lr0=4.588e-04, train_time=0.892
+[ip-10-0-216-33:0/16] 2024-02-27 14:38:31,329 (trainer:762) INFO: 5epoch:train:11501-12000batch: iter_time=1.475e-04, forward_time=0.210, loss_ctc=67.486, loss_att=47.029, acc=0.707, loss=53.166, backward_time=0.233, grad_norm=31.119, clip=100.000, loss_scale=2.252e+15, optim_step_time=0.075, optim0_lr0=4.572e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-27 14:45:57,128 (trainer:762) INFO: 5epoch:train:12001-12500batch: iter_time=1.483e-04, forward_time=0.209, loss_ctc=68.263, loss_att=47.686, acc=0.702, loss=53.859, backward_time=0.231, grad_norm=33.056, clip=100.000, loss_scale=4.504e+15, optim_step_time=0.075, optim0_lr0=4.556e-04, train_time=0.891
+[ip-10-0-216-33:0/16] 2024-02-27 14:53:31,370 (trainer:762) INFO: 5epoch:train:12501-13000batch: iter_time=1.474e-04, forward_time=0.209, loss_ctc=69.519, loss_att=47.872, acc=0.698, loss=54.366, backward_time=0.232, grad_norm=31.249, clip=100.000, loss_scale=4.504e+15, optim_step_time=0.075, optim0_lr0=4.541e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 15:01:06,605 (trainer:762) INFO: 5epoch:train:13001-13500batch: iter_time=1.480e-04, forward_time=0.209, loss_ctc=70.029, loss_att=48.234, acc=0.712, loss=54.772, backward_time=0.237, grad_norm=31.055, clip=100.000, loss_scale=4.504e+15, optim_step_time=0.075, optim0_lr0=4.525e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-27 15:08:43,091 (trainer:762) INFO: 5epoch:train:13501-14000batch: iter_time=1.515e-04, forward_time=0.210, loss_ctc=67.284, loss_att=46.431, acc=0.701, loss=52.687, backward_time=0.240, grad_norm=31.950, clip=100.000, loss_scale=4.504e+15, optim_step_time=0.075, optim0_lr0=4.510e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-27 15:16:13,577 (trainer:762) INFO: 5epoch:train:14001-14500batch: iter_time=1.565e-04, forward_time=0.209, loss_ctc=68.770, loss_att=47.324, acc=0.707, loss=53.757, backward_time=0.237, grad_norm=33.846, clip=100.000, loss_scale=9.007e+15, optim_step_time=0.075, optim0_lr0=4.495e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-27 15:23:48,672 (trainer:762) INFO: 5epoch:train:14501-15000batch: iter_time=1.540e-04, forward_time=0.210, loss_ctc=70.077, loss_att=47.984, acc=0.697, loss=54.612, backward_time=0.239, grad_norm=33.259, clip=100.000, loss_scale=9.007e+15, optim_step_time=0.075, optim0_lr0=4.480e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-27 15:37:51,128 (trainer:361) INFO: 5epoch results: [train] iter_time=0.096, forward_time=0.209, loss_ctc=72.900, loss_att=51.390, acc=0.683, loss=57.843, backward_time=0.235, grad_norm=34.876, clip=100.000, loss_scale=1.792e+15, optim_step_time=0.076, optim0_lr0=4.721e-04, train_time=1.007, time=4 hours, 11 minutes and 59.87 seconds, total_count=75000, gpu_max_cached_mem_GB=36.488, [valid] loss_ctc=83.757, cer_ctc=0.462, loss_att=52.534, acc=0.573, cer=0.420, wer=1.000, loss=61.901, time=13 minutes and 46.25 seconds, total_count=10665, gpu_max_cached_mem_GB=36.488
+[ip-10-0-216-33:0/16] 2024-02-27 15:38:00,792 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-27 15:38:00,799 (trainer:290) INFO: 6/45epoch started. Estimated time to finish: 1 week, 7 hours and 38 minutes
+[ip-10-0-216-33:0/16] 2024-02-27 15:38:00,807 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 15:38:37,182 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 15:38:45,275 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf1e6d30>)
+[ip-10-0-216-33:0/16] 2024-02-27 15:38:45,275 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-27 15:38:45,279 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 15:52:45,687 (trainer:762) INFO: 6epoch:train:1-500batch: iter_time=0.845, forward_time=0.208, loss_ctc=67.586, loss_att=46.720, acc=0.705, loss=52.980, backward_time=0.231, grad_norm=33.341, clip=100.000, loss_scale=9.007e+15, optim_step_time=0.076, optim0_lr0=4.465e-04, train_time=1.769
+[ip-10-0-216-33:0/16] 2024-02-27 16:00:23,649 (trainer:762) INFO: 6epoch:train:501-1000batch: iter_time=1.580e-04, forward_time=0.209, loss_ctc=70.295, loss_att=48.599, acc=0.707, loss=55.108, backward_time=0.232, grad_norm=32.098, clip=100.000, loss_scale=9.007e+15, optim_step_time=0.076, optim0_lr0=4.450e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-27 16:08:01,595 (trainer:762) INFO: 6epoch:train:1001-1500batch: iter_time=1.583e-04, forward_time=0.209, loss_ctc=68.225, loss_att=46.862, acc=0.704, loss=53.271, backward_time=0.224, grad_norm=32.100, clip=100.000, loss_scale=1.801e+16, optim_step_time=0.076, optim0_lr0=4.435e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-27 16:15:37,945 (trainer:762) INFO: 6epoch:train:1501-2000batch: iter_time=1.614e-04, forward_time=0.208, loss_ctc=63.373, loss_att=43.921, acc=0.725, loss=49.757, backward_time=0.232, grad_norm=29.323, clip=100.000, loss_scale=1.801e+16, optim_step_time=0.076, optim0_lr0=4.421e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-27 16:23:14,089 (trainer:762) INFO: 6epoch:train:2001-2500batch: iter_time=1.565e-04, forward_time=0.209, loss_ctc=61.601, loss_att=43.094, acc=0.723, loss=48.647, backward_time=0.232, grad_norm=30.103, clip=100.000, loss_scale=1.801e+16, optim_step_time=0.076, optim0_lr0=4.406e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-27 16:30:50,219 (trainer:762) INFO: 6epoch:train:2501-3000batch: iter_time=1.576e-04, forward_time=0.209, loss_ctc=65.737, loss_att=45.271, acc=0.712, loss=51.411, backward_time=0.236, grad_norm=30.929, clip=100.000, loss_scale=1.801e+16, optim_step_time=0.076, optim0_lr0=4.392e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-27 16:38:29,122 (trainer:762) INFO: 6epoch:train:3001-3500batch: iter_time=1.471e-04, forward_time=0.208, loss_ctc=64.790, loss_att=44.976, acc=0.718, loss=50.920, backward_time=0.232, grad_norm=31.579, clip=100.000, loss_scale=3.603e+16, optim_step_time=0.076, optim0_lr0=4.378e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-02-27 16:46:01,558 (trainer:762) INFO: 6epoch:train:3501-4000batch: iter_time=1.464e-04, forward_time=0.209, loss_ctc=68.739, loss_att=46.938, acc=0.718, loss=53.478, backward_time=0.237, grad_norm=31.106, clip=100.000, loss_scale=3.603e+16, optim_step_time=0.075, optim0_lr0=4.364e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-27 16:53:34,499 (trainer:762) INFO: 6epoch:train:4001-4500batch: iter_time=1.471e-04, forward_time=0.208, loss_ctc=61.897, loss_att=42.818, acc=0.724, loss=48.542, backward_time=0.237, grad_norm=30.098, clip=100.000, loss_scale=3.603e+16, optim_step_time=0.075, optim0_lr0=4.351e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-27 17:01:20,979 (trainer:762) INFO: 6epoch:train:4501-5000batch: iter_time=0.003, forward_time=0.209, loss_ctc=64.274, loss_att=44.586, acc=0.725, loss=50.493, backward_time=0.235, grad_norm=32.178, clip=100.000, loss_scale=3.603e+16, optim_step_time=0.075, optim0_lr0=4.337e-04, train_time=0.933
+[ip-10-0-216-33:0/16] 2024-02-27 17:01:25,688 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 17:02:02,483 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 17:02:10,505 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0401d90>)
+[ip-10-0-216-33:0/16] 2024-02-27 17:02:10,506 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-27 17:02:10,510 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 17:17:00,706 (trainer:762) INFO: 6epoch:train:5001-5500batch: iter_time=0.888, forward_time=0.209, loss_ctc=65.050, loss_att=44.363, acc=0.718, loss=50.569, backward_time=0.239, grad_norm=32.010, clip=100.000, loss_scale=7.206e+16, optim_step_time=0.075, optim0_lr0=4.323e-04, train_time=1.879
+[ip-10-0-216-33:0/16] 2024-02-27 17:24:40,574 (trainer:762) INFO: 6epoch:train:5501-6000batch: iter_time=1.535e-04, forward_time=0.209, loss_ctc=67.182, loss_att=46.296, acc=0.719, loss=52.562, backward_time=0.233, grad_norm=32.510, clip=100.000, loss_scale=7.206e+16, optim_step_time=0.075, optim0_lr0=4.310e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-27 17:32:08,392 (trainer:762) INFO: 6epoch:train:6001-6500batch: iter_time=1.501e-04, forward_time=0.208, loss_ctc=64.818, loss_att=44.459, acc=0.717, loss=50.567, backward_time=0.237, grad_norm=31.276, clip=100.000, loss_scale=7.206e+16, optim_step_time=0.075, optim0_lr0=4.297e-04, train_time=0.895
+[ip-10-0-216-33:0/16] 2024-02-27 17:39:48,309 (trainer:762) INFO: 6epoch:train:6501-7000batch: iter_time=1.505e-04, forward_time=0.209, loss_ctc=61.070, loss_att=42.062, acc=0.735, loss=47.764, backward_time=0.237, grad_norm=35.136, clip=100.000, loss_scale=7.206e+16, optim_step_time=0.075, optim0_lr0=4.283e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-02-27 17:47:18,171 (trainer:762) INFO: 6epoch:train:7001-7500batch: iter_time=1.489e-04, forward_time=0.209, loss_ctc=59.201, loss_att=41.252, acc=0.734, loss=46.636, backward_time=0.232, grad_norm=28.837, clip=100.000, loss_scale=1.441e+17, optim_step_time=0.075, optim0_lr0=4.270e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-02-27 17:54:50,177 (trainer:762) INFO: 6epoch:train:7501-8000batch: iter_time=1.515e-04, forward_time=0.210, loss_ctc=63.383, loss_att=43.480, acc=0.722, loss=49.451, backward_time=0.240, grad_norm=30.185, clip=100.000, loss_scale=1.441e+17, optim_step_time=0.075, optim0_lr0=4.258e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-27 18:02:29,623 (trainer:762) INFO: 6epoch:train:8001-8500batch: iter_time=1.476e-04, forward_time=0.208, loss_ctc=62.483, loss_att=43.059, acc=0.728, loss=48.886, backward_time=0.232, grad_norm=30.214, clip=100.000, loss_scale=1.441e+17, optim_step_time=0.075, optim0_lr0=4.245e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-27 18:09:58,035 (trainer:762) INFO: 6epoch:train:8501-9000batch: iter_time=1.477e-04, forward_time=0.210, loss_ctc=66.005, loss_att=45.109, acc=0.728, loss=51.378, backward_time=0.235, grad_norm=32.121, clip=100.000, loss_scale=1.441e+17, optim_step_time=0.075, optim0_lr0=4.232e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-02-27 18:17:34,801 (trainer:762) INFO: 6epoch:train:9001-9500batch: iter_time=1.494e-04, forward_time=0.208, loss_ctc=59.733, loss_att=41.128, acc=0.733, loss=46.709, backward_time=0.237, grad_norm=31.719, clip=100.000, loss_scale=2.882e+17, optim_step_time=0.075, optim0_lr0=4.219e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-27 18:25:09,337 (trainer:762) INFO: 6epoch:train:9501-10000batch: iter_time=1.470e-04, forward_time=0.209, loss_ctc=62.050, loss_att=42.895, acc=0.735, loss=48.642, backward_time=0.241, grad_norm=30.371, clip=100.000, loss_scale=2.882e+17, optim_step_time=0.075, optim0_lr0=4.207e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-27 18:25:14,560 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 18:25:51,816 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 18:26:00,018 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e744ebee0>)
+[ip-10-0-216-33:0/16] 2024-02-27 18:26:00,018 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-27 18:26:00,023 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 18:40:36,691 (trainer:762) INFO: 6epoch:train:10001-10500batch: iter_time=0.906, forward_time=0.209, loss_ctc=62.941, loss_att=42.752, acc=0.727, loss=48.809, backward_time=0.233, grad_norm=31.245, clip=100.000, loss_scale=2.882e+17, optim_step_time=0.075, optim0_lr0=4.195e-04, train_time=1.854
+[ip-10-0-216-33:0/16] 2024-02-27 18:48:14,742 (trainer:762) INFO: 6epoch:train:10501-11000batch: iter_time=1.533e-04, forward_time=0.209, loss_ctc=64.964, loss_att=44.693, acc=0.729, loss=50.774, backward_time=0.235, grad_norm=31.216, clip=100.000, loss_scale=2.882e+17, optim_step_time=0.075, optim0_lr0=4.182e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-27 18:55:49,684 (trainer:762) INFO: 6epoch:train:11001-11500batch: iter_time=1.575e-04, forward_time=0.210, loss_ctc=62.915, loss_att=43.124, acc=0.726, loss=49.061, backward_time=0.234, grad_norm=30.338, clip=100.000, loss_scale=5.765e+17, optim_step_time=0.075, optim0_lr0=4.170e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-27 19:03:23,605 (trainer:762) INFO: 6epoch:train:11501-12000batch: iter_time=1.470e-04, forward_time=0.209, loss_ctc=58.921, loss_att=40.451, acc=0.745, loss=45.992, backward_time=0.237, grad_norm=28.504, clip=100.000, loss_scale=5.765e+17, optim_step_time=0.075, optim0_lr0=4.158e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 19:11:01,160 (trainer:762) INFO: 6epoch:train:12001-12500batch: iter_time=1.536e-04, forward_time=0.210, loss_ctc=57.176, loss_att=39.822, acc=0.742, loss=45.028, backward_time=0.238, grad_norm=29.198, clip=100.000, loss_scale=5.765e+17, optim_step_time=0.075, optim0_lr0=4.146e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-27 19:18:32,432 (trainer:762) INFO: 6epoch:train:12501-13000batch: iter_time=1.523e-04, forward_time=0.208, loss_ctc=61.335, loss_att=41.977, acc=0.731, loss=47.784, backward_time=0.241, grad_norm=30.590, clip=100.000, loss_scale=5.765e+17, optim_step_time=0.075, optim0_lr0=4.134e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-02-27 19:26:11,918 (trainer:762) INFO: 6epoch:train:13001-13500batch: iter_time=1.573e-04, forward_time=0.210, loss_ctc=60.353, loss_att=41.645, acc=0.738, loss=47.257, backward_time=0.230, grad_norm=29.761, clip=100.000, loss_scale=1.153e+18, optim_step_time=0.075, optim0_lr0=4.123e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-27 19:33:43,453 (trainer:762) INFO: 6epoch:train:13501-14000batch: iter_time=1.522e-04, forward_time=0.209, loss_ctc=64.273, loss_att=43.639, acc=0.736, loss=49.829, backward_time=0.235, grad_norm=30.641, clip=100.000, loss_scale=1.153e+18, optim_step_time=0.075, optim0_lr0=4.111e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-27 19:41:18,927 (trainer:762) INFO: 6epoch:train:14001-14500batch: iter_time=1.514e-04, forward_time=0.210, loss_ctc=57.865, loss_att=39.840, acc=0.740, loss=45.248, backward_time=0.236, grad_norm=28.762, clip=100.000, loss_scale=1.153e+18, optim_step_time=0.075, optim0_lr0=4.100e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-27 19:48:47,743 (trainer:762) INFO: 6epoch:train:14501-15000batch: iter_time=1.483e-04, forward_time=0.209, loss_ctc=59.996, loss_att=41.406, acc=0.743, loss=46.983, backward_time=0.240, grad_norm=29.913, clip=100.000, loss_scale=1.153e+18, optim_step_time=0.075, optim0_lr0=4.088e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-02-27 20:07:33,786 (trainer:361) INFO: 6epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=63.274, loss_att=43.575, acc=0.726, loss=49.484, backward_time=0.235, grad_norm=30.913, clip=100.000, loss_scale=3.056e+17, optim_step_time=0.075, optim0_lr0=4.268e-04, train_time=1.003, time=4 hours, 11 minutes and 2.93 seconds, total_count=90000, gpu_max_cached_mem_GB=36.488, [valid] loss_ctc=69.057, cer_ctc=0.374, loss_att=47.036, acc=0.613, cer=0.398, wer=1.000, loss=53.642, time=18 minutes and 29.77 seconds, total_count=12798, gpu_max_cached_mem_GB=36.488
+[ip-10-0-216-33:0/16] 2024-02-27 20:07:43,212 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-27 20:07:43,231 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/1epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-27 20:07:43,232 (trainer:290) INFO: 7/45epoch started. Estimated time to finish: 1 week, 4 hours and 3 minutes
+[ip-10-0-216-33:0/16] 2024-02-27 20:07:43,240 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 20:08:19,191 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 20:08:27,606 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5eceadd1f0>)
+[ip-10-0-216-33:0/16] 2024-02-27 20:08:27,606 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-27 20:08:27,611 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 20:22:23,156 (trainer:762) INFO: 7epoch:train:1-500batch: iter_time=0.839, forward_time=0.210, loss_ctc=62.995, loss_att=43.472, acc=0.738, loss=49.329, backward_time=0.232, grad_norm=39.714, clip=100.000, loss_scale=2.306e+18, optim_step_time=0.075, optim0_lr0=4.077e-04, train_time=1.760
+[ip-10-0-216-33:0/16] 2024-02-27 20:30:00,005 (trainer:762) INFO: 7epoch:train:501-1000batch: iter_time=1.529e-04, forward_time=0.208, loss_ctc=60.657, loss_att=41.588, acc=0.744, loss=47.309, backward_time=0.237, grad_norm=29.663, clip=100.000, loss_scale=2.306e+18, optim_step_time=0.075, optim0_lr0=4.066e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-27 20:37:39,562 (trainer:762) INFO: 7epoch:train:1001-1500batch: iter_time=1.538e-04, forward_time=0.210, loss_ctc=57.001, loss_att=39.554, acc=0.756, loss=44.788, backward_time=0.240, grad_norm=28.101, clip=100.000, loss_scale=2.306e+18, optim_step_time=0.075, optim0_lr0=4.054e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-27 20:45:13,804 (trainer:762) INFO: 7epoch:train:1501-2000batch: iter_time=1.519e-04, forward_time=0.209, loss_ctc=57.672, loss_att=40.320, acc=0.746, loss=45.526, backward_time=0.238, grad_norm=28.277, clip=100.000, loss_scale=2.306e+18, optim_step_time=0.075, optim0_lr0=4.043e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 20:52:50,037 (trainer:762) INFO: 7epoch:train:2001-2500batch: iter_time=1.504e-04, forward_time=0.208, loss_ctc=63.232, loss_att=43.016, acc=0.738, loss=49.081, backward_time=0.237, grad_norm=33.016, clip=100.000, loss_scale=4.612e+18, optim_step_time=0.075, optim0_lr0=4.032e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-27 21:00:26,890 (trainer:762) INFO: 7epoch:train:2501-3000batch: iter_time=1.514e-04, forward_time=0.209, loss_ctc=58.012, loss_att=39.809, acc=0.744, loss=45.270, backward_time=0.239, grad_norm=29.078, clip=100.000, loss_scale=4.612e+18, optim_step_time=0.075, optim0_lr0=4.021e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-27 21:08:07,561 (trainer:762) INFO: 7epoch:train:3001-3500batch: iter_time=0.004, forward_time=0.209, loss_ctc=60.592, loss_att=41.807, acc=0.737, loss=47.443, backward_time=0.240, grad_norm=30.753, clip=100.000, loss_scale=4.612e+18, optim_step_time=0.075, optim0_lr0=4.011e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-02-27 21:15:42,499 (trainer:762) INFO: 7epoch:train:3501-4000batch: iter_time=1.508e-04, forward_time=0.208, loss_ctc=62.557, loss_att=42.758, acc=0.740, loss=48.698, backward_time=0.237, grad_norm=31.295, clip=100.000, loss_scale=4.612e+18, optim_step_time=0.075, optim0_lr0=4.000e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-27 21:23:23,811 (trainer:762) INFO: 7epoch:train:4001-4500batch: iter_time=1.492e-04, forward_time=0.209, loss_ctc=59.029, loss_att=40.827, acc=0.753, loss=46.288, backward_time=0.236, grad_norm=29.177, clip=100.000, loss_scale=9.223e+18, optim_step_time=0.075, optim0_lr0=3.989e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-02-27 21:30:59,725 (trainer:762) INFO: 7epoch:train:4501-5000batch: iter_time=1.531e-04, forward_time=0.208, loss_ctc=59.757, loss_att=40.638, acc=0.744, loss=46.374, backward_time=0.227, grad_norm=29.281, clip=100.000, loss_scale=9.223e+18, optim_step_time=0.076, optim0_lr0=3.979e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-27 21:31:05,439 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 21:31:41,945 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 21:31:50,434 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ea105f130>)
+[ip-10-0-216-33:0/16] 2024-02-27 21:31:50,434 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-27 21:31:50,439 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 21:46:06,579 (trainer:762) INFO: 7epoch:train:5001-5500batch: iter_time=0.882, forward_time=0.209, loss_ctc=60.957, loss_att=42.119, acc=0.746, loss=47.770, backward_time=0.235, grad_norm=29.848, clip=100.000, loss_scale=9.223e+18, optim_step_time=0.075, optim0_lr0=3.968e-04, train_time=1.813
+[ip-10-0-216-33:0/16] 2024-02-27 21:53:38,177 (trainer:762) INFO: 7epoch:train:5501-6000batch: iter_time=1.494e-04, forward_time=0.210, loss_ctc=58.646, loss_att=40.111, acc=0.751, loss=45.671, backward_time=0.231, grad_norm=32.839, clip=100.000, loss_scale=9.223e+18, optim_step_time=0.075, optim0_lr0=3.958e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-27 22:01:15,502 (trainer:762) INFO: 7epoch:train:6001-6500batch: iter_time=0.008, forward_time=0.210, loss_ctc=55.707, loss_att=38.482, acc=0.763, loss=43.649, backward_time=0.233, grad_norm=28.117, clip=100.000, loss_scale=1.845e+19, optim_step_time=0.075, optim0_lr0=3.948e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-27 22:08:46,811 (trainer:762) INFO: 7epoch:train:6501-7000batch: iter_time=1.456e-04, forward_time=0.208, loss_ctc=56.054, loss_att=39.295, acc=0.751, loss=44.323, backward_time=0.237, grad_norm=28.317, clip=100.000, loss_scale=1.845e+19, optim_step_time=0.075, optim0_lr0=3.937e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-02-27 22:16:21,198 (trainer:762) INFO: 7epoch:train:7001-7500batch: iter_time=1.473e-04, forward_time=0.209, loss_ctc=60.521, loss_att=41.714, acc=0.744, loss=47.356, backward_time=0.232, grad_norm=34.682, clip=100.000, loss_scale=1.845e+19, optim_step_time=0.075, optim0_lr0=3.927e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-27 22:23:56,377 (trainer:762) INFO: 7epoch:train:7501-8000batch: iter_time=1.477e-04, forward_time=0.208, loss_ctc=56.190, loss_att=38.820, acc=0.749, loss=44.031, backward_time=0.232, grad_norm=29.071, clip=100.000, loss_scale=1.845e+19, optim_step_time=0.075, optim0_lr0=3.917e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-27 22:31:36,304 (trainer:762) INFO: 7epoch:train:8001-8500batch: iter_time=1.470e-04, forward_time=0.209, loss_ctc=58.922, loss_att=40.494, acc=0.744, loss=46.023, backward_time=0.233, grad_norm=29.829, clip=100.000, loss_scale=3.689e+19, optim_step_time=0.075, optim0_lr0=3.907e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-02-27 22:39:08,003 (trainer:762) INFO: 7epoch:train:8501-9000batch: iter_time=1.475e-04, forward_time=0.209, loss_ctc=60.680, loss_att=41.491, acc=0.746, loss=47.247, backward_time=0.231, grad_norm=30.221, clip=100.000, loss_scale=3.689e+19, optim_step_time=0.075, optim0_lr0=3.897e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-27 22:46:42,292 (trainer:762) INFO: 7epoch:train:9001-9500batch: iter_time=1.435e-04, forward_time=0.210, loss_ctc=57.246, loss_att=39.603, acc=0.759, loss=44.896, backward_time=0.234, grad_norm=27.989, clip=100.000, loss_scale=3.689e+19, optim_step_time=0.075, optim0_lr0=3.888e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-27 22:54:15,034 (trainer:762) INFO: 7epoch:train:9501-10000batch: iter_time=1.434e-04, forward_time=0.207, loss_ctc=57.692, loss_att=39.336, acc=0.751, loss=44.842, backward_time=0.235, grad_norm=30.214, clip=100.000, loss_scale=3.689e+19, optim_step_time=0.075, optim0_lr0=3.878e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-27 22:54:21,388 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-27 22:54:58,747 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-27 22:55:07,402 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e787fafd0>)
+[ip-10-0-216-33:0/16] 2024-02-27 22:55:07,402 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-27 22:55:07,407 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-27 23:09:34,986 (trainer:762) INFO: 7epoch:train:10001-10500batch: iter_time=0.895, forward_time=0.209, loss_ctc=59.693, loss_att=41.072, acc=0.751, loss=46.658, backward_time=0.236, grad_norm=30.837, clip=100.000, loss_scale=7.379e+19, optim_step_time=0.075, optim0_lr0=3.868e-04, train_time=1.840
+[ip-10-0-216-33:0/16] 2024-02-27 23:17:09,418 (trainer:762) INFO: 7epoch:train:10501-11000batch: iter_time=1.502e-04, forward_time=0.210, loss_ctc=57.517, loss_att=39.300, acc=0.755, loss=44.765, backward_time=0.240, grad_norm=29.002, clip=100.000, loss_scale=7.379e+19, optim_step_time=0.075, optim0_lr0=3.859e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-27 23:24:40,185 (trainer:762) INFO: 7epoch:train:11001-11500batch: iter_time=1.531e-04, forward_time=0.209, loss_ctc=54.011, loss_att=37.379, acc=0.769, loss=42.368, backward_time=0.233, grad_norm=26.912, clip=100.000, loss_scale=7.379e+19, optim_step_time=0.075, optim0_lr0=3.849e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-27 23:32:13,712 (trainer:762) INFO: 7epoch:train:11501-12000batch: iter_time=1.489e-04, forward_time=0.210, loss_ctc=55.046, loss_att=38.435, acc=0.756, loss=43.418, backward_time=0.234, grad_norm=27.379, clip=100.000, loss_scale=7.379e+19, optim_step_time=0.075, optim0_lr0=3.840e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-27 23:39:50,930 (trainer:762) INFO: 7epoch:train:12001-12500batch: iter_time=1.485e-04, forward_time=0.208, loss_ctc=59.010, loss_att=40.716, acc=0.749, loss=46.204, backward_time=0.235, grad_norm=33.270, clip=100.000, loss_scale=1.476e+20, optim_step_time=0.075, optim0_lr0=3.830e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-27 23:47:21,732 (trainer:762) INFO: 7epoch:train:12501-13000batch: iter_time=1.488e-04, forward_time=0.209, loss_ctc=54.867, loss_att=37.925, acc=0.755, loss=43.008, backward_time=0.236, grad_norm=28.485, clip=100.000, loss_scale=1.476e+20, optim_step_time=0.075, optim0_lr0=3.821e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-27 23:55:02,691 (trainer:762) INFO: 7epoch:train:13001-13500batch: iter_time=1.454e-04, forward_time=0.208, loss_ctc=57.374, loss_att=39.616, acc=0.749, loss=44.944, backward_time=0.235, grad_norm=29.680, clip=100.000, loss_scale=1.476e+20, optim_step_time=0.075, optim0_lr0=3.812e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-02-28 00:02:39,713 (trainer:762) INFO: 7epoch:train:13501-14000batch: iter_time=1.496e-04, forward_time=0.209, loss_ctc=59.246, loss_att=40.659, acc=0.751, loss=46.235, backward_time=0.237, grad_norm=29.520, clip=100.000, loss_scale=1.476e+20, optim_step_time=0.075, optim0_lr0=3.802e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-28 00:10:20,529 (trainer:762) INFO: 7epoch:train:14001-14500batch: iter_time=1.485e-04, forward_time=0.207, loss_ctc=56.111, loss_att=38.776, acc=0.764, loss=43.976, backward_time=0.230, grad_norm=29.550, clip=100.000, loss_scale=2.951e+20, optim_step_time=0.076, optim0_lr0=3.793e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-02-28 00:17:57,666 (trainer:762) INFO: 7epoch:train:14501-15000batch: iter_time=1.455e-04, forward_time=0.209, loss_ctc=56.817, loss_att=38.683, acc=0.754, loss=44.123, backward_time=0.233, grad_norm=30.193, clip=100.000, loss_scale=2.951e+20, optim_step_time=0.075, optim0_lr0=3.784e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-28 00:32:33,141 (trainer:361) INFO: 7epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=58.460, loss_att=40.261, acc=0.750, loss=45.720, backward_time=0.235, grad_norm=30.144, clip=100.000, loss_scale=5.872e+19, optim_step_time=0.075, optim0_lr0=3.925e-04, train_time=1.001, time=4 hours, 10 minutes and 41.98 seconds, total_count=105000, gpu_max_cached_mem_GB=36.488, [valid] loss_ctc=64.304, cer_ctc=0.346, loss_att=43.848, acc=0.644, cer=0.313, wer=0.999, loss=49.985, time=14 minutes and 7.63 seconds, total_count=14931, gpu_max_cached_mem_GB=36.488
+[ip-10-0-216-33:0/16] 2024-02-28 00:32:42,437 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-28 00:32:42,457 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/2epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-28 00:32:42,458 (trainer:290) INFO: 8/45epoch started. Estimated time to finish: 6 days, 23 hours and 40 minutes
+[ip-10-0-216-33:0/16] 2024-02-28 00:32:42,465 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 00:33:18,158 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 00:33:26,007 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecd9bc400>)
+[ip-10-0-216-33:0/16] 2024-02-28 00:33:26,008 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-28 00:33:26,012 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 00:47:26,538 (trainer:762) INFO: 8epoch:train:1-500batch: iter_time=0.843, forward_time=0.206, loss_ctc=61.629, loss_att=41.608, acc=0.742, loss=47.614, backward_time=0.240, grad_norm=32.378, clip=100.000, loss_scale=2.951e+20, optim_step_time=0.075, optim0_lr0=3.775e-04, train_time=1.768
+[ip-10-0-216-33:0/16] 2024-02-28 00:55:05,189 (trainer:762) INFO: 8epoch:train:501-1000batch: iter_time=1.423e-04, forward_time=0.208, loss_ctc=56.314, loss_att=38.416, acc=0.759, loss=43.785, backward_time=0.228, grad_norm=28.765, clip=100.000, loss_scale=2.951e+20, optim_step_time=0.075, optim0_lr0=3.766e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-28 01:02:43,846 (trainer:762) INFO: 8epoch:train:1001-1500batch: iter_time=1.476e-04, forward_time=0.209, loss_ctc=55.902, loss_att=38.559, acc=0.758, loss=43.762, backward_time=0.235, grad_norm=28.687, clip=100.000, loss_scale=5.903e+20, optim_step_time=0.075, optim0_lr0=3.757e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-28 01:10:16,933 (trainer:762) INFO: 8epoch:train:1501-2000batch: iter_time=1.430e-04, forward_time=0.208, loss_ctc=55.030, loss_att=38.249, acc=0.754, loss=43.284, backward_time=0.239, grad_norm=36.102, clip=100.000, loss_scale=5.903e+20, optim_step_time=0.075, optim0_lr0=3.749e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-28 01:17:49,778 (trainer:762) INFO: 8epoch:train:2001-2500batch: iter_time=1.499e-04, forward_time=0.210, loss_ctc=59.827, loss_att=40.777, acc=0.747, loss=46.492, backward_time=0.237, grad_norm=32.459, clip=100.000, loss_scale=5.903e+20, optim_step_time=0.075, optim0_lr0=3.740e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-28 01:25:23,263 (trainer:762) INFO: 8epoch:train:2501-3000batch: iter_time=1.461e-04, forward_time=0.209, loss_ctc=57.704, loss_att=39.050, acc=0.751, loss=44.646, backward_time=0.233, grad_norm=30.979, clip=100.000, loss_scale=5.903e+20, optim_step_time=0.075, optim0_lr0=3.731e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-28 01:32:56,662 (trainer:762) INFO: 8epoch:train:3001-3500batch: iter_time=1.463e-04, forward_time=0.209, loss_ctc=60.076, loss_att=40.578, acc=0.754, loss=46.427, backward_time=0.234, grad_norm=43.863, clip=100.000, loss_scale=1.181e+21, optim_step_time=0.075, optim0_lr0=3.722e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-28 01:40:36,153 (trainer:762) INFO: 8epoch:train:3501-4000batch: iter_time=1.487e-04, forward_time=0.208, loss_ctc=58.596, loss_att=40.013, acc=0.751, loss=45.588, backward_time=0.233, grad_norm=30.268, clip=100.000, loss_scale=1.181e+21, optim_step_time=0.075, optim0_lr0=3.714e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-28 01:48:17,342 (trainer:762) INFO: 8epoch:train:4001-4500batch: iter_time=1.459e-04, forward_time=0.208, loss_ctc=58.604, loss_att=39.909, acc=0.755, loss=45.518, backward_time=0.234, grad_norm=30.349, clip=100.000, loss_scale=1.181e+21, optim_step_time=0.075, optim0_lr0=3.705e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-02-28 01:55:55,745 (trainer:762) INFO: 8epoch:train:4501-5000batch: iter_time=1.486e-04, forward_time=0.207, loss_ctc=55.452, loss_att=37.753, acc=0.761, loss=43.063, backward_time=0.238, grad_norm=32.389, clip=100.000, loss_scale=1.181e+21, optim_step_time=0.075, optim0_lr0=3.697e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-28 01:56:00,211 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 01:56:37,246 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 01:56:45,227 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5eceb77af0>)
+[ip-10-0-216-33:0/16] 2024-02-28 01:56:45,227 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-28 01:56:45,231 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 02:11:00,103 (trainer:762) INFO: 8epoch:train:5001-5500batch: iter_time=0.881, forward_time=0.208, loss_ctc=59.940, loss_att=40.458, acc=0.749, loss=46.302, backward_time=0.236, grad_norm=30.593, clip=100.000, loss_scale=2.361e+21, optim_step_time=0.075, optim0_lr0=3.689e-04, train_time=1.808
+[ip-10-0-216-33:0/16] 2024-02-28 02:18:34,855 (trainer:762) INFO: 8epoch:train:5501-6000batch: iter_time=1.424e-04, forward_time=0.208, loss_ctc=55.152, loss_att=37.439, acc=0.765, loss=42.753, backward_time=0.235, grad_norm=27.387, clip=100.000, loss_scale=2.361e+21, optim_step_time=0.075, optim0_lr0=3.680e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 02:26:09,459 (trainer:762) INFO: 8epoch:train:6001-6500batch: iter_time=1.420e-04, forward_time=0.208, loss_ctc=54.700, loss_att=37.689, acc=0.764, loss=42.792, backward_time=0.234, grad_norm=28.088, clip=100.000, loss_scale=2.361e+21, optim_step_time=0.075, optim0_lr0=3.672e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 02:33:38,113 (trainer:762) INFO: 8epoch:train:6501-7000batch: iter_time=1.418e-04, forward_time=0.210, loss_ctc=53.886, loss_att=37.269, acc=0.760, loss=42.254, backward_time=0.228, grad_norm=28.899, clip=100.000, loss_scale=2.361e+21, optim_step_time=0.075, optim0_lr0=3.664e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-02-28 02:41:15,721 (trainer:762) INFO: 8epoch:train:7001-7500batch: iter_time=1.442e-04, forward_time=0.206, loss_ctc=58.757, loss_att=39.703, acc=0.753, loss=45.419, backward_time=0.240, grad_norm=32.552, clip=100.000, loss_scale=4.722e+21, optim_step_time=0.075, optim0_lr0=3.656e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-28 02:48:55,806 (trainer:762) INFO: 8epoch:train:7501-8000batch: iter_time=1.495e-04, forward_time=0.207, loss_ctc=56.292, loss_att=38.018, acc=0.757, loss=43.500, backward_time=0.233, grad_norm=29.519, clip=100.000, loss_scale=4.722e+21, optim_step_time=0.075, optim0_lr0=3.647e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-02-28 02:56:37,392 (trainer:762) INFO: 8epoch:train:8001-8500batch: iter_time=1.455e-04, forward_time=0.209, loss_ctc=58.666, loss_att=39.527, acc=0.760, loss=45.269, backward_time=0.227, grad_norm=34.102, clip=100.000, loss_scale=4.722e+21, optim_step_time=0.075, optim0_lr0=3.639e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-02-28 03:04:15,471 (trainer:762) INFO: 8epoch:train:8501-9000batch: iter_time=0.002, forward_time=0.207, loss_ctc=57.105, loss_att=39.081, acc=0.756, loss=44.488, backward_time=0.235, grad_norm=30.053, clip=100.000, loss_scale=4.722e+21, optim_step_time=0.075, optim0_lr0=3.631e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-28 03:11:54,423 (trainer:762) INFO: 8epoch:train:9001-9500batch: iter_time=1.446e-04, forward_time=0.208, loss_ctc=57.147, loss_att=38.837, acc=0.760, loss=44.330, backward_time=0.235, grad_norm=30.883, clip=100.000, loss_scale=9.445e+21, optim_step_time=0.075, optim0_lr0=3.623e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-02-28 03:19:32,102 (trainer:762) INFO: 8epoch:train:9501-10000batch: iter_time=1.424e-04, forward_time=0.207, loss_ctc=54.132, loss_att=36.826, acc=0.766, loss=42.018, backward_time=0.233, grad_norm=28.885, clip=100.000, loss_scale=9.445e+21, optim_step_time=0.075, optim0_lr0=3.615e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-28 03:19:36,744 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 03:20:13,983 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 03:20:22,126 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5eff293e50>)
+[ip-10-0-216-33:0/16] 2024-02-28 03:20:22,127 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-28 03:20:22,131 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 03:34:42,426 (trainer:762) INFO: 8epoch:train:10001-10500batch: iter_time=0.886, forward_time=0.208, loss_ctc=58.790, loss_att=39.735, acc=0.754, loss=45.452, backward_time=0.227, grad_norm=31.441, clip=100.000, loss_scale=9.445e+21, optim_step_time=0.075, optim0_lr0=3.608e-04, train_time=1.820
+[ip-10-0-216-33:0/16] 2024-02-28 03:42:22,322 (trainer:762) INFO: 8epoch:train:10501-11000batch: iter_time=1.503e-04, forward_time=0.207, loss_ctc=53.757, loss_att=36.617, acc=0.769, loss=41.759, backward_time=0.233, grad_norm=29.520, clip=100.000, loss_scale=9.445e+21, optim_step_time=0.075, optim0_lr0=3.600e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-02-28 03:50:01,648 (trainer:762) INFO: 8epoch:train:11001-11500batch: iter_time=1.491e-04, forward_time=0.209, loss_ctc=53.725, loss_att=36.938, acc=0.769, loss=41.974, backward_time=0.237, grad_norm=35.040, clip=100.000, loss_scale=1.889e+22, optim_step_time=0.075, optim0_lr0=3.592e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-02-28 03:57:42,279 (trainer:762) INFO: 8epoch:train:11501-12000batch: iter_time=1.473e-04, forward_time=0.208, loss_ctc=52.884, loss_att=36.669, acc=0.764, loss=41.533, backward_time=0.237, grad_norm=30.513, clip=100.000, loss_scale=1.889e+22, optim_step_time=0.075, optim0_lr0=3.584e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-02-28 04:05:22,948 (trainer:762) INFO: 8epoch:train:12001-12500batch: iter_time=1.488e-04, forward_time=0.207, loss_ctc=57.456, loss_att=38.959, acc=0.757, loss=44.509, backward_time=0.240, grad_norm=30.584, clip=100.000, loss_scale=1.889e+22, optim_step_time=0.075, optim0_lr0=3.577e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-02-28 04:12:56,726 (trainer:762) INFO: 8epoch:train:12501-13000batch: iter_time=1.513e-04, forward_time=0.209, loss_ctc=55.207, loss_att=37.313, acc=0.761, loss=42.681, backward_time=0.235, grad_norm=29.365, clip=100.000, loss_scale=1.889e+22, optim_step_time=0.075, optim0_lr0=3.569e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-28 04:20:34,225 (trainer:762) INFO: 8epoch:train:13001-13500batch: iter_time=1.478e-04, forward_time=0.209, loss_ctc=57.401, loss_att=38.592, acc=0.764, loss=44.235, backward_time=0.231, grad_norm=29.859, clip=100.000, loss_scale=3.778e+22, optim_step_time=0.075, optim0_lr0=3.562e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-28 04:28:10,992 (trainer:762) INFO: 8epoch:train:13501-14000batch: iter_time=1.471e-04, forward_time=0.208, loss_ctc=56.060, loss_att=38.319, acc=0.761, loss=43.641, backward_time=0.236, grad_norm=30.422, clip=100.000, loss_scale=3.778e+22, optim_step_time=0.075, optim0_lr0=3.554e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-28 04:35:46,024 (trainer:762) INFO: 8epoch:train:14001-14500batch: iter_time=1.448e-04, forward_time=0.210, loss_ctc=55.922, loss_att=38.105, acc=0.765, loss=43.450, backward_time=0.243, grad_norm=30.300, clip=100.000, loss_scale=3.778e+22, optim_step_time=0.075, optim0_lr0=3.547e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 04:43:16,717 (trainer:762) INFO: 8epoch:train:14501-15000batch: iter_time=1.481e-04, forward_time=0.209, loss_ctc=53.201, loss_att=36.226, acc=0.770, loss=41.318, backward_time=0.236, grad_norm=28.763, clip=100.000, loss_scale=3.778e+22, optim_step_time=0.075, optim0_lr0=3.539e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-28 04:57:23,326 (trainer:361) INFO: 8epoch results: [train] iter_time=0.087, forward_time=0.208, loss_ctc=56.644, loss_att=38.574, acc=0.759, loss=43.995, backward_time=0.235, grad_norm=31.100, clip=100.000, loss_scale=1.002e+22, optim_step_time=0.075, optim0_lr0=3.653e-04, train_time=1.002, time=4 hours, 10 minutes and 49.45 seconds, total_count=120000, gpu_max_cached_mem_GB=36.488, [valid] loss_ctc=60.437, cer_ctc=0.333, loss_att=44.204, acc=0.648, cer=0.377, wer=1.000, loss=49.074, time=13 minutes and 51.01 seconds, total_count=17064, gpu_max_cached_mem_GB=36.488
+[ip-10-0-216-33:0/16] 2024-02-28 04:57:33,153 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-28 04:57:33,168 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/3epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-28 04:57:33,169 (trainer:290) INFO: 9/45epoch started. Estimated time to finish: 6 days, 19 hours and 16 minutes
+[ip-10-0-216-33:0/16] 2024-02-28 04:57:33,176 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 04:58:09,077 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 04:58:17,478 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ece9c65e0>)
+[ip-10-0-216-33:0/16] 2024-02-28 04:58:17,478 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-28 04:58:17,483 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 05:12:11,693 (trainer:762) INFO: 9epoch:train:1-500batch: iter_time=0.836, forward_time=0.209, loss_ctc=51.998, loss_att=35.799, acc=0.774, loss=40.659, backward_time=0.229, grad_norm=28.062, clip=100.000, loss_scale=7.556e+22, optim_step_time=0.075, optim0_lr0=3.532e-04, train_time=1.757
+[ip-10-0-216-33:0/16] 2024-02-28 05:19:43,279 (trainer:762) INFO: 9epoch:train:501-1000batch: iter_time=1.466e-04, forward_time=0.209, loss_ctc=54.912, loss_att=37.388, acc=0.761, loss=42.645, backward_time=0.230, grad_norm=29.987, clip=100.000, loss_scale=7.556e+22, optim_step_time=0.075, optim0_lr0=3.525e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-28 05:27:19,301 (trainer:762) INFO: 9epoch:train:1001-1500batch: iter_time=1.473e-04, forward_time=0.208, loss_ctc=58.066, loss_att=39.094, acc=0.762, loss=44.786, backward_time=0.229, grad_norm=31.793, clip=100.000, loss_scale=7.556e+22, optim_step_time=0.075, optim0_lr0=3.517e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-28 05:34:52,623 (trainer:762) INFO: 9epoch:train:1501-2000batch: iter_time=1.456e-04, forward_time=0.210, loss_ctc=56.046, loss_att=38.131, acc=0.766, loss=43.505, backward_time=0.233, grad_norm=32.375, clip=100.000, loss_scale=7.556e+22, optim_step_time=0.075, optim0_lr0=3.510e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-28 05:42:26,349 (trainer:762) INFO: 9epoch:train:2001-2500batch: iter_time=1.436e-04, forward_time=0.208, loss_ctc=54.287, loss_att=36.973, acc=0.765, loss=42.167, backward_time=0.232, grad_norm=29.008, clip=100.000, loss_scale=1.511e+23, optim_step_time=0.075, optim0_lr0=3.503e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-28 05:50:04,575 (trainer:762) INFO: 9epoch:train:2501-3000batch: iter_time=1.461e-04, forward_time=0.208, loss_ctc=54.092, loss_att=37.111, acc=0.775, loss=42.205, backward_time=0.232, grad_norm=28.601, clip=100.000, loss_scale=1.511e+23, optim_step_time=0.076, optim0_lr0=3.496e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-28 05:57:41,255 (trainer:762) INFO: 9epoch:train:3001-3500batch: iter_time=1.449e-04, forward_time=0.208, loss_ctc=52.614, loss_att=35.717, acc=0.765, loss=40.786, backward_time=0.235, grad_norm=29.473, clip=100.000, loss_scale=1.511e+23, optim_step_time=0.075, optim0_lr0=3.489e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-28 06:05:22,814 (trainer:762) INFO: 9epoch:train:3501-4000batch: iter_time=1.482e-04, forward_time=0.205, loss_ctc=56.100, loss_att=38.178, acc=0.766, loss=43.554, backward_time=0.228, grad_norm=29.356, clip=100.000, loss_scale=1.511e+23, optim_step_time=0.076, optim0_lr0=3.482e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-02-28 06:12:56,206 (trainer:762) INFO: 9epoch:train:4001-4500batch: iter_time=1.489e-04, forward_time=0.209, loss_ctc=51.624, loss_att=35.487, acc=0.773, loss=40.328, backward_time=0.231, grad_norm=27.249, clip=100.000, loss_scale=3.022e+23, optim_step_time=0.076, optim0_lr0=3.475e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-28 06:20:31,307 (trainer:762) INFO: 9epoch:train:4501-5000batch: iter_time=1.456e-04, forward_time=0.208, loss_ctc=53.344, loss_att=36.526, acc=0.770, loss=41.571, backward_time=0.238, grad_norm=28.544, clip=100.000, loss_scale=3.022e+23, optim_step_time=0.076, optim0_lr0=3.468e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 06:20:36,416 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 06:21:13,513 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 06:21:22,108 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecff718b0>)
+[ip-10-0-216-33:0/16] 2024-02-28 06:21:22,108 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-28 06:21:22,112 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 06:35:44,340 (trainer:762) INFO: 9epoch:train:5001-5500batch: iter_time=0.870, forward_time=0.208, loss_ctc=51.039, loss_att=35.002, acc=0.778, loss=39.813, backward_time=0.235, grad_norm=26.892, clip=100.000, loss_scale=3.022e+23, optim_step_time=0.075, optim0_lr0=3.461e-04, train_time=1.826
+[ip-10-0-216-33:0/16] 2024-02-28 06:43:17,073 (trainer:762) INFO: 9epoch:train:5501-6000batch: iter_time=1.431e-04, forward_time=0.210, loss_ctc=53.501, loss_att=36.548, acc=0.765, loss=41.634, backward_time=0.236, grad_norm=33.185, clip=100.000, loss_scale=3.022e+23, optim_step_time=0.075, optim0_lr0=3.454e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-28 06:50:48,814 (trainer:762) INFO: 9epoch:train:6001-6500batch: iter_time=1.433e-04, forward_time=0.210, loss_ctc=56.592, loss_att=37.999, acc=0.767, loss=43.577, backward_time=0.230, grad_norm=31.168, clip=100.000, loss_scale=6.045e+23, optim_step_time=0.075, optim0_lr0=3.447e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-28 06:58:23,923 (trainer:762) INFO: 9epoch:train:6501-7000batch: iter_time=1.412e-04, forward_time=0.209, loss_ctc=54.595, loss_att=37.241, acc=0.771, loss=42.447, backward_time=0.230, grad_norm=39.097, clip=100.000, loss_scale=6.045e+23, optim_step_time=0.075, optim0_lr0=3.440e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 07:06:04,526 (trainer:762) INFO: 9epoch:train:7001-7500batch: iter_time=1.430e-04, forward_time=0.208, loss_ctc=53.319, loss_att=36.310, acc=0.769, loss=41.413, backward_time=0.232, grad_norm=28.730, clip=100.000, loss_scale=6.045e+23, optim_step_time=0.075, optim0_lr0=3.433e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-02-28 07:13:38,086 (trainer:762) INFO: 9epoch:train:7501-8000batch: iter_time=1.449e-04, forward_time=0.208, loss_ctc=52.983, loss_att=36.254, acc=0.779, loss=41.273, backward_time=0.240, grad_norm=28.424, clip=100.000, loss_scale=6.045e+23, optim_step_time=0.075, optim0_lr0=3.427e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-28 07:21:16,553 (trainer:762) INFO: 9epoch:train:8001-8500batch: iter_time=1.435e-04, forward_time=0.208, loss_ctc=51.710, loss_att=35.149, acc=0.769, loss=40.117, backward_time=0.237, grad_norm=30.131, clip=100.000, loss_scale=1.209e+24, optim_step_time=0.075, optim0_lr0=3.420e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-28 07:28:53,125 (trainer:762) INFO: 9epoch:train:8501-9000batch: iter_time=1.458e-04, forward_time=0.208, loss_ctc=54.977, loss_att=37.431, acc=0.770, loss=42.695, backward_time=0.237, grad_norm=29.380, clip=100.000, loss_scale=1.209e+24, optim_step_time=0.075, optim0_lr0=3.413e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-28 07:36:25,302 (trainer:762) INFO: 9epoch:train:9001-9500batch: iter_time=1.419e-04, forward_time=0.210, loss_ctc=50.752, loss_att=34.864, acc=0.776, loss=39.630, backward_time=0.231, grad_norm=31.511, clip=100.000, loss_scale=1.209e+24, optim_step_time=0.076, optim0_lr0=3.407e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-28 07:43:56,842 (trainer:762) INFO: 9epoch:train:9501-10000batch: iter_time=1.417e-04, forward_time=0.208, loss_ctc=52.238, loss_att=35.854, acc=0.774, loss=40.769, backward_time=0.229, grad_norm=28.566, clip=100.000, loss_scale=1.209e+24, optim_step_time=0.075, optim0_lr0=3.400e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-28 07:44:01,912 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 07:44:38,881 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 07:44:47,488 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f14094dc0>)
+[ip-10-0-216-33:0/16] 2024-02-28 07:44:47,489 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-28 07:44:47,493 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 07:59:14,385 (trainer:762) INFO: 9epoch:train:10001-10500batch: iter_time=0.867, forward_time=0.209, loss_ctc=50.401, loss_att=34.563, acc=0.781, loss=39.314, backward_time=0.234, grad_norm=26.697, clip=100.000, loss_scale=2.418e+24, optim_step_time=0.075, optim0_lr0=3.394e-04, train_time=1.835
+[ip-10-0-216-33:0/16] 2024-02-28 08:06:52,252 (trainer:762) INFO: 9epoch:train:10501-11000batch: iter_time=1.501e-04, forward_time=0.208, loss_ctc=52.666, loss_att=35.954, acc=0.768, loss=40.967, backward_time=0.234, grad_norm=28.222, clip=100.000, loss_scale=2.418e+24, optim_step_time=0.075, optim0_lr0=3.387e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-28 08:14:31,059 (trainer:762) INFO: 9epoch:train:11001-11500batch: iter_time=1.534e-04, forward_time=0.207, loss_ctc=55.787, loss_att=37.522, acc=0.770, loss=43.001, backward_time=0.231, grad_norm=30.183, clip=100.000, loss_scale=2.418e+24, optim_step_time=0.075, optim0_lr0=3.381e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-28 08:22:08,307 (trainer:762) INFO: 9epoch:train:11501-12000batch: iter_time=1.469e-04, forward_time=0.209, loss_ctc=53.691, loss_att=36.616, acc=0.775, loss=41.739, backward_time=0.234, grad_norm=29.588, clip=100.000, loss_scale=2.418e+24, optim_step_time=0.075, optim0_lr0=3.374e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-28 08:29:48,077 (trainer:762) INFO: 9epoch:train:12001-12500batch: iter_time=1.530e-04, forward_time=0.208, loss_ctc=52.705, loss_att=35.783, acc=0.772, loss=40.860, backward_time=0.236, grad_norm=29.020, clip=100.000, loss_scale=4.836e+24, optim_step_time=0.075, optim0_lr0=3.368e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-28 08:37:21,464 (trainer:762) INFO: 9epoch:train:12501-13000batch: iter_time=1.536e-04, forward_time=0.210, loss_ctc=52.290, loss_att=35.832, acc=0.782, loss=40.769, backward_time=0.235, grad_norm=27.796, clip=100.000, loss_scale=4.836e+24, optim_step_time=0.075, optim0_lr0=3.361e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-28 08:45:01,274 (trainer:762) INFO: 9epoch:train:13001-13500batch: iter_time=1.524e-04, forward_time=0.207, loss_ctc=50.608, loss_att=34.517, acc=0.773, loss=39.345, backward_time=0.231, grad_norm=28.941, clip=100.000, loss_scale=4.836e+24, optim_step_time=0.075, optim0_lr0=3.355e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-28 08:52:39,081 (trainer:762) INFO: 9epoch:train:13501-14000batch: iter_time=1.495e-04, forward_time=0.209, loss_ctc=53.940, loss_att=36.718, acc=0.774, loss=41.885, backward_time=0.235, grad_norm=30.108, clip=100.000, loss_scale=4.836e+24, optim_step_time=0.075, optim0_lr0=3.349e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-28 09:00:13,592 (trainer:762) INFO: 9epoch:train:14001-14500batch: iter_time=1.535e-04, forward_time=0.209, loss_ctc=49.844, loss_att=34.289, acc=0.779, loss=38.955, backward_time=0.232, grad_norm=27.753, clip=100.000, loss_scale=9.671e+24, optim_step_time=0.075, optim0_lr0=3.343e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 09:07:50,776 (trainer:762) INFO: 9epoch:train:14501-15000batch: iter_time=1.465e-04, forward_time=0.209, loss_ctc=51.679, loss_att=35.474, acc=0.776, loss=40.336, backward_time=0.237, grad_norm=28.156, clip=100.000, loss_scale=9.671e+24, optim_step_time=0.075, optim0_lr0=3.336e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-28 09:22:08,669 (trainer:361) INFO: 9epoch results: [train] iter_time=0.086, forward_time=0.208, loss_ctc=53.280, loss_att=36.344, acc=0.771, loss=41.425, backward_time=0.233, grad_norm=29.600, clip=100.000, loss_scale=1.924e+24, optim_step_time=0.075, optim0_lr0=3.431e-04, train_time=1.001, time=4 hours, 10 minutes and 33.55 seconds, total_count=135000, gpu_max_cached_mem_GB=36.488, [valid] loss_ctc=54.791, cer_ctc=0.304, loss_att=39.448, acc=0.673, cer=0.318, wer=0.999, loss=44.051, time=14 minutes and 1.72 seconds, total_count=19197, gpu_max_cached_mem_GB=36.488
+[ip-10-0-216-33:0/16] 2024-02-28 09:22:17,984 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-28 09:22:18,017 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/4epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-28 09:22:18,018 (trainer:290) INFO: 10/45epoch started. Estimated time to finish: 6 days, 14 hours and 51 minutes
+[ip-10-0-216-33:0/16] 2024-02-28 09:22:18,025 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 09:22:53,663 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 09:23:02,102 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0082df0>)
+[ip-10-0-216-33:0/16] 2024-02-28 09:23:02,102 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-28 09:23:02,107 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 09:36:56,356 (trainer:762) INFO: 10epoch:train:1-500batch: iter_time=0.843, forward_time=0.209, loss_ctc=49.222, loss_att=33.626, acc=0.777, loss=38.305, backward_time=0.235, grad_norm=27.948, clip=100.000, loss_scale=9.671e+24, optim_step_time=0.075, optim0_lr0=3.330e-04, train_time=1.756
+[ip-10-0-216-33:0/16] 2024-02-28 09:44:30,626 (trainer:762) INFO: 10epoch:train:501-1000batch: iter_time=1.468e-04, forward_time=0.208, loss_ctc=51.510, loss_att=35.495, acc=0.771, loss=40.299, backward_time=0.235, grad_norm=29.339, clip=100.000, loss_scale=9.671e+24, optim_step_time=0.075, optim0_lr0=3.324e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-28 09:52:02,008 (trainer:762) INFO: 10epoch:train:1001-1500batch: iter_time=1.485e-04, forward_time=0.210, loss_ctc=49.845, loss_att=34.186, acc=0.780, loss=38.884, backward_time=0.229, grad_norm=27.975, clip=100.000, loss_scale=1.934e+25, optim_step_time=0.075, optim0_lr0=3.318e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-28 09:59:38,948 (trainer:762) INFO: 10epoch:train:1501-2000batch: iter_time=1.468e-04, forward_time=0.208, loss_ctc=51.972, loss_att=35.383, acc=0.772, loss=40.360, backward_time=0.233, grad_norm=29.522, clip=100.000, loss_scale=1.934e+25, optim_step_time=0.075, optim0_lr0=3.312e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-28 10:07:13,635 (trainer:762) INFO: 10epoch:train:2001-2500batch: iter_time=1.467e-04, forward_time=0.208, loss_ctc=51.252, loss_att=35.130, acc=0.773, loss=39.967, backward_time=0.236, grad_norm=29.638, clip=100.000, loss_scale=1.934e+25, optim_step_time=0.075, optim0_lr0=3.306e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 10:14:51,514 (trainer:762) INFO: 10epoch:train:2501-3000batch: iter_time=1.469e-04, forward_time=0.209, loss_ctc=54.867, loss_att=37.606, acc=0.770, loss=42.784, backward_time=0.239, grad_norm=30.668, clip=100.000, loss_scale=1.934e+25, optim_step_time=0.075, optim0_lr0=3.300e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-28 10:22:32,895 (trainer:762) INFO: 10epoch:train:3001-3500batch: iter_time=1.477e-04, forward_time=0.208, loss_ctc=54.902, loss_att=37.192, acc=0.770, loss=42.505, backward_time=0.235, grad_norm=29.310, clip=100.000, loss_scale=3.869e+25, optim_step_time=0.075, optim0_lr0=3.294e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-02-28 10:30:08,323 (trainer:762) INFO: 10epoch:train:3501-4000batch: iter_time=1.489e-04, forward_time=0.207, loss_ctc=48.229, loss_att=33.432, acc=0.779, loss=37.871, backward_time=0.238, grad_norm=27.236, clip=100.000, loss_scale=3.869e+25, optim_step_time=0.075, optim0_lr0=3.288e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-28 10:37:49,310 (trainer:762) INFO: 10epoch:train:4001-4500batch: iter_time=1.482e-04, forward_time=0.207, loss_ctc=51.069, loss_att=34.591, acc=0.771, loss=39.534, backward_time=0.242, grad_norm=29.088, clip=100.000, loss_scale=3.869e+25, optim_step_time=0.075, optim0_lr0=3.282e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-02-28 10:45:26,709 (trainer:762) INFO: 10epoch:train:4501-5000batch: iter_time=1.479e-04, forward_time=0.207, loss_ctc=52.465, loss_att=35.664, acc=0.769, loss=40.705, backward_time=0.241, grad_norm=28.895, clip=100.000, loss_scale=3.869e+25, optim_step_time=0.075, optim0_lr0=3.276e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-28 10:45:31,310 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 10:46:08,100 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 10:46:16,784 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed2d0c070>)
+[ip-10-0-216-33:0/16] 2024-02-28 10:46:16,784 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-28 10:46:16,788 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 11:00:38,084 (trainer:762) INFO: 10epoch:train:5001-5500batch: iter_time=0.888, forward_time=0.207, loss_ctc=48.445, loss_att=33.302, acc=0.781, loss=37.845, backward_time=0.234, grad_norm=28.652, clip=100.000, loss_scale=7.737e+25, optim_step_time=0.075, optim0_lr0=3.270e-04, train_time=1.823
+[ip-10-0-216-33:0/16] 2024-02-28 11:08:14,551 (trainer:762) INFO: 10epoch:train:5501-6000batch: iter_time=1.442e-04, forward_time=0.209, loss_ctc=50.861, loss_att=35.048, acc=0.774, loss=39.792, backward_time=0.234, grad_norm=32.199, clip=100.000, loss_scale=7.737e+25, optim_step_time=0.075, optim0_lr0=3.265e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-28 11:15:50,123 (trainer:762) INFO: 10epoch:train:6001-6500batch: iter_time=1.472e-04, forward_time=0.209, loss_ctc=48.968, loss_att=33.647, acc=0.784, loss=38.243, backward_time=0.238, grad_norm=27.746, clip=100.000, loss_scale=7.737e+25, optim_step_time=0.075, optim0_lr0=3.259e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-28 11:23:22,067 (trainer:762) INFO: 10epoch:train:6501-7000batch: iter_time=1.424e-04, forward_time=0.209, loss_ctc=50.981, loss_att=34.670, acc=0.776, loss=39.563, backward_time=0.234, grad_norm=28.774, clip=100.000, loss_scale=7.737e+25, optim_step_time=0.075, optim0_lr0=3.253e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-28 11:31:00,973 (trainer:762) INFO: 10epoch:train:7001-7500batch: iter_time=1.428e-04, forward_time=0.207, loss_ctc=50.277, loss_att=34.486, acc=0.777, loss=39.223, backward_time=0.230, grad_norm=29.192, clip=100.000, loss_scale=1.547e+26, optim_step_time=0.075, optim0_lr0=3.247e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-02-28 11:38:42,792 (trainer:762) INFO: 10epoch:train:7501-8000batch: iter_time=1.454e-04, forward_time=0.208, loss_ctc=53.761, loss_att=36.904, acc=0.773, loss=41.961, backward_time=0.233, grad_norm=37.100, clip=100.000, loss_scale=1.547e+26, optim_step_time=0.075, optim0_lr0=3.242e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-02-28 11:46:26,225 (trainer:762) INFO: 10epoch:train:8001-8500batch: iter_time=1.487e-04, forward_time=0.208, loss_ctc=53.676, loss_att=36.525, acc=0.774, loss=41.670, backward_time=0.237, grad_norm=29.093, clip=100.000, loss_scale=1.547e+26, optim_step_time=0.075, optim0_lr0=3.236e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-02-28 11:54:00,672 (trainer:762) INFO: 10epoch:train:8501-9000batch: iter_time=1.494e-04, forward_time=0.207, loss_ctc=47.630, loss_att=32.958, acc=0.782, loss=37.360, backward_time=0.229, grad_norm=28.446, clip=100.000, loss_scale=1.547e+26, optim_step_time=0.075, optim0_lr0=3.230e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 12:01:39,022 (trainer:762) INFO: 10epoch:train:9001-9500batch: iter_time=1.443e-04, forward_time=0.208, loss_ctc=50.325, loss_att=34.043, acc=0.774, loss=38.928, backward_time=0.240, grad_norm=29.568, clip=100.000, loss_scale=3.095e+26, optim_step_time=0.075, optim0_lr0=3.225e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-28 12:09:13,083 (trainer:762) INFO: 10epoch:train:9501-10000batch: iter_time=1.469e-04, forward_time=0.207, loss_ctc=51.776, loss_att=35.214, acc=0.772, loss=40.183, backward_time=0.238, grad_norm=29.709, clip=100.000, loss_scale=3.095e+26, optim_step_time=0.075, optim0_lr0=3.219e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-28 12:09:17,856 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 12:09:54,848 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 12:10:03,315 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f03f4d970>)
+[ip-10-0-216-33:0/16] 2024-02-28 12:10:03,316 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-28 12:10:03,320 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 12:24:24,833 (trainer:762) INFO: 10epoch:train:10001-10500batch: iter_time=0.883, forward_time=0.207, loss_ctc=47.727, loss_att=32.901, acc=0.784, loss=37.349, backward_time=0.230, grad_norm=27.715, clip=100.000, loss_scale=3.095e+26, optim_step_time=0.075, optim0_lr0=3.214e-04, train_time=1.823
+[ip-10-0-216-33:0/16] 2024-02-28 12:31:58,077 (trainer:762) INFO: 10epoch:train:10501-11000batch: iter_time=1.521e-04, forward_time=0.210, loss_ctc=50.143, loss_att=34.525, acc=0.777, loss=39.211, backward_time=0.233, grad_norm=31.115, clip=100.000, loss_scale=3.095e+26, optim_step_time=0.075, optim0_lr0=3.208e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-28 12:39:34,777 (trainer:762) INFO: 10epoch:train:11001-11500batch: iter_time=1.535e-04, forward_time=0.207, loss_ctc=48.356, loss_att=33.199, acc=0.787, loss=37.746, backward_time=0.234, grad_norm=28.197, clip=100.000, loss_scale=6.190e+26, optim_step_time=0.075, optim0_lr0=3.203e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-28 12:47:17,728 (trainer:762) INFO: 10epoch:train:11501-12000batch: iter_time=1.508e-04, forward_time=0.208, loss_ctc=50.405, loss_att=34.340, acc=0.778, loss=39.160, backward_time=0.237, grad_norm=28.449, clip=100.000, loss_scale=6.190e+26, optim_step_time=0.075, optim0_lr0=3.197e-04, train_time=0.926
+[ip-10-0-216-33:0/16] 2024-02-28 12:54:48,967 (trainer:762) INFO: 10epoch:train:12001-12500batch: iter_time=1.503e-04, forward_time=0.209, loss_ctc=49.608, loss_att=33.977, acc=0.780, loss=38.667, backward_time=0.234, grad_norm=37.166, clip=100.000, loss_scale=6.190e+26, optim_step_time=0.075, optim0_lr0=3.192e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-02-28 13:02:21,492 (trainer:762) INFO: 10epoch:train:12501-13000batch: iter_time=1.511e-04, forward_time=0.211, loss_ctc=52.907, loss_att=36.317, acc=0.776, loss=41.294, backward_time=0.234, grad_norm=30.522, clip=100.000, loss_scale=6.190e+26, optim_step_time=0.075, optim0_lr0=3.186e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-28 13:09:56,429 (trainer:762) INFO: 10epoch:train:13001-13500batch: iter_time=1.476e-04, forward_time=0.209, loss_ctc=52.953, loss_att=36.095, acc=0.776, loss=41.152, backward_time=0.235, grad_norm=29.083, clip=100.000, loss_scale=1.238e+27, optim_step_time=0.075, optim0_lr0=3.181e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 13:17:28,449 (trainer:762) INFO: 10epoch:train:13501-14000batch: iter_time=1.447e-04, forward_time=0.209, loss_ctc=47.021, loss_att=32.561, acc=0.785, loss=36.899, backward_time=0.233, grad_norm=27.403, clip=100.000, loss_scale=1.238e+27, optim_step_time=0.075, optim0_lr0=3.176e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-28 13:25:03,137 (trainer:762) INFO: 10epoch:train:14001-14500batch: iter_time=1.523e-04, forward_time=0.209, loss_ctc=49.487, loss_att=33.566, acc=0.778, loss=38.342, backward_time=0.240, grad_norm=27.951, clip=100.000, loss_scale=1.238e+27, optim_step_time=0.075, optim0_lr0=3.170e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 13:32:42,521 (trainer:762) INFO: 10epoch:train:14501-15000batch: iter_time=1.486e-04, forward_time=0.208, loss_ctc=51.052, loss_att=34.714, acc=0.775, loss=39.615, backward_time=0.233, grad_norm=28.879, clip=100.000, loss_scale=1.238e+27, optim_step_time=0.075, optim0_lr0=3.165e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-28 13:46:53,735 (trainer:361) INFO: 10epoch results: [train] iter_time=0.087, forward_time=0.208, loss_ctc=50.723, loss_att=34.710, acc=0.776, loss=39.514, backward_time=0.235, grad_norm=29.553, clip=100.000, loss_scale=3.282e+26, optim_step_time=0.075, optim0_lr0=3.246e-04, train_time=1.001, time=4 hours, 10 minutes and 39.37 seconds, total_count=150000, gpu_max_cached_mem_GB=36.488, [valid] loss_ctc=53.295, cer_ctc=0.294, loss_att=38.250, acc=0.679, cer=0.349, wer=1.000, loss=42.764, time=13 minutes and 56.01 seconds, total_count=21330, gpu_max_cached_mem_GB=36.488
+[ip-10-0-216-33:0/16] 2024-02-28 13:47:03,608 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-28 13:47:03,631 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/5epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-28 13:47:03,631 (trainer:290) INFO: 11/45epoch started. Estimated time to finish: 6 days, 10 hours and 26 minutes
+[ip-10-0-216-33:0/16] 2024-02-28 13:47:03,640 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 13:47:40,150 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 13:47:48,634 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f657290f490>)
+[ip-10-0-216-33:0/16] 2024-02-28 13:47:48,634 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-28 13:47:48,639 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 14:01:52,564 (trainer:762) INFO: 11epoch:train:1-500batch: iter_time=0.820, forward_time=0.209, loss_ctc=51.023, loss_att=34.981, acc=0.784, loss=39.793, backward_time=0.231, grad_norm=29.352, clip=100.000, loss_scale=2.476e+27, optim_step_time=0.075, optim0_lr0=3.160e-04, train_time=1.778
+[ip-10-0-216-33:0/16] 2024-02-28 14:09:29,447 (trainer:762) INFO: 11epoch:train:501-1000batch: iter_time=1.487e-04, forward_time=0.209, loss_ctc=50.240, loss_att=34.482, acc=0.783, loss=39.209, backward_time=0.236, grad_norm=28.985, clip=100.000, loss_scale=2.476e+27, optim_step_time=0.075, optim0_lr0=3.154e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-28 14:17:07,122 (trainer:762) INFO: 11epoch:train:1001-1500batch: iter_time=1.486e-04, forward_time=0.209, loss_ctc=50.301, loss_att=34.292, acc=0.772, loss=39.095, backward_time=0.235, grad_norm=29.901, clip=100.000, loss_scale=2.476e+27, optim_step_time=0.075, optim0_lr0=3.149e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-28 14:24:48,402 (trainer:762) INFO: 11epoch:train:1501-2000batch: iter_time=1.448e-04, forward_time=0.207, loss_ctc=49.760, loss_att=34.030, acc=0.789, loss=38.749, backward_time=0.235, grad_norm=27.658, clip=100.000, loss_scale=2.476e+27, optim_step_time=0.075, optim0_lr0=3.144e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-02-28 14:32:26,378 (trainer:762) INFO: 11epoch:train:2001-2500batch: iter_time=1.449e-04, forward_time=0.209, loss_ctc=51.963, loss_att=35.699, acc=0.781, loss=40.578, backward_time=0.233, grad_norm=31.300, clip=100.000, loss_scale=4.952e+27, optim_step_time=0.076, optim0_lr0=3.139e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-28 14:40:01,686 (trainer:762) INFO: 11epoch:train:2501-3000batch: iter_time=1.490e-04, forward_time=0.209, loss_ctc=50.326, loss_att=34.357, acc=0.779, loss=39.148, backward_time=0.230, grad_norm=29.879, clip=100.000, loss_scale=4.952e+27, optim_step_time=0.075, optim0_lr0=3.134e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 14:47:33,375 (trainer:762) INFO: 11epoch:train:3001-3500batch: iter_time=1.466e-04, forward_time=0.209, loss_ctc=51.458, loss_att=35.234, acc=0.778, loss=40.101, backward_time=0.233, grad_norm=101.998, clip=100.000, loss_scale=4.952e+27, optim_step_time=0.076, optim0_lr0=3.129e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-28 14:55:07,832 (trainer:762) INFO: 11epoch:train:3501-4000batch: iter_time=1.456e-04, forward_time=0.211, loss_ctc=51.695, loss_att=34.959, acc=0.796, loss=39.980, backward_time=0.232, grad_norm=28.639, clip=100.000, loss_scale=4.952e+27, optim_step_time=0.075, optim0_lr0=3.123e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 15:02:47,857 (trainer:762) INFO: 11epoch:train:4001-4500batch: iter_time=1.449e-04, forward_time=0.209, loss_ctc=52.412, loss_att=35.434, acc=0.787, loss=40.527, backward_time=0.234, grad_norm=28.813, clip=100.000, loss_scale=9.904e+27, optim_step_time=0.075, optim0_lr0=3.118e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-02-28 15:10:17,314 (trainer:762) INFO: 11epoch:train:4501-5000batch: iter_time=1.437e-04, forward_time=0.207, loss_ctc=48.465, loss_att=33.363, acc=0.784, loss=37.893, backward_time=0.228, grad_norm=27.865, clip=100.000, loss_scale=9.904e+27, optim_step_time=0.075, optim0_lr0=3.113e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-02-28 15:10:22,314 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 15:10:59,328 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 15:11:07,327 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfe08c40>)
+[ip-10-0-216-33:0/16] 2024-02-28 15:11:07,328 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-28 15:11:07,332 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 15:25:21,356 (trainer:762) INFO: 11epoch:train:5001-5500batch: iter_time=0.869, forward_time=0.208, loss_ctc=49.561, loss_att=34.006, acc=0.788, loss=38.672, backward_time=0.226, grad_norm=31.965, clip=100.000, loss_scale=9.904e+27, optim_step_time=0.075, optim0_lr0=3.108e-04, train_time=1.808
+[ip-10-0-216-33:0/16] 2024-02-28 15:33:03,460 (trainer:762) INFO: 11epoch:train:5501-6000batch: iter_time=1.381e-04, forward_time=0.206, loss_ctc=49.070, loss_att=33.789, acc=0.786, loss=38.373, backward_time=0.232, grad_norm=28.946, clip=100.000, loss_scale=9.904e+27, optim_step_time=0.075, optim0_lr0=3.103e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-02-28 15:40:43,446 (trainer:762) INFO: 11epoch:train:6001-6500batch: iter_time=1.421e-04, forward_time=0.206, loss_ctc=49.250, loss_att=33.772, acc=0.775, loss=38.416, backward_time=0.232, grad_norm=29.553, clip=100.000, loss_scale=1.981e+28, optim_step_time=0.075, optim0_lr0=3.098e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-02-28 15:48:19,427 (trainer:762) INFO: 11epoch:train:6501-7000batch: iter_time=1.446e-04, forward_time=0.209, loss_ctc=49.307, loss_att=33.598, acc=0.792, loss=38.310, backward_time=0.232, grad_norm=27.584, clip=100.000, loss_scale=1.981e+28, optim_step_time=0.075, optim0_lr0=3.093e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-28 15:55:54,273 (trainer:762) INFO: 11epoch:train:7001-7500batch: iter_time=1.427e-04, forward_time=0.208, loss_ctc=50.966, loss_att=34.997, acc=0.784, loss=39.787, backward_time=0.233, grad_norm=30.642, clip=100.000, loss_scale=1.981e+28, optim_step_time=0.075, optim0_lr0=3.089e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 16:03:26,812 (trainer:762) INFO: 11epoch:train:7501-8000batch: iter_time=1.428e-04, forward_time=0.209, loss_ctc=49.737, loss_att=33.968, acc=0.781, loss=38.699, backward_time=0.235, grad_norm=38.244, clip=100.000, loss_scale=1.981e+28, optim_step_time=0.075, optim0_lr0=3.084e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-28 16:11:07,217 (trainer:762) INFO: 11epoch:train:8001-8500batch: iter_time=1.412e-04, forward_time=0.207, loss_ctc=50.499, loss_att=34.483, acc=0.781, loss=39.288, backward_time=0.242, grad_norm=29.388, clip=100.000, loss_scale=3.961e+28, optim_step_time=0.075, optim0_lr0=3.079e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-02-28 16:18:51,021 (trainer:762) INFO: 11epoch:train:8501-9000batch: iter_time=1.420e-04, forward_time=0.209, loss_ctc=50.719, loss_att=34.463, acc=0.797, loss=39.340, backward_time=0.234, grad_norm=27.851, clip=100.000, loss_scale=3.961e+28, optim_step_time=0.075, optim0_lr0=3.074e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-02-28 16:26:26,275 (trainer:762) INFO: 11epoch:train:9001-9500batch: iter_time=1.408e-04, forward_time=0.209, loss_ctc=51.516, loss_att=34.877, acc=0.790, loss=39.869, backward_time=0.233, grad_norm=28.977, clip=100.000, loss_scale=3.961e+28, optim_step_time=0.075, optim0_lr0=3.069e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 16:34:02,139 (trainer:762) INFO: 11epoch:train:9501-10000batch: iter_time=1.458e-04, forward_time=0.208, loss_ctc=47.564, loss_att=32.790, acc=0.787, loss=37.222, backward_time=0.234, grad_norm=27.489, clip=100.000, loss_scale=3.961e+28, optim_step_time=0.075, optim0_lr0=3.064e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-28 16:34:06,862 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 16:34:44,344 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 16:34:52,348 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f05ea8c40>)
+[ip-10-0-216-33:0/16] 2024-02-28 16:34:52,348 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-28 16:34:52,352 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 16:49:23,483 (trainer:762) INFO: 11epoch:train:10001-10500batch: iter_time=0.880, forward_time=0.209, loss_ctc=48.987, loss_att=33.680, acc=0.790, loss=38.272, backward_time=0.237, grad_norm=29.332, clip=100.000, loss_scale=7.923e+28, optim_step_time=0.075, optim0_lr0=3.059e-04, train_time=1.842
+[ip-10-0-216-33:0/16] 2024-02-28 16:57:01,861 (trainer:762) INFO: 11epoch:train:10501-11000batch: iter_time=1.447e-04, forward_time=0.209, loss_ctc=48.610, loss_att=33.447, acc=0.788, loss=37.996, backward_time=0.231, grad_norm=29.548, clip=100.000, loss_scale=7.923e+28, optim_step_time=0.075, optim0_lr0=3.055e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-28 17:04:37,244 (trainer:762) INFO: 11epoch:train:11001-11500batch: iter_time=1.487e-04, forward_time=0.208, loss_ctc=48.705, loss_att=33.422, acc=0.776, loss=38.007, backward_time=0.233, grad_norm=30.429, clip=100.000, loss_scale=7.923e+28, optim_step_time=0.075, optim0_lr0=3.050e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-28 17:12:13,348 (trainer:762) INFO: 11epoch:train:11501-12000batch: iter_time=1.494e-04, forward_time=0.210, loss_ctc=48.848, loss_att=33.359, acc=0.793, loss=38.006, backward_time=0.235, grad_norm=28.340, clip=100.000, loss_scale=7.923e+28, optim_step_time=0.075, optim0_lr0=3.045e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-28 17:19:53,845 (trainer:762) INFO: 11epoch:train:12001-12500batch: iter_time=1.500e-04, forward_time=0.207, loss_ctc=50.119, loss_att=34.571, acc=0.787, loss=39.236, backward_time=0.235, grad_norm=30.465, clip=100.000, loss_scale=1.585e+29, optim_step_time=0.075, optim0_lr0=3.041e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-02-28 17:27:26,757 (trainer:762) INFO: 11epoch:train:12501-13000batch: iter_time=1.478e-04, forward_time=0.209, loss_ctc=49.308, loss_att=33.609, acc=0.783, loss=38.319, backward_time=0.229, grad_norm=29.202, clip=100.000, loss_scale=1.585e+29, optim_step_time=0.075, optim0_lr0=3.036e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-28 17:35:01,184 (trainer:762) INFO: 11epoch:train:13001-13500batch: iter_time=1.505e-04, forward_time=0.210, loss_ctc=49.874, loss_att=34.156, acc=0.783, loss=38.871, backward_time=0.233, grad_norm=30.569, clip=100.000, loss_scale=1.585e+29, optim_step_time=0.075, optim0_lr0=3.031e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 17:42:36,532 (trainer:762) INFO: 11epoch:train:13501-14000batch: iter_time=1.492e-04, forward_time=0.210, loss_ctc=50.247, loss_att=34.143, acc=0.800, loss=38.974, backward_time=0.233, grad_norm=27.997, clip=100.000, loss_scale=1.585e+29, optim_step_time=0.075, optim0_lr0=3.027e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 17:50:06,316 (trainer:762) INFO: 11epoch:train:14001-14500batch: iter_time=1.489e-04, forward_time=0.211, loss_ctc=50.867, loss_att=34.511, acc=0.792, loss=39.418, backward_time=0.236, grad_norm=28.992, clip=100.000, loss_scale=3.169e+29, optim_step_time=0.075, optim0_lr0=3.022e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-02-28 17:57:40,171 (trainer:762) INFO: 11epoch:train:14501-15000batch: iter_time=1.478e-04, forward_time=0.209, loss_ctc=47.004, loss_att=32.523, acc=0.788, loss=36.867, backward_time=0.230, grad_norm=28.879, clip=100.000, loss_scale=3.169e+29, optim_step_time=0.075, optim0_lr0=3.017e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-28 18:11:56,580 (trainer:361) INFO: 11epoch results: [train] iter_time=0.086, forward_time=0.209, loss_ctc=49.947, loss_att=34.166, acc=0.786, loss=38.900, backward_time=0.233, grad_norm=31.959, clip=100.000, loss_scale=6.305e+28, optim_step_time=0.075, optim0_lr0=3.087e-04, train_time=1.002, time=4 hours, 10 minutes and 53.73 seconds, total_count=165000, gpu_max_cached_mem_GB=36.488, [valid] loss_ctc=50.615, cer_ctc=0.282, loss_att=38.464, acc=0.680, cer=0.302, wer=1.000, loss=42.110, time=13 minutes and 58.9 seconds, total_count=23463, gpu_max_cached_mem_GB=36.488
+[ip-10-0-216-33:0/16] 2024-02-28 18:12:05,725 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-28 18:12:05,744 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/6epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-28 18:12:05,745 (trainer:290) INFO: 12/45epoch started. Estimated time to finish: 6 days, 6 hours and 2 minutes
+[ip-10-0-216-33:0/16] 2024-02-28 18:12:05,753 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 18:12:41,862 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 18:12:50,305 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e71f0cc10>)
+[ip-10-0-216-33:0/16] 2024-02-28 18:12:50,305 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-28 18:12:50,310 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 18:26:50,167 (trainer:762) INFO: 12epoch:train:1-500batch: iter_time=0.835, forward_time=0.210, loss_ctc=47.246, loss_att=32.445, acc=0.792, loss=36.885, backward_time=0.229, grad_norm=29.129, clip=100.000, loss_scale=3.169e+29, optim_step_time=0.076, optim0_lr0=3.013e-04, train_time=1.769
+[ip-10-0-216-33:0/16] 2024-02-28 18:34:28,848 (trainer:762) INFO: 12epoch:train:501-1000batch: iter_time=1.467e-04, forward_time=0.208, loss_ctc=50.504, loss_att=34.956, acc=0.793, loss=39.620, backward_time=0.232, grad_norm=30.207, clip=100.000, loss_scale=3.169e+29, optim_step_time=0.075, optim0_lr0=3.008e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-28 18:42:08,251 (trainer:762) INFO: 12epoch:train:1001-1500batch: iter_time=1.504e-04, forward_time=0.209, loss_ctc=50.480, loss_att=34.388, acc=0.790, loss=39.215, backward_time=0.235, grad_norm=31.515, clip=100.000, loss_scale=6.338e+29, optim_step_time=0.075, optim0_lr0=3.004e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-28 18:49:44,337 (trainer:762) INFO: 12epoch:train:1501-2000batch: iter_time=1.473e-04, forward_time=0.209, loss_ctc=48.986, loss_att=33.253, acc=0.790, loss=37.973, backward_time=0.237, grad_norm=29.009, clip=100.000, loss_scale=6.338e+29, optim_step_time=0.075, optim0_lr0=2.999e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-28 18:57:22,359 (trainer:762) INFO: 12epoch:train:2001-2500batch: iter_time=1.460e-04, forward_time=0.207, loss_ctc=47.194, loss_att=33.091, acc=0.791, loss=37.322, backward_time=0.233, grad_norm=29.337, clip=100.000, loss_scale=6.338e+29, optim_step_time=0.075, optim0_lr0=2.995e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-28 19:04:57,656 (trainer:762) INFO: 12epoch:train:2501-3000batch: iter_time=1.460e-04, forward_time=0.209, loss_ctc=49.786, loss_att=33.638, acc=0.785, loss=38.482, backward_time=0.235, grad_norm=29.645, clip=100.000, loss_scale=6.338e+29, optim_step_time=0.075, optim0_lr0=2.990e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 19:12:32,824 (trainer:762) INFO: 12epoch:train:3001-3500batch: iter_time=1.464e-04, forward_time=0.207, loss_ctc=48.230, loss_att=32.904, acc=0.782, loss=37.502, backward_time=0.232, grad_norm=30.739, clip=100.000, loss_scale=1.268e+30, optim_step_time=0.075, optim0_lr0=2.986e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 19:20:05,623 (trainer:762) INFO: 12epoch:train:3501-4000batch: iter_time=1.527e-04, forward_time=0.208, loss_ctc=49.262, loss_att=33.455, acc=0.779, loss=38.197, backward_time=0.238, grad_norm=44.256, clip=100.000, loss_scale=1.268e+30, optim_step_time=0.075, optim0_lr0=2.981e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-28 19:27:37,863 (trainer:762) INFO: 12epoch:train:4001-4500batch: iter_time=1.518e-04, forward_time=0.210, loss_ctc=50.647, loss_att=34.064, acc=0.791, loss=39.039, backward_time=0.231, grad_norm=29.950, clip=100.000, loss_scale=1.268e+30, optim_step_time=0.075, optim0_lr0=2.977e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-28 19:35:10,926 (trainer:762) INFO: 12epoch:train:4501-5000batch: iter_time=1.485e-04, forward_time=0.208, loss_ctc=47.005, loss_att=32.194, acc=0.785, loss=36.638, backward_time=0.235, grad_norm=28.931, clip=100.000, loss_scale=1.268e+30, optim_step_time=0.075, optim0_lr0=2.973e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-28 19:35:15,558 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 19:35:52,989 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 19:36:01,914 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed015fa60>)
+[ip-10-0-216-33:0/16] 2024-02-28 19:36:01,914 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-28 19:36:01,919 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 19:50:23,013 (trainer:762) INFO: 12epoch:train:5001-5500batch: iter_time=0.874, forward_time=0.208, loss_ctc=46.579, loss_att=31.918, acc=0.795, loss=36.316, backward_time=0.243, grad_norm=30.743, clip=100.000, loss_scale=2.535e+30, optim_step_time=0.075, optim0_lr0=2.968e-04, train_time=1.824
+[ip-10-0-216-33:0/16] 2024-02-28 19:58:03,683 (trainer:762) INFO: 12epoch:train:5501-6000batch: iter_time=1.478e-04, forward_time=0.209, loss_ctc=49.734, loss_att=34.408, acc=0.796, loss=39.006, backward_time=0.239, grad_norm=29.235, clip=100.000, loss_scale=2.535e+30, optim_step_time=0.075, optim0_lr0=2.964e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-02-28 20:05:40,374 (trainer:762) INFO: 12epoch:train:6001-6500batch: iter_time=1.480e-04, forward_time=0.209, loss_ctc=49.761, loss_att=33.965, acc=0.792, loss=38.704, backward_time=0.233, grad_norm=30.579, clip=100.000, loss_scale=2.535e+30, optim_step_time=0.075, optim0_lr0=2.960e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-28 20:13:19,967 (trainer:762) INFO: 12epoch:train:6501-7000batch: iter_time=1.422e-04, forward_time=0.208, loss_ctc=48.641, loss_att=33.008, acc=0.792, loss=37.698, backward_time=0.235, grad_norm=28.963, clip=100.000, loss_scale=2.535e+30, optim_step_time=0.075, optim0_lr0=2.955e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-28 20:20:56,847 (trainer:762) INFO: 12epoch:train:7001-7500batch: iter_time=1.471e-04, forward_time=0.208, loss_ctc=46.871, loss_att=32.937, acc=0.792, loss=37.117, backward_time=0.238, grad_norm=39.204, clip=100.000, loss_scale=5.071e+30, optim_step_time=0.075, optim0_lr0=2.951e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-28 20:28:34,722 (trainer:762) INFO: 12epoch:train:7501-8000batch: iter_time=1.470e-04, forward_time=0.206, loss_ctc=49.382, loss_att=33.338, acc=0.786, loss=38.151, backward_time=0.236, grad_norm=30.060, clip=100.000, loss_scale=5.071e+30, optim_step_time=0.075, optim0_lr0=2.947e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-28 20:36:09,645 (trainer:762) INFO: 12epoch:train:8001-8500batch: iter_time=1.430e-04, forward_time=0.208, loss_ctc=47.602, loss_att=32.568, acc=0.784, loss=37.079, backward_time=0.238, grad_norm=29.811, clip=100.000, loss_scale=5.071e+30, optim_step_time=0.075, optim0_lr0=2.942e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 20:43:39,058 (trainer:762) INFO: 12epoch:train:8501-9000batch: iter_time=1.444e-04, forward_time=0.209, loss_ctc=48.398, loss_att=32.909, acc=0.782, loss=37.556, backward_time=0.231, grad_norm=29.612, clip=100.000, loss_scale=5.071e+30, optim_step_time=0.075, optim0_lr0=2.938e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-02-28 20:51:13,915 (trainer:762) INFO: 12epoch:train:9001-9500batch: iter_time=1.437e-04, forward_time=0.209, loss_ctc=49.953, loss_att=33.638, acc=0.794, loss=38.532, backward_time=0.233, grad_norm=29.698, clip=100.000, loss_scale=1.014e+31, optim_step_time=0.075, optim0_lr0=2.934e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 20:58:51,375 (trainer:762) INFO: 12epoch:train:9501-10000batch: iter_time=1.428e-04, forward_time=0.206, loss_ctc=46.660, loss_att=31.731, acc=0.787, loss=36.210, backward_time=0.231, grad_norm=28.998, clip=100.000, loss_scale=1.014e+31, optim_step_time=0.075, optim0_lr0=2.930e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-28 20:58:56,587 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 20:59:33,501 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 20:59:42,294 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed2dd2f10>)
+[ip-10-0-216-33:0/16] 2024-02-28 20:59:42,294 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-28 20:59:42,299 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 21:14:04,831 (trainer:762) INFO: 12epoch:train:10001-10500batch: iter_time=0.890, forward_time=0.208, loss_ctc=46.086, loss_att=31.543, acc=0.797, loss=35.906, backward_time=0.235, grad_norm=28.965, clip=100.000, loss_scale=1.014e+31, optim_step_time=0.075, optim0_lr0=2.926e-04, train_time=1.827
+[ip-10-0-216-33:0/16] 2024-02-28 21:21:37,044 (trainer:762) INFO: 12epoch:train:10501-11000batch: iter_time=1.453e-04, forward_time=0.210, loss_ctc=48.939, loss_att=33.764, acc=0.798, loss=38.317, backward_time=0.236, grad_norm=39.947, clip=100.000, loss_scale=1.014e+31, optim_step_time=0.075, optim0_lr0=2.921e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-28 21:29:08,341 (trainer:762) INFO: 12epoch:train:11001-11500batch: iter_time=1.553e-04, forward_time=0.209, loss_ctc=48.674, loss_att=33.321, acc=0.795, loss=37.927, backward_time=0.234, grad_norm=29.252, clip=100.000, loss_scale=2.028e+31, optim_step_time=0.075, optim0_lr0=2.917e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-02-28 21:36:42,209 (trainer:762) INFO: 12epoch:train:11501-12000batch: iter_time=1.518e-04, forward_time=0.211, loss_ctc=47.974, loss_att=32.523, acc=0.794, loss=37.158, backward_time=0.236, grad_norm=28.828, clip=100.000, loss_scale=2.028e+31, optim_step_time=0.075, optim0_lr0=2.913e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-28 21:44:13,693 (trainer:762) INFO: 12epoch:train:12001-12500batch: iter_time=1.464e-04, forward_time=0.209, loss_ctc=46.404, loss_att=32.637, acc=0.795, loss=36.767, backward_time=0.230, grad_norm=30.410, clip=100.000, loss_scale=2.028e+31, optim_step_time=0.075, optim0_lr0=2.909e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-28 21:51:46,450 (trainer:762) INFO: 12epoch:train:12501-13000batch: iter_time=1.515e-04, forward_time=0.210, loss_ctc=48.466, loss_att=32.848, acc=0.789, loss=37.533, backward_time=0.235, grad_norm=29.161, clip=100.000, loss_scale=2.028e+31, optim_step_time=0.075, optim0_lr0=2.905e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-28 21:59:18,428 (trainer:762) INFO: 12epoch:train:13001-13500batch: iter_time=1.466e-04, forward_time=0.208, loss_ctc=46.989, loss_att=32.240, acc=0.787, loss=36.665, backward_time=0.235, grad_norm=30.940, clip=100.000, loss_scale=4.056e+31, optim_step_time=0.075, optim0_lr0=2.901e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-28 22:06:52,815 (trainer:762) INFO: 12epoch:train:13501-14000batch: iter_time=1.515e-04, forward_time=0.210, loss_ctc=48.058, loss_att=32.719, acc=0.783, loss=37.320, backward_time=0.238, grad_norm=30.501, clip=100.000, loss_scale=4.056e+31, optim_step_time=0.075, optim0_lr0=2.897e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 22:14:27,805 (trainer:762) INFO: 12epoch:train:14001-14500batch: iter_time=1.494e-04, forward_time=0.209, loss_ctc=49.487, loss_att=33.286, acc=0.795, loss=38.146, backward_time=0.236, grad_norm=29.518, clip=100.000, loss_scale=4.056e+31, optim_step_time=0.075, optim0_lr0=2.893e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-28 22:22:00,945 (trainer:762) INFO: 12epoch:train:14501-15000batch: iter_time=1.477e-04, forward_time=0.210, loss_ctc=46.186, loss_att=31.526, acc=0.789, loss=35.924, backward_time=0.239, grad_norm=29.768, clip=100.000, loss_scale=4.056e+31, optim_step_time=0.075, optim0_lr0=2.889e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-28 22:36:13,917 (trainer:361) INFO: 12epoch results: [train] iter_time=0.087, forward_time=0.209, loss_ctc=48.339, loss_att=33.040, acc=0.790, loss=37.630, backward_time=0.235, grad_norm=30.897, clip=100.000, loss_scale=1.075e+31, optim_step_time=0.075, optim0_lr0=2.950e-04, train_time=0.999, time=4 hours, 10 minutes and 11.47 seconds, total_count=180000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=50.169, cer_ctc=0.275, loss_att=36.519, acc=0.690, cer=0.348, wer=1.000, loss=40.614, time=13 minutes and 56.28 seconds, total_count=25596, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-02-28 22:36:23,367 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-28 22:36:23,385 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/7epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-28 22:36:23,386 (trainer:290) INFO: 13/45epoch started. Estimated time to finish: 6 days, 1 hour and 36 minutes
+[ip-10-0-216-33:0/16] 2024-02-28 22:36:23,393 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 22:36:59,528 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-28 22:37:07,590 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfcc7100>)
+[ip-10-0-216-33:0/16] 2024-02-28 22:37:07,590 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-28 22:37:07,595 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-28 22:51:08,150 (trainer:762) INFO: 13epoch:train:1-500batch: iter_time=0.831, forward_time=0.209, loss_ctc=46.619, loss_att=32.154, acc=0.789, loss=36.494, backward_time=0.231, grad_norm=31.718, clip=100.000, loss_scale=8.113e+31, optim_step_time=0.075, optim0_lr0=2.885e-04, train_time=1.769
+[ip-10-0-216-33:0/16] 2024-02-28 22:58:45,791 (trainer:762) INFO: 13epoch:train:501-1000batch: iter_time=1.499e-04, forward_time=0.207, loss_ctc=47.888, loss_att=32.394, acc=0.786, loss=37.042, backward_time=0.236, grad_norm=30.907, clip=100.000, loss_scale=8.113e+31, optim_step_time=0.075, optim0_lr0=2.881e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-28 23:06:21,725 (trainer:762) INFO: 13epoch:train:1001-1500batch: iter_time=1.534e-04, forward_time=0.209, loss_ctc=49.048, loss_att=33.552, acc=0.788, loss=38.201, backward_time=0.234, grad_norm=31.765, clip=100.000, loss_scale=8.113e+31, optim_step_time=0.075, optim0_lr0=2.877e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-28 23:13:54,009 (trainer:762) INFO: 13epoch:train:1501-2000batch: iter_time=1.534e-04, forward_time=0.210, loss_ctc=49.792, loss_att=33.618, acc=0.791, loss=38.470, backward_time=0.235, grad_norm=32.460, clip=100.000, loss_scale=8.113e+31, optim_step_time=0.075, optim0_lr0=2.873e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-28 23:21:28,141 (trainer:762) INFO: 13epoch:train:2001-2500batch: iter_time=1.514e-04, forward_time=0.208, loss_ctc=49.856, loss_att=34.040, acc=0.787, loss=38.785, backward_time=0.230, grad_norm=30.904, clip=100.000, loss_scale=1.623e+32, optim_step_time=0.075, optim0_lr0=2.869e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-28 23:29:02,590 (trainer:762) INFO: 13epoch:train:2501-3000batch: iter_time=1.504e-04, forward_time=0.209, loss_ctc=47.701, loss_att=32.480, acc=0.798, loss=37.046, backward_time=0.234, grad_norm=30.379, clip=100.000, loss_scale=1.623e+32, optim_step_time=0.075, optim0_lr0=2.865e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 23:36:36,996 (trainer:762) INFO: 13epoch:train:3001-3500batch: iter_time=1.509e-04, forward_time=0.210, loss_ctc=48.650, loss_att=33.054, acc=0.789, loss=37.733, backward_time=0.239, grad_norm=30.431, clip=100.000, loss_scale=1.623e+32, optim_step_time=0.075, optim0_lr0=2.861e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-28 23:44:09,541 (trainer:762) INFO: 13epoch:train:3501-4000batch: iter_time=1.485e-04, forward_time=0.209, loss_ctc=48.180, loss_att=33.595, acc=0.791, loss=37.970, backward_time=0.232, grad_norm=30.940, clip=100.000, loss_scale=1.623e+32, optim_step_time=0.075, optim0_lr0=2.857e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-28 23:51:43,255 (trainer:762) INFO: 13epoch:train:4001-4500batch: iter_time=1.484e-04, forward_time=0.210, loss_ctc=48.972, loss_att=33.520, acc=0.795, loss=38.155, backward_time=0.234, grad_norm=30.186, clip=100.000, loss_scale=3.245e+32, optim_step_time=0.075, optim0_lr0=2.853e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-28 23:59:17,445 (trainer:762) INFO: 13epoch:train:4501-5000batch: iter_time=1.479e-04, forward_time=0.209, loss_ctc=47.018, loss_att=32.249, acc=0.803, loss=36.680, backward_time=0.237, grad_norm=27.930, clip=100.000, loss_scale=3.245e+32, optim_step_time=0.075, optim0_lr0=2.849e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-28 23:59:22,570 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-28 23:59:59,351 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 00:00:08,112 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed042be50>)
+[ip-10-0-216-33:0/16] 2024-02-29 00:00:08,112 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-29 00:00:08,116 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 00:14:32,389 (trainer:762) INFO: 13epoch:train:5001-5500batch: iter_time=0.868, forward_time=0.209, loss_ctc=46.190, loss_att=31.729, acc=0.791, loss=36.067, backward_time=0.232, grad_norm=30.632, clip=100.000, loss_scale=3.245e+32, optim_step_time=0.075, optim0_lr0=2.846e-04, train_time=1.830
+[ip-10-0-216-33:0/16] 2024-02-29 00:22:03,956 (trainer:762) INFO: 13epoch:train:5501-6000batch: iter_time=1.446e-04, forward_time=0.208, loss_ctc=47.189, loss_att=31.970, acc=0.788, loss=36.536, backward_time=0.234, grad_norm=29.986, clip=100.000, loss_scale=3.245e+32, optim_step_time=0.075, optim0_lr0=2.842e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-29 00:29:34,663 (trainer:762) INFO: 13epoch:train:6001-6500batch: iter_time=1.427e-04, forward_time=0.210, loss_ctc=47.840, loss_att=32.885, acc=0.792, loss=37.372, backward_time=0.232, grad_norm=30.200, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=2.838e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-29 00:37:08,877 (trainer:762) INFO: 13epoch:train:6501-7000batch: iter_time=1.411e-04, forward_time=0.209, loss_ctc=48.897, loss_att=33.104, acc=0.794, loss=37.842, backward_time=0.230, grad_norm=30.965, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=2.834e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 00:44:39,577 (trainer:762) INFO: 13epoch:train:7001-7500batch: iter_time=1.401e-04, forward_time=0.210, loss_ctc=49.134, loss_att=33.479, acc=0.790, loss=38.176, backward_time=0.237, grad_norm=29.618, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=2.830e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-02-29 00:52:17,365 (trainer:762) INFO: 13epoch:train:7501-8000batch: iter_time=1.444e-04, forward_time=0.208, loss_ctc=46.855, loss_att=32.039, acc=0.801, loss=36.484, backward_time=0.234, grad_norm=28.698, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=2.827e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-29 00:59:49,880 (trainer:762) INFO: 13epoch:train:8001-8500batch: iter_time=1.412e-04, forward_time=0.210, loss_ctc=47.998, loss_att=32.705, acc=0.791, loss=37.293, backward_time=0.233, grad_norm=30.273, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.823e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 01:07:24,720 (trainer:762) INFO: 13epoch:train:8501-9000batch: iter_time=1.402e-04, forward_time=0.208, loss_ctc=47.542, loss_att=33.050, acc=0.794, loss=37.398, backward_time=0.236, grad_norm=30.590, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.819e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-29 01:15:01,020 (trainer:762) INFO: 13epoch:train:9001-9500batch: iter_time=1.425e-04, forward_time=0.209, loss_ctc=48.364, loss_att=33.032, acc=0.797, loss=37.632, backward_time=0.231, grad_norm=29.406, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.815e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-29 01:22:36,449 (trainer:762) INFO: 13epoch:train:9501-10000batch: iter_time=1.401e-04, forward_time=0.207, loss_ctc=46.435, loss_att=31.867, acc=0.805, loss=36.237, backward_time=0.235, grad_norm=28.688, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.812e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-29 01:22:41,486 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 01:23:19,908 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 01:23:29,241 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f05ce5b50>)
+[ip-10-0-216-33:0/16] 2024-02-29 01:23:29,242 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-29 01:23:29,246 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 01:37:43,649 (trainer:762) INFO: 13epoch:train:10001-10500batch: iter_time=0.884, forward_time=0.209, loss_ctc=45.527, loss_att=31.304, acc=0.793, loss=35.571, backward_time=0.228, grad_norm=30.135, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.808e-04, train_time=1.814
+[ip-10-0-216-33:0/16] 2024-02-29 01:45:17,701 (trainer:762) INFO: 13epoch:train:10501-11000batch: iter_time=1.442e-04, forward_time=0.209, loss_ctc=46.644, loss_att=31.692, acc=0.790, loss=36.177, backward_time=0.232, grad_norm=29.379, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.804e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 01:52:46,802 (trainer:762) INFO: 13epoch:train:11001-11500batch: iter_time=1.440e-04, forward_time=0.209, loss_ctc=47.699, loss_att=32.645, acc=0.794, loss=37.161, backward_time=0.237, grad_norm=32.783, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.801e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-02-29 02:00:24,914 (trainer:762) INFO: 13epoch:train:11501-12000batch: iter_time=1.471e-04, forward_time=0.209, loss_ctc=48.470, loss_att=32.888, acc=0.795, loss=37.563, backward_time=0.240, grad_norm=31.877, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.797e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-29 02:08:02,104 (trainer:762) INFO: 13epoch:train:12001-12500batch: iter_time=1.489e-04, forward_time=0.206, loss_ctc=48.605, loss_att=33.304, acc=0.791, loss=37.894, backward_time=0.230, grad_norm=30.608, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.793e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-29 02:15:46,780 (trainer:762) INFO: 13epoch:train:12501-13000batch: iter_time=1.490e-04, forward_time=0.206, loss_ctc=46.666, loss_att=31.917, acc=0.802, loss=36.341, backward_time=0.228, grad_norm=29.809, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.790e-04, train_time=0.929
+[ip-10-0-216-33:0/16] 2024-02-29 02:17:14,968 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 02:23:30,807 (trainer:762) INFO: 13epoch:train:13001-13500batch: iter_time=1.493e-04, forward_time=0.206, loss_ctc=47.620, loss_att=32.406, acc=0.793, loss=36.970, backward_time=0.232, grad_norm=30.534, clip=100.000, loss_scale=3.080e+33, optim_step_time=0.075, optim0_lr0=2.786e-04, train_time=0.928
+[ip-10-0-216-33:0/16] 2024-02-29 02:31:14,451 (trainer:762) INFO: 13epoch:train:13501-14000batch: iter_time=1.497e-04, forward_time=0.208, loss_ctc=47.310, loss_att=33.002, acc=0.795, loss=37.294, backward_time=0.232, grad_norm=31.058, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.782e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-02-29 02:38:51,477 (trainer:762) INFO: 13epoch:train:14001-14500batch: iter_time=1.468e-04, forward_time=0.209, loss_ctc=48.371, loss_att=33.037, acc=0.798, loss=37.638, backward_time=0.233, grad_norm=38.660, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.779e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-29 02:46:33,411 (trainer:762) INFO: 13epoch:train:14501-15000batch: iter_time=1.494e-04, forward_time=0.210, loss_ctc=46.188, loss_att=31.659, acc=0.806, loss=36.018, backward_time=0.231, grad_norm=31.145, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.775e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-02-29 03:00:36,200 (trainer:361) INFO: 13epoch results: [train] iter_time=0.086, forward_time=0.209, loss_ctc=47.776, loss_att=32.679, acc=0.794, loss=37.208, backward_time=0.233, grad_norm=30.755, clip=100.000, loss_scale=1.390e+33, optim_step_time=0.075, optim0_lr0=2.829e-04, train_time=1.000, time=4 hours, 10 minutes and 24.75 seconds, total_count=195000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=48.938, cer_ctc=0.271, loss_att=36.502, acc=0.693, cer=0.344, wer=1.000, loss=40.233, time=13 minutes and 47.76 seconds, total_count=27729, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-02-29 03:00:45,821 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-29 03:00:45,868 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/8epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-29 03:00:45,868 (trainer:290) INFO: 14/45epoch started. Estimated time to finish: 5 days, 21 hours and 10 minutes
+[ip-10-0-216-33:0/16] 2024-02-29 03:00:45,876 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 03:01:21,624 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 03:01:29,822 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed03d9670>)
+[ip-10-0-216-33:0/16] 2024-02-29 03:01:29,822 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-29 03:01:29,827 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 03:15:23,829 (trainer:762) INFO: 14epoch:train:1-500batch: iter_time=0.846, forward_time=0.210, loss_ctc=48.457, loss_att=33.156, acc=0.801, loss=37.746, backward_time=0.231, grad_norm=30.604, clip=100.000, loss_scale=4.704e+33, optim_step_time=0.075, optim0_lr0=2.772e-04, train_time=1.756
+[ip-10-0-216-33:0/16] 2024-02-29 03:22:58,749 (trainer:762) INFO: 14epoch:train:501-1000batch: iter_time=1.505e-04, forward_time=0.210, loss_ctc=47.934, loss_att=32.812, acc=0.795, loss=37.349, backward_time=0.232, grad_norm=30.876, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.768e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-29 03:30:31,281 (trainer:762) INFO: 14epoch:train:1001-1500batch: iter_time=1.524e-04, forward_time=0.209, loss_ctc=47.084, loss_att=32.496, acc=0.796, loss=36.872, backward_time=0.237, grad_norm=31.510, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.765e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 03:38:07,836 (trainer:762) INFO: 14epoch:train:1501-2000batch: iter_time=1.464e-04, forward_time=0.209, loss_ctc=47.602, loss_att=32.975, acc=0.801, loss=37.363, backward_time=0.234, grad_norm=31.192, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.761e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-29 03:40:46,854 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 03:45:41,813 (trainer:762) INFO: 14epoch:train:2001-2500batch: iter_time=1.476e-04, forward_time=0.210, loss_ctc=48.408, loss_att=33.114, acc=0.794, loss=37.702, backward_time=0.236, grad_norm=32.107, clip=100.000, loss_scale=6.035e+33, optim_step_time=0.075, optim0_lr0=2.758e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 03:53:19,854 (trainer:762) INFO: 14epoch:train:2501-3000batch: iter_time=1.507e-04, forward_time=0.208, loss_ctc=49.734, loss_att=34.081, acc=0.793, loss=38.777, backward_time=0.235, grad_norm=34.660, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.754e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-29 04:00:52,939 (trainer:762) INFO: 14epoch:train:3001-3500batch: iter_time=1.503e-04, forward_time=0.211, loss_ctc=47.953, loss_att=32.901, acc=0.797, loss=37.417, backward_time=0.232, grad_norm=31.518, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.751e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-29 04:08:24,686 (trainer:762) INFO: 14epoch:train:3501-4000batch: iter_time=1.491e-04, forward_time=0.209, loss_ctc=48.551, loss_att=32.881, acc=0.798, loss=37.582, backward_time=0.236, grad_norm=32.140, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.747e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-29 04:13:00,589 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 04:16:01,616 (trainer:762) INFO: 14epoch:train:4001-4500batch: iter_time=1.513e-04, forward_time=0.210, loss_ctc=46.760, loss_att=31.891, acc=0.802, loss=36.352, backward_time=0.235, grad_norm=31.374, clip=100.000, loss_scale=6.493e+33, optim_step_time=0.075, optim0_lr0=2.744e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-29 04:23:33,926 (trainer:762) INFO: 14epoch:train:4501-5000batch: iter_time=1.473e-04, forward_time=0.210, loss_ctc=47.024, loss_att=32.518, acc=0.801, loss=36.870, backward_time=0.229, grad_norm=30.417, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.740e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-29 04:23:39,039 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 04:24:15,567 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 04:24:23,499 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecde0d550>)
+[ip-10-0-216-33:0/16] 2024-02-29 04:24:23,499 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-29 04:24:23,504 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 04:38:48,896 (trainer:762) INFO: 14epoch:train:5001-5500batch: iter_time=0.895, forward_time=0.208, loss_ctc=47.610, loss_att=32.632, acc=0.803, loss=37.126, backward_time=0.236, grad_norm=30.402, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.737e-04, train_time=1.830
+[ip-10-0-216-33:0/16] 2024-02-29 04:39:42,436 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 04:46:20,601 (trainer:762) INFO: 14epoch:train:5501-6000batch: iter_time=1.471e-04, forward_time=0.209, loss_ctc=47.504, loss_att=32.547, acc=0.797, loss=37.034, backward_time=0.232, grad_norm=31.093, clip=100.000, loss_scale=2.893e+33, optim_step_time=0.075, optim0_lr0=2.734e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-29 04:53:54,868 (trainer:762) INFO: 14epoch:train:6001-6500batch: iter_time=1.463e-04, forward_time=0.210, loss_ctc=46.615, loss_att=32.142, acc=0.797, loss=36.484, backward_time=0.229, grad_norm=31.035, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.730e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 05:01:27,717 (trainer:762) INFO: 14epoch:train:6501-7000batch: iter_time=1.443e-04, forward_time=0.209, loss_ctc=46.759, loss_att=32.500, acc=0.802, loss=36.778, backward_time=0.232, grad_norm=31.182, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.727e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 05:09:04,644 (trainer:762) INFO: 14epoch:train:7001-7500batch: iter_time=1.454e-04, forward_time=0.208, loss_ctc=47.823, loss_att=32.745, acc=0.797, loss=37.268, backward_time=0.229, grad_norm=31.536, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.723e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-29 05:16:38,575 (trainer:762) INFO: 14epoch:train:7501-8000batch: iter_time=1.460e-04, forward_time=0.208, loss_ctc=48.897, loss_att=33.662, acc=0.794, loss=38.232, backward_time=0.228, grad_norm=34.101, clip=100.000, loss_scale=4.891e+33, optim_step_time=0.075, optim0_lr0=2.720e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 05:22:55,549 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 05:24:08,694 (trainer:762) INFO: 14epoch:train:8001-8500batch: iter_time=1.473e-04, forward_time=0.210, loss_ctc=47.713, loss_att=32.725, acc=0.798, loss=37.222, backward_time=0.229, grad_norm=30.919, clip=100.000, loss_scale=4.771e+33, optim_step_time=0.075, optim0_lr0=2.717e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-02-29 05:31:38,615 (trainer:762) INFO: 14epoch:train:8501-9000batch: iter_time=1.470e-04, forward_time=0.210, loss_ctc=48.059, loss_att=32.601, acc=0.799, loss=37.239, backward_time=0.232, grad_norm=31.573, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.713e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-02-29 05:39:12,104 (trainer:762) INFO: 14epoch:train:9001-9500batch: iter_time=1.450e-04, forward_time=0.209, loss_ctc=45.856, loss_att=31.445, acc=0.804, loss=35.769, backward_time=0.236, grad_norm=30.398, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.710e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-29 05:46:41,426 (trainer:762) INFO: 14epoch:train:9501-10000batch: iter_time=1.448e-04, forward_time=0.210, loss_ctc=46.686, loss_att=32.130, acc=0.803, loss=36.497, backward_time=0.229, grad_norm=29.878, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.707e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-02-29 05:46:46,299 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 05:47:23,462 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 05:47:31,422 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f0350d370>)
+[ip-10-0-216-33:0/16] 2024-02-29 05:47:31,423 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-29 05:47:31,427 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 06:02:01,665 (trainer:762) INFO: 14epoch:train:10001-10500batch: iter_time=0.884, forward_time=0.210, loss_ctc=47.312, loss_att=32.392, acc=0.804, loss=36.868, backward_time=0.232, grad_norm=30.633, clip=100.000, loss_scale=3.017e+33, optim_step_time=0.075, optim0_lr0=2.703e-04, train_time=1.840
+[ip-10-0-216-33:0/16] 2024-02-29 06:09:35,811 (trainer:762) INFO: 14epoch:train:10501-11000batch: iter_time=1.532e-04, forward_time=0.209, loss_ctc=47.094, loss_att=32.303, acc=0.798, loss=36.740, backward_time=0.229, grad_norm=30.746, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.700e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 06:17:09,889 (trainer:762) INFO: 14epoch:train:11001-11500batch: iter_time=1.539e-04, forward_time=0.210, loss_ctc=46.141, loss_att=31.839, acc=0.799, loss=36.130, backward_time=0.234, grad_norm=29.979, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.697e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 06:24:41,430 (trainer:762) INFO: 14epoch:train:11501-12000batch: iter_time=1.490e-04, forward_time=0.210, loss_ctc=46.445, loss_att=32.172, acc=0.804, loss=36.454, backward_time=0.232, grad_norm=30.652, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.694e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-29 06:26:26,759 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 06:32:18,062 (trainer:762) INFO: 14epoch:train:12001-12500batch: iter_time=1.499e-04, forward_time=0.209, loss_ctc=47.285, loss_att=32.466, acc=0.798, loss=36.912, backward_time=0.237, grad_norm=30.464, clip=100.000, loss_scale=3.194e+33, optim_step_time=0.075, optim0_lr0=2.690e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-29 06:39:59,419 (trainer:762) INFO: 14epoch:train:12501-13000batch: iter_time=1.516e-04, forward_time=0.208, loss_ctc=48.607, loss_att=33.428, acc=0.796, loss=37.982, backward_time=0.234, grad_norm=32.923, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.687e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-02-29 06:47:41,398 (trainer:762) INFO: 14epoch:train:13001-13500batch: iter_time=1.539e-04, forward_time=0.208, loss_ctc=47.222, loss_att=32.356, acc=0.800, loss=36.816, backward_time=0.229, grad_norm=29.806, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.684e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-02-29 06:55:19,417 (trainer:762) INFO: 14epoch:train:13501-14000batch: iter_time=1.539e-04, forward_time=0.207, loss_ctc=47.426, loss_att=32.212, acc=0.801, loss=36.776, backward_time=0.231, grad_norm=31.068, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.681e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-29 07:02:58,079 (trainer:762) INFO: 14epoch:train:14001-14500batch: iter_time=1.534e-04, forward_time=0.208, loss_ctc=45.572, loss_att=31.395, acc=0.804, loss=35.648, backward_time=0.234, grad_norm=42.193, clip=100.000, loss_scale=4.590e+33, optim_step_time=0.075, optim0_lr0=2.677e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-29 07:10:36,438 (trainer:762) INFO: 14epoch:train:14501-15000batch: iter_time=1.503e-04, forward_time=0.209, loss_ctc=46.146, loss_att=31.907, acc=0.804, loss=36.179, backward_time=0.234, grad_norm=29.988, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.674e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-29 07:25:08,680 (trainer:361) INFO: 14epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=47.409, loss_att=32.548, acc=0.799, loss=37.006, backward_time=0.232, grad_norm=31.566, clip=100.000, loss_scale=4.209e+33, optim_step_time=0.075, optim0_lr0=2.722e-04, train_time=0.999, time=4 hours, 10 minutes and 6.22 seconds, total_count=210000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=46.186, cer_ctc=0.256, loss_att=35.478, acc=0.700, cer=0.337, wer=1.000, loss=38.690, time=14 minutes and 16.28 seconds, total_count=29862, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-02-29 07:25:18,556 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-29 07:25:18,576 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/9epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-29 07:25:18,576 (trainer:290) INFO: 15/45epoch started. Estimated time to finish: 5 days, 16 hours and 45 minutes
+[ip-10-0-216-33:0/16] 2024-02-29 07:25:18,584 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 07:25:54,414 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 07:26:02,262 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecef7c460>)
+[ip-10-0-216-33:0/16] 2024-02-29 07:26:02,263 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-29 07:26:02,267 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 07:39:56,183 (trainer:762) INFO: 15epoch:train:1-500batch: iter_time=0.839, forward_time=0.209, loss_ctc=46.119, loss_att=31.499, acc=0.796, loss=35.885, backward_time=0.237, grad_norm=31.479, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.671e-04, train_time=1.755
+[ip-10-0-216-33:0/16] 2024-02-29 07:40:33,891 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 07:47:33,496 (trainer:762) INFO: 15epoch:train:501-1000batch: iter_time=1.505e-04, forward_time=0.210, loss_ctc=47.063, loss_att=32.784, acc=0.793, loss=37.068, backward_time=0.235, grad_norm=32.957, clip=100.000, loss_scale=2.809e+33, optim_step_time=0.075, optim0_lr0=2.668e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-29 07:55:10,981 (trainer:762) INFO: 15epoch:train:1001-1500batch: iter_time=1.530e-04, forward_time=0.210, loss_ctc=47.556, loss_att=32.834, acc=0.803, loss=37.250, backward_time=0.239, grad_norm=30.602, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.665e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-29 08:02:52,919 (trainer:762) INFO: 15epoch:train:1501-2000batch: iter_time=1.544e-04, forward_time=0.207, loss_ctc=48.309, loss_att=33.155, acc=0.799, loss=37.701, backward_time=0.232, grad_norm=31.081, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.662e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-02-29 08:10:34,465 (trainer:762) INFO: 15epoch:train:2001-2500batch: iter_time=1.493e-04, forward_time=0.208, loss_ctc=45.815, loss_att=31.394, acc=0.799, loss=35.720, backward_time=0.231, grad_norm=30.358, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.658e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-02-29 08:18:19,091 (trainer:762) INFO: 15epoch:train:2501-3000batch: iter_time=1.484e-04, forward_time=0.206, loss_ctc=48.081, loss_att=32.840, acc=0.796, loss=37.412, backward_time=0.231, grad_norm=34.974, clip=100.000, loss_scale=4.974e+33, optim_step_time=0.075, optim0_lr0=2.655e-04, train_time=0.929
+[ip-10-0-216-33:0/16] 2024-02-29 08:25:59,129 (trainer:762) INFO: 15epoch:train:3001-3500batch: iter_time=1.488e-04, forward_time=0.208, loss_ctc=46.593, loss_att=32.330, acc=0.791, loss=36.609, backward_time=0.239, grad_norm=33.152, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.652e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-02-29 08:29:26,323 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 08:33:39,376 (trainer:762) INFO: 15epoch:train:3501-4000batch: iter_time=1.512e-04, forward_time=0.209, loss_ctc=46.609, loss_att=32.244, acc=0.796, loss=36.553, backward_time=0.238, grad_norm=32.874, clip=100.000, loss_scale=3.751e+33, optim_step_time=0.075, optim0_lr0=2.649e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-02-29 08:41:23,546 (trainer:762) INFO: 15epoch:train:4001-4500batch: iter_time=1.524e-04, forward_time=0.207, loss_ctc=46.331, loss_att=32.137, acc=0.798, loss=36.395, backward_time=0.236, grad_norm=31.982, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.646e-04, train_time=0.928
+[ip-10-0-216-33:0/16] 2024-02-29 08:48:58,356 (trainer:762) INFO: 15epoch:train:4501-5000batch: iter_time=1.482e-04, forward_time=0.209, loss_ctc=45.081, loss_att=31.207, acc=0.800, loss=35.369, backward_time=0.235, grad_norm=31.673, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.643e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-29 08:49:03,071 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 08:49:39,637 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 08:49:47,631 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0536d30>)
+[ip-10-0-216-33:0/16] 2024-02-29 08:49:47,631 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-29 08:49:47,636 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 09:06:14,753 (trainer:762) INFO: 15epoch:train:5001-5500batch: iter_time=1.159, forward_time=0.209, loss_ctc=45.306, loss_att=30.955, acc=0.799, loss=35.260, backward_time=0.229, grad_norm=30.401, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.640e-04, train_time=2.073
+[ip-10-0-216-33:0/16] 2024-02-29 09:13:52,781 (trainer:762) INFO: 15epoch:train:5501-6000batch: iter_time=1.435e-04, forward_time=0.209, loss_ctc=46.330, loss_att=32.165, acc=0.796, loss=36.415, backward_time=0.235, grad_norm=32.285, clip=100.000, loss_scale=4.034e+33, optim_step_time=0.075, optim0_lr0=2.637e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-29 09:21:30,845 (trainer:762) INFO: 15epoch:train:6001-6500batch: iter_time=1.487e-04, forward_time=0.209, loss_ctc=47.084, loss_att=32.375, acc=0.805, loss=36.788, backward_time=0.239, grad_norm=30.458, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.634e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-29 09:29:09,795 (trainer:762) INFO: 15epoch:train:6501-7000batch: iter_time=1.454e-04, forward_time=0.209, loss_ctc=47.678, loss_att=32.734, acc=0.802, loss=37.217, backward_time=0.233, grad_norm=31.896, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.631e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-02-29 09:36:46,148 (trainer:762) INFO: 15epoch:train:7001-7500batch: iter_time=1.493e-04, forward_time=0.207, loss_ctc=45.469, loss_att=31.183, acc=0.801, loss=35.469, backward_time=0.234, grad_norm=30.086, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.628e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-29 09:42:09,803 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 09:44:27,449 (trainer:762) INFO: 15epoch:train:7501-8000batch: iter_time=1.503e-04, forward_time=0.208, loss_ctc=47.592, loss_att=32.550, acc=0.797, loss=37.062, backward_time=0.235, grad_norm=32.658, clip=100.000, loss_scale=6.503e+33, optim_step_time=0.075, optim0_lr0=2.625e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-02-29 09:51:58,898 (trainer:762) INFO: 15epoch:train:8001-8500batch: iter_time=1.503e-04, forward_time=0.208, loss_ctc=45.755, loss_att=31.793, acc=0.794, loss=35.982, backward_time=0.233, grad_norm=31.877, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.622e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-29 09:59:31,436 (trainer:762) INFO: 15epoch:train:8501-9000batch: iter_time=1.505e-04, forward_time=0.210, loss_ctc=46.400, loss_att=32.045, acc=0.798, loss=36.352, backward_time=0.242, grad_norm=33.589, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.619e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 10:00:17,495 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 10:07:07,011 (trainer:762) INFO: 15epoch:train:9001-9500batch: iter_time=1.486e-04, forward_time=0.208, loss_ctc=45.676, loss_att=31.652, acc=0.801, loss=35.859, backward_time=0.240, grad_norm=32.166, clip=100.000, loss_scale=2.851e+33, optim_step_time=0.075, optim0_lr0=2.616e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-29 10:14:36,779 (trainer:762) INFO: 15epoch:train:9501-10000batch: iter_time=1.450e-04, forward_time=0.210, loss_ctc=44.660, loss_att=30.909, acc=0.801, loss=35.035, backward_time=0.234, grad_norm=36.789, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.613e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-02-29 10:14:41,613 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 10:15:18,993 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 10:15:26,935 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5efc668cd0>)
+[ip-10-0-216-33:0/16] 2024-02-29 10:15:26,936 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-29 10:15:26,940 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 10:29:51,771 (trainer:762) INFO: 15epoch:train:10001-10500batch: iter_time=0.868, forward_time=0.208, loss_ctc=45.340, loss_att=30.977, acc=0.800, loss=35.286, backward_time=0.235, grad_norm=31.875, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.610e-04, train_time=1.830
+[ip-10-0-216-33:0/16] 2024-02-29 10:37:28,110 (trainer:762) INFO: 15epoch:train:10501-11000batch: iter_time=1.483e-04, forward_time=0.208, loss_ctc=45.994, loss_att=32.062, acc=0.797, loss=36.241, backward_time=0.239, grad_norm=33.527, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.607e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-29 10:45:00,899 (trainer:762) INFO: 15epoch:train:11001-11500batch: iter_time=1.511e-04, forward_time=0.209, loss_ctc=46.931, loss_att=32.299, acc=0.806, loss=36.689, backward_time=0.231, grad_norm=33.846, clip=100.000, loss_scale=4.933e+33, optim_step_time=0.075, optim0_lr0=2.604e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 10:52:40,701 (trainer:762) INFO: 15epoch:train:11501-12000batch: iter_time=1.520e-04, forward_time=0.208, loss_ctc=47.390, loss_att=32.516, acc=0.803, loss=36.978, backward_time=0.231, grad_norm=31.445, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.601e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-02-29 11:00:17,876 (trainer:762) INFO: 15epoch:train:12001-12500batch: iter_time=1.450e-04, forward_time=0.209, loss_ctc=45.182, loss_att=30.923, acc=0.802, loss=35.201, backward_time=0.232, grad_norm=30.653, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.598e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-29 11:02:44,731 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 11:07:55,294 (trainer:762) INFO: 15epoch:train:12501-13000batch: iter_time=1.454e-04, forward_time=0.209, loss_ctc=47.292, loss_att=32.283, acc=0.798, loss=36.785, backward_time=0.236, grad_norm=33.023, clip=100.000, loss_scale=3.423e+33, optim_step_time=0.075, optim0_lr0=2.595e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-29 11:15:33,221 (trainer:762) INFO: 15epoch:train:13001-13500batch: iter_time=1.460e-04, forward_time=0.207, loss_ctc=45.603, loss_att=31.726, acc=0.795, loss=35.889, backward_time=0.231, grad_norm=32.876, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.592e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-29 11:23:08,668 (trainer:762) INFO: 15epoch:train:13501-14000batch: iter_time=1.504e-04, forward_time=0.209, loss_ctc=45.787, loss_att=31.705, acc=0.799, loss=35.930, backward_time=0.231, grad_norm=32.341, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.589e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-29 11:30:45,005 (trainer:762) INFO: 15epoch:train:14001-14500batch: iter_time=1.467e-04, forward_time=0.209, loss_ctc=45.690, loss_att=31.580, acc=0.802, loss=35.813, backward_time=0.234, grad_norm=31.903, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.586e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-29 11:38:19,611 (trainer:762) INFO: 15epoch:train:14501-15000batch: iter_time=1.441e-04, forward_time=0.207, loss_ctc=44.266, loss_att=30.735, acc=0.802, loss=34.794, backward_time=0.235, grad_norm=31.053, clip=100.000, loss_scale=4.362e+33, optim_step_time=0.075, optim0_lr0=2.583e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-29 11:52:35,968 (trainer:361) INFO: 15epoch results: [train] iter_time=0.096, forward_time=0.208, loss_ctc=46.299, loss_att=31.920, acc=0.799, loss=36.234, backward_time=0.235, grad_norm=32.196, clip=100.000, loss_scale=3.851e+33, optim_step_time=0.075, optim0_lr0=2.627e-04, train_time=1.012, time=4 hours, 13 minutes and 15.81 seconds, total_count=225000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=46.031, cer_ctc=0.254, loss_att=35.781, acc=0.703, cer=0.341, wer=1.000, loss=38.856, time=14 minutes and 1.24 seconds, total_count=31995, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-02-29 11:52:45,582 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-29 11:52:45,601 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/10epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-29 11:52:45,601 (trainer:290) INFO: 16/45epoch started. Estimated time to finish: 5 days, 12 hours and 26 minutes
+[ip-10-0-216-33:0/16] 2024-02-29 11:52:45,608 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 11:53:21,588 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 11:53:29,872 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf610d90>)
+[ip-10-0-216-33:0/16] 2024-02-29 11:53:29,872 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-29 11:53:29,877 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 12:07:23,652 (trainer:762) INFO: 16epoch:train:1-500batch: iter_time=0.833, forward_time=0.210, loss_ctc=45.586, loss_att=31.106, acc=0.808, loss=35.450, backward_time=0.233, grad_norm=32.273, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.581e-04, train_time=1.756
+[ip-10-0-216-33:0/16] 2024-02-29 12:15:02,054 (trainer:762) INFO: 16epoch:train:501-1000batch: iter_time=1.477e-04, forward_time=0.208, loss_ctc=44.590, loss_att=30.760, acc=0.798, loss=34.909, backward_time=0.229, grad_norm=33.143, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.578e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-29 12:15:32,222 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 12:22:40,227 (trainer:762) INFO: 16epoch:train:1001-1500batch: iter_time=1.483e-04, forward_time=0.209, loss_ctc=49.477, loss_att=34.189, acc=0.794, loss=38.776, backward_time=0.227, grad_norm=35.199, clip=100.000, loss_scale=2.763e+33, optim_step_time=0.075, optim0_lr0=2.575e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-29 12:30:18,629 (trainer:762) INFO: 16epoch:train:1501-2000batch: iter_time=1.457e-04, forward_time=0.209, loss_ctc=46.816, loss_att=32.081, acc=0.803, loss=36.501, backward_time=0.240, grad_norm=33.325, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.572e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-29 12:37:53,278 (trainer:762) INFO: 16epoch:train:2001-2500batch: iter_time=1.485e-04, forward_time=0.209, loss_ctc=44.470, loss_att=30.456, acc=0.809, loss=34.660, backward_time=0.236, grad_norm=30.769, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.569e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-29 12:45:34,053 (trainer:762) INFO: 16epoch:train:2501-3000batch: iter_time=1.501e-04, forward_time=0.209, loss_ctc=48.605, loss_att=33.627, acc=0.801, loss=38.121, backward_time=0.238, grad_norm=35.051, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.566e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-02-29 12:49:14,658 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 12:53:11,106 (trainer:762) INFO: 16epoch:train:3001-3500batch: iter_time=1.452e-04, forward_time=0.210, loss_ctc=48.203, loss_att=33.129, acc=0.798, loss=37.651, backward_time=0.233, grad_norm=32.731, clip=100.000, loss_scale=3.684e+33, optim_step_time=0.075, optim0_lr0=2.564e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-02-29 13:00:51,916 (trainer:762) INFO: 16epoch:train:3501-4000batch: iter_time=1.462e-04, forward_time=0.206, loss_ctc=49.027, loss_att=32.877, acc=0.798, loss=37.722, backward_time=0.235, grad_norm=32.418, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.561e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-02-29 13:08:34,120 (trainer:762) INFO: 16epoch:train:4001-4500batch: iter_time=1.455e-04, forward_time=0.209, loss_ctc=45.388, loss_att=31.030, acc=0.808, loss=35.337, backward_time=0.231, grad_norm=30.972, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.558e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-02-29 13:16:11,658 (trainer:762) INFO: 16epoch:train:4501-5000batch: iter_time=1.478e-04, forward_time=0.208, loss_ctc=47.713, loss_att=32.581, acc=0.798, loss=37.120, backward_time=0.233, grad_norm=32.607, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.555e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-29 13:16:16,197 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 13:16:53,470 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 13:17:02,297 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfcfbf10>)
+[ip-10-0-216-33:0/16] 2024-02-29 13:17:02,298 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-29 13:17:02,302 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 13:31:20,942 (trainer:762) INFO: 16epoch:train:5001-5500batch: iter_time=0.876, forward_time=0.209, loss_ctc=44.777, loss_att=30.527, acc=0.810, loss=34.802, backward_time=0.231, grad_norm=30.894, clip=100.000, loss_scale=3.931e+33, optim_step_time=0.075, optim0_lr0=2.552e-04, train_time=1.818
+[ip-10-0-216-33:0/16] 2024-02-29 13:38:53,721 (trainer:762) INFO: 16epoch:train:5501-6000batch: iter_time=1.428e-04, forward_time=0.208, loss_ctc=43.873, loss_att=30.173, acc=0.800, loss=34.283, backward_time=0.237, grad_norm=33.859, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.550e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 13:46:23,381 (trainer:762) INFO: 16epoch:train:6001-6500batch: iter_time=1.462e-04, forward_time=0.210, loss_ctc=49.051, loss_att=33.798, acc=0.797, loss=38.374, backward_time=0.237, grad_norm=33.185, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.547e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-02-29 13:53:55,881 (trainer:762) INFO: 16epoch:train:6501-7000batch: iter_time=1.436e-04, forward_time=0.209, loss_ctc=46.206, loss_att=31.597, acc=0.805, loss=35.980, backward_time=0.238, grad_norm=31.791, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.544e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 14:01:25,819 (trainer:762) INFO: 16epoch:train:7001-7500batch: iter_time=1.398e-04, forward_time=0.210, loss_ctc=43.965, loss_att=30.172, acc=0.810, loss=34.310, backward_time=0.237, grad_norm=31.106, clip=100.000, loss_scale=7.861e+33, optim_step_time=0.075, optim0_lr0=2.541e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-02-29 14:01:35,131 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 14:08:57,749 (trainer:762) INFO: 16epoch:train:7501-8000batch: iter_time=1.427e-04, forward_time=0.209, loss_ctc=47.733, loss_att=33.123, acc=0.804, loss=37.506, backward_time=0.237, grad_norm=31.903, clip=100.000, loss_scale=5.286e+33, optim_step_time=0.075, optim0_lr0=2.539e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-29 14:16:33,066 (trainer:762) INFO: 16epoch:train:8001-8500batch: iter_time=1.457e-04, forward_time=0.210, loss_ctc=47.652, loss_att=32.681, acc=0.800, loss=37.172, backward_time=0.233, grad_norm=32.193, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.536e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-29 14:23:45,676 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 14:24:13,186 (trainer:762) INFO: 16epoch:train:8501-9000batch: iter_time=1.428e-04, forward_time=0.207, loss_ctc=48.539, loss_att=32.466, acc=0.801, loss=37.288, backward_time=0.232, grad_norm=32.641, clip=100.000, loss_scale=5.036e+33, optim_step_time=0.075, optim0_lr0=2.533e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-02-29 14:31:48,248 (trainer:762) INFO: 16epoch:train:9001-9500batch: iter_time=1.405e-04, forward_time=0.210, loss_ctc=44.877, loss_att=30.729, acc=0.809, loss=34.973, backward_time=0.229, grad_norm=30.372, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.531e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-29 14:39:27,489 (trainer:762) INFO: 16epoch:train:9501-10000batch: iter_time=1.403e-04, forward_time=0.206, loss_ctc=46.766, loss_att=32.006, acc=0.800, loss=36.434, backward_time=0.229, grad_norm=32.226, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.528e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-02-29 14:39:32,321 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 14:40:10,350 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 14:40:19,258 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f08b99fa0>)
+[ip-10-0-216-33:0/16] 2024-02-29 14:40:19,258 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-29 14:40:19,263 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 14:50:48,562 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 14:58:23,585 (trainer:762) INFO: 16epoch:train:10001-10500batch: iter_time=1.313, forward_time=0.208, loss_ctc=44.460, loss_att=30.292, acc=0.811, loss=34.543, backward_time=0.233, grad_norm=31.395, clip=100.000, loss_scale=1.301e+33, optim_step_time=0.075, optim0_lr0=2.525e-04, train_time=2.272
+[ip-10-0-216-33:0/16] 2024-02-29 15:05:57,620 (trainer:762) INFO: 16epoch:train:10501-11000batch: iter_time=1.494e-04, forward_time=0.209, loss_ctc=43.599, loss_att=30.075, acc=0.801, loss=34.132, backward_time=0.230, grad_norm=33.606, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.523e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 15:13:36,355 (trainer:762) INFO: 16epoch:train:11001-11500batch: iter_time=1.501e-04, forward_time=0.207, loss_ctc=48.706, loss_att=33.602, acc=0.798, loss=38.133, backward_time=0.232, grad_norm=32.752, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.520e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-29 15:21:08,884 (trainer:762) INFO: 16epoch:train:11501-12000batch: iter_time=1.464e-04, forward_time=0.210, loss_ctc=46.103, loss_att=31.414, acc=0.806, loss=35.821, backward_time=0.232, grad_norm=32.872, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.517e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 15:28:40,086 (trainer:762) INFO: 16epoch:train:12001-12500batch: iter_time=1.467e-04, forward_time=0.209, loss_ctc=43.752, loss_att=30.075, acc=0.811, loss=34.178, backward_time=0.231, grad_norm=30.701, clip=100.000, loss_scale=2.591e+33, optim_step_time=0.075, optim0_lr0=2.515e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-02-29 15:36:16,253 (trainer:762) INFO: 16epoch:train:12501-13000batch: iter_time=1.487e-04, forward_time=0.210, loss_ctc=47.597, loss_att=33.026, acc=0.805, loss=37.397, backward_time=0.230, grad_norm=31.558, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.512e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-29 15:43:51,970 (trainer:762) INFO: 16epoch:train:13001-13500batch: iter_time=1.529e-04, forward_time=0.210, loss_ctc=47.430, loss_att=32.531, acc=0.802, loss=37.001, backward_time=0.236, grad_norm=31.637, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.509e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-29 15:51:28,310 (trainer:762) INFO: 16epoch:train:13501-14000batch: iter_time=1.479e-04, forward_time=0.209, loss_ctc=48.417, loss_att=32.452, acc=0.801, loss=37.242, backward_time=0.231, grad_norm=32.235, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.507e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-29 15:59:01,207 (trainer:762) INFO: 16epoch:train:14001-14500batch: iter_time=1.478e-04, forward_time=0.210, loss_ctc=44.262, loss_att=30.388, acc=0.812, loss=34.551, backward_time=0.229, grad_norm=30.453, clip=100.000, loss_scale=5.182e+33, optim_step_time=0.075, optim0_lr0=2.504e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-29 16:06:34,850 (trainer:762) INFO: 16epoch:train:14501-15000batch: iter_time=1.478e-04, forward_time=0.210, loss_ctc=46.758, loss_att=32.038, acc=0.801, loss=36.454, backward_time=0.233, grad_norm=32.396, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.501e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-02-29 16:20:48,058 (trainer:361) INFO: 16epoch results: [train] iter_time=0.101, forward_time=0.209, loss_ctc=46.480, loss_att=31.833, acc=0.803, loss=36.227, backward_time=0.233, grad_norm=32.275, clip=100.000, loss_scale=3.548e+33, optim_step_time=0.075, optim0_lr0=2.540e-04, train_time=1.015, time=4 hours, 14 minutes and 4.34 seconds, total_count=240000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=43.907, cer_ctc=0.245, loss_att=32.328, acc=0.727, cer=0.289, wer=0.998, loss=35.802, time=13 minutes and 57.8 seconds, total_count=34128, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-02-29 16:20:57,536 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-29 16:20:57,559 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/11epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-29 16:20:57,559 (trainer:290) INFO: 17/45epoch started. Estimated time to finish: 5 days, 8 hours and 8 minutes
+[ip-10-0-216-33:0/16] 2024-02-29 16:20:57,566 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 16:21:33,444 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 16:21:41,312 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf4ea0d0>)
+[ip-10-0-216-33:0/16] 2024-02-29 16:21:41,313 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-29 16:21:41,317 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 16:32:28,123 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 16:35:34,685 (trainer:762) INFO: 17epoch:train:1-500batch: iter_time=0.845, forward_time=0.210, loss_ctc=44.419, loss_att=30.623, acc=0.806, loss=34.762, backward_time=0.235, grad_norm=31.517, clip=100.000, loss_scale=4.121e+33, optim_step_time=0.075, optim0_lr0=2.499e-04, train_time=1.754
+[ip-10-0-216-33:0/16] 2024-02-29 16:43:10,549 (trainer:762) INFO: 17epoch:train:501-1000batch: iter_time=1.492e-04, forward_time=0.208, loss_ctc=43.193, loss_att=29.559, acc=0.807, loss=33.649, backward_time=0.233, grad_norm=30.523, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.076, optim0_lr0=2.496e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-29 16:50:43,142 (trainer:762) INFO: 17epoch:train:1001-1500batch: iter_time=1.521e-04, forward_time=0.209, loss_ctc=44.344, loss_att=30.363, acc=0.802, loss=34.557, backward_time=0.233, grad_norm=33.539, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.494e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 16:58:19,755 (trainer:762) INFO: 17epoch:train:1501-2000batch: iter_time=1.477e-04, forward_time=0.209, loss_ctc=45.780, loss_att=31.445, acc=0.803, loss=35.745, backward_time=0.238, grad_norm=32.079, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.491e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-02-29 17:05:52,452 (trainer:762) INFO: 17epoch:train:2001-2500batch: iter_time=1.446e-04, forward_time=0.208, loss_ctc=43.906, loss_att=30.219, acc=0.801, loss=34.325, backward_time=0.240, grad_norm=32.683, clip=100.000, loss_scale=3.666e+33, optim_step_time=0.076, optim0_lr0=2.488e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 17:13:26,772 (trainer:762) INFO: 17epoch:train:2501-3000batch: iter_time=1.516e-04, forward_time=0.210, loss_ctc=47.110, loss_att=32.326, acc=0.802, loss=36.761, backward_time=0.232, grad_norm=33.149, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.076, optim0_lr0=2.486e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 17:21:05,355 (trainer:762) INFO: 17epoch:train:3001-3500batch: iter_time=1.474e-04, forward_time=0.209, loss_ctc=47.253, loss_att=32.316, acc=0.805, loss=36.797, backward_time=0.236, grad_norm=34.024, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.076, optim0_lr0=2.483e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-29 17:28:40,213 (trainer:762) INFO: 17epoch:train:3501-4000batch: iter_time=1.448e-04, forward_time=0.209, loss_ctc=45.861, loss_att=31.239, acc=0.805, loss=35.626, backward_time=0.239, grad_norm=32.415, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.076, optim0_lr0=2.481e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-29 17:36:18,450 (trainer:762) INFO: 17epoch:train:4001-4500batch: iter_time=1.488e-04, forward_time=0.210, loss_ctc=45.194, loss_att=31.308, acc=0.804, loss=35.474, backward_time=0.235, grad_norm=37.064, clip=100.000, loss_scale=7.332e+33, optim_step_time=0.075, optim0_lr0=2.478e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-02-29 17:40:51,753 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 17:43:51,108 (trainer:762) INFO: 17epoch:train:4501-5000batch: iter_time=1.451e-04, forward_time=0.210, loss_ctc=44.455, loss_att=30.365, acc=0.805, loss=34.592, backward_time=0.234, grad_norm=33.685, clip=100.000, loss_scale=8.356e+33, optim_step_time=0.075, optim0_lr0=2.476e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 17:43:55,914 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 17:44:33,252 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 17:44:41,252 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed4e57c10>)
+[ip-10-0-216-33:0/16] 2024-02-29 17:44:41,253 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-29 17:44:41,258 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 17:59:10,998 (trainer:762) INFO: 17epoch:train:5001-5500batch: iter_time=0.881, forward_time=0.209, loss_ctc=44.000, loss_att=30.526, acc=0.807, loss=34.568, backward_time=0.230, grad_norm=31.533, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.473e-04, train_time=1.840
+[ip-10-0-216-33:0/16] 2024-02-29 18:06:49,657 (trainer:762) INFO: 17epoch:train:5501-6000batch: iter_time=1.436e-04, forward_time=0.206, loss_ctc=42.599, loss_att=29.250, acc=0.809, loss=33.255, backward_time=0.237, grad_norm=30.896, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.471e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-02-29 18:14:21,787 (trainer:762) INFO: 17epoch:train:6001-6500batch: iter_time=1.470e-04, forward_time=0.209, loss_ctc=44.240, loss_att=30.187, acc=0.803, loss=34.403, backward_time=0.235, grad_norm=33.668, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.468e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-29 18:21:55,853 (trainer:762) INFO: 17epoch:train:6501-7000batch: iter_time=1.514e-04, forward_time=0.207, loss_ctc=45.145, loss_att=31.001, acc=0.805, loss=35.245, backward_time=0.232, grad_norm=32.209, clip=100.000, loss_scale=7.217e+33, optim_step_time=0.075, optim0_lr0=2.466e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 18:29:24,866 (trainer:762) INFO: 17epoch:train:7001-7500batch: iter_time=1.522e-04, forward_time=0.208, loss_ctc=43.583, loss_att=30.092, acc=0.801, loss=34.139, backward_time=0.234, grad_norm=32.585, clip=100.000, loss_scale=1.038e+34, optim_step_time=0.075, optim0_lr0=2.463e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-02-29 18:33:34,561 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 18:36:56,755 (trainer:762) INFO: 17epoch:train:7501-8000batch: iter_time=1.534e-04, forward_time=0.209, loss_ctc=46.432, loss_att=31.922, acc=0.804, loss=36.275, backward_time=0.235, grad_norm=33.129, clip=100.000, loss_scale=8.043e+33, optim_step_time=0.075, optim0_lr0=2.461e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-29 18:44:30,967 (trainer:762) INFO: 17epoch:train:8001-8500batch: iter_time=1.503e-04, forward_time=0.210, loss_ctc=46.788, loss_att=32.068, acc=0.807, loss=36.484, backward_time=0.239, grad_norm=32.685, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.458e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 18:51:59,397 (trainer:762) INFO: 17epoch:train:8501-9000batch: iter_time=1.538e-04, forward_time=0.210, loss_ctc=45.428, loss_att=30.924, acc=0.807, loss=35.276, backward_time=0.234, grad_norm=31.077, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.456e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-02-29 18:57:41,054 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 18:59:31,788 (trainer:762) INFO: 17epoch:train:9001-9500batch: iter_time=1.490e-04, forward_time=0.209, loss_ctc=44.566, loss_att=30.966, acc=0.806, loss=35.046, backward_time=0.232, grad_norm=32.361, clip=100.000, loss_scale=4.558e+33, optim_step_time=0.075, optim0_lr0=2.453e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 19:07:04,356 (trainer:762) INFO: 17epoch:train:9501-10000batch: iter_time=1.437e-04, forward_time=0.209, loss_ctc=43.944, loss_att=30.080, acc=0.806, loss=34.239, backward_time=0.239, grad_norm=33.593, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.451e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 19:07:09,976 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 19:07:46,651 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 19:07:54,747 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f040599d0>)
+[ip-10-0-216-33:0/16] 2024-02-29 19:07:54,748 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-29 19:07:54,752 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 19:22:17,782 (trainer:762) INFO: 17epoch:train:10001-10500batch: iter_time=0.873, forward_time=0.211, loss_ctc=43.669, loss_att=30.331, acc=0.808, loss=34.333, backward_time=0.237, grad_norm=31.640, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.448e-04, train_time=1.827
+[ip-10-0-216-33:0/16] 2024-02-29 19:29:50,337 (trainer:762) INFO: 17epoch:train:10501-11000batch: iter_time=1.503e-04, forward_time=0.208, loss_ctc=42.399, loss_att=29.104, acc=0.810, loss=33.093, backward_time=0.233, grad_norm=30.816, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.446e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 19:37:22,238 (trainer:762) INFO: 17epoch:train:11001-11500batch: iter_time=1.532e-04, forward_time=0.210, loss_ctc=43.426, loss_att=29.880, acc=0.805, loss=33.944, backward_time=0.237, grad_norm=31.936, clip=100.000, loss_scale=3.230e+33, optim_step_time=0.075, optim0_lr0=2.443e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-02-29 19:40:46,105 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 19:45:00,071 (trainer:762) INFO: 17epoch:train:11501-12000batch: iter_time=1.481e-04, forward_time=0.208, loss_ctc=44.807, loss_att=30.784, acc=0.806, loss=34.991, backward_time=0.232, grad_norm=32.041, clip=100.000, loss_scale=3.741e+33, optim_step_time=0.075, optim0_lr0=2.441e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-02-29 19:52:33,048 (trainer:762) INFO: 17epoch:train:12001-12500batch: iter_time=1.493e-04, forward_time=0.211, loss_ctc=43.122, loss_att=29.873, acc=0.802, loss=33.848, backward_time=0.230, grad_norm=31.579, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.439e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-29 20:00:08,317 (trainer:762) INFO: 17epoch:train:12501-13000batch: iter_time=1.464e-04, forward_time=0.209, loss_ctc=46.245, loss_att=31.869, acc=0.804, loss=36.182, backward_time=0.234, grad_norm=33.401, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.436e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-29 20:07:51,463 (trainer:762) INFO: 17epoch:train:13001-13500batch: iter_time=0.002, forward_time=0.209, loss_ctc=46.492, loss_att=31.844, acc=0.807, loss=36.238, backward_time=0.238, grad_norm=33.016, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.434e-04, train_time=0.926
+[ip-10-0-216-33:0/16] 2024-02-29 20:15:26,012 (trainer:762) INFO: 17epoch:train:13501-14000batch: iter_time=1.474e-04, forward_time=0.210, loss_ctc=45.072, loss_att=30.757, acc=0.807, loss=35.052, backward_time=0.226, grad_norm=31.618, clip=100.000, loss_scale=4.045e+33, optim_step_time=0.075, optim0_lr0=2.431e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-29 20:22:58,984 (trainer:762) INFO: 17epoch:train:14001-14500batch: iter_time=1.518e-04, forward_time=0.209, loss_ctc=44.353, loss_att=30.884, acc=0.806, loss=34.925, backward_time=0.235, grad_norm=32.295, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.429e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-29 20:26:25,221 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 20:30:34,484 (trainer:762) INFO: 17epoch:train:14501-15000batch: iter_time=1.480e-04, forward_time=0.210, loss_ctc=43.543, loss_att=29.897, acc=0.807, loss=33.991, backward_time=0.237, grad_norm=32.584, clip=100.000, loss_scale=3.777e+33, optim_step_time=0.075, optim0_lr0=2.427e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-29 20:45:00,576 (trainer:361) INFO: 17epoch results: [train] iter_time=0.087, forward_time=0.209, loss_ctc=44.712, loss_att=30.733, acc=0.805, loss=34.927, backward_time=0.235, grad_norm=32.511, clip=100.000, loss_scale=4.618e+33, optim_step_time=0.075, optim0_lr0=2.462e-04, train_time=0.998, time=4 hours, 9 minutes and 52.01 seconds, total_count=255000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=44.011, cer_ctc=0.240, loss_att=33.366, acc=0.721, cer=0.296, wer=0.998, loss=36.560, time=14 minutes and 10.68 seconds, total_count=36261, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-02-29 20:45:11,054 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-02-29 20:45:11,077 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/12epoch.pth
+[ip-10-0-216-33:0/16] 2024-02-29 20:45:11,078 (trainer:290) INFO: 18/45epoch started. Estimated time to finish: 5 days, 3 hours and 41 minutes
+[ip-10-0-216-33:0/16] 2024-02-29 20:45:11,085 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 20:45:47,196 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 20:45:55,045 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfc83280>)
+[ip-10-0-216-33:0/16] 2024-02-29 20:45:55,045 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-02-29 20:45:55,049 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 20:59:53,250 (trainer:762) INFO: 18epoch:train:1-500batch: iter_time=0.843, forward_time=0.210, loss_ctc=47.330, loss_att=32.290, acc=0.813, loss=36.802, backward_time=0.229, grad_norm=32.370, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.424e-04, train_time=1.764
+[ip-10-0-216-33:0/16] 2024-02-29 21:07:28,982 (trainer:762) INFO: 18epoch:train:501-1000batch: iter_time=1.498e-04, forward_time=0.210, loss_ctc=46.655, loss_att=32.317, acc=0.801, loss=36.619, backward_time=0.236, grad_norm=35.183, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.422e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-29 21:15:00,755 (trainer:762) INFO: 18epoch:train:1001-1500batch: iter_time=1.500e-04, forward_time=0.209, loss_ctc=43.989, loss_att=30.015, acc=0.804, loss=34.207, backward_time=0.228, grad_norm=31.923, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.420e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-29 21:22:34,052 (trainer:762) INFO: 18epoch:train:1501-2000batch: iter_time=1.489e-04, forward_time=0.210, loss_ctc=43.761, loss_att=29.997, acc=0.810, loss=34.126, backward_time=0.242, grad_norm=32.475, clip=100.000, loss_scale=4.008e+33, optim_step_time=0.075, optim0_lr0=2.417e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-29 21:30:04,954 (trainer:762) INFO: 18epoch:train:2001-2500batch: iter_time=1.471e-04, forward_time=0.210, loss_ctc=45.856, loss_att=31.442, acc=0.801, loss=35.766, backward_time=0.235, grad_norm=32.977, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.415e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-02-29 21:33:39,735 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 21:37:36,524 (trainer:762) INFO: 18epoch:train:2501-3000batch: iter_time=1.473e-04, forward_time=0.209, loss_ctc=45.251, loss_att=30.738, acc=0.803, loss=35.092, backward_time=0.233, grad_norm=33.480, clip=100.000, loss_scale=3.829e+33, optim_step_time=0.075, optim0_lr0=2.412e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-02-29 21:45:11,838 (trainer:762) INFO: 18epoch:train:3001-3500batch: iter_time=1.520e-04, forward_time=0.209, loss_ctc=42.940, loss_att=29.202, acc=0.810, loss=33.324, backward_time=0.239, grad_norm=31.638, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.410e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-02-29 21:52:46,331 (trainer:762) INFO: 18epoch:train:3501-4000batch: iter_time=1.504e-04, forward_time=0.210, loss_ctc=43.931, loss_att=29.574, acc=0.813, loss=33.881, backward_time=0.237, grad_norm=31.627, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.408e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-02-29 22:00:20,226 (trainer:762) INFO: 18epoch:train:4001-4500batch: iter_time=1.573e-04, forward_time=0.209, loss_ctc=45.080, loss_att=31.655, acc=0.801, loss=35.683, backward_time=0.236, grad_norm=34.147, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.405e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-02-29 22:07:03,476 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 22:07:55,961 (trainer:762) INFO: 18epoch:train:4501-5000batch: iter_time=1.457e-04, forward_time=0.209, loss_ctc=44.952, loss_att=30.811, acc=0.806, loss=35.053, backward_time=0.233, grad_norm=30.845, clip=100.000, loss_scale=3.652e+33, optim_step_time=0.075, optim0_lr0=2.403e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-02-29 22:08:00,736 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 22:08:37,629 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 22:08:45,603 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf347880>)
+[ip-10-0-216-33:0/16] 2024-02-29 22:08:45,603 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-02-29 22:08:45,607 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 22:23:12,430 (trainer:762) INFO: 18epoch:train:5001-5500batch: iter_time=0.895, forward_time=0.208, loss_ctc=46.883, loss_att=31.989, acc=0.814, loss=36.457, backward_time=0.233, grad_norm=38.814, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.401e-04, train_time=1.833
+[ip-10-0-216-33:0/16] 2024-02-29 22:30:45,701 (trainer:762) INFO: 18epoch:train:5501-6000batch: iter_time=1.472e-04, forward_time=0.209, loss_ctc=46.410, loss_att=32.141, acc=0.803, loss=36.422, backward_time=0.235, grad_norm=35.189, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.399e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-02-29 22:38:26,629 (trainer:762) INFO: 18epoch:train:6001-6500batch: iter_time=1.487e-04, forward_time=0.205, loss_ctc=43.634, loss_att=29.830, acc=0.805, loss=33.971, backward_time=0.235, grad_norm=30.855, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.396e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-02-29 22:46:05,623 (trainer:762) INFO: 18epoch:train:6501-7000batch: iter_time=1.494e-04, forward_time=0.207, loss_ctc=43.538, loss_att=29.796, acc=0.811, loss=33.919, backward_time=0.236, grad_norm=30.094, clip=100.000, loss_scale=2.897e+33, optim_step_time=0.075, optim0_lr0=2.394e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-02-29 22:49:35,811 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-02-29 22:53:35,414 (trainer:762) INFO: 18epoch:train:7001-7500batch: iter_time=1.448e-04, forward_time=0.208, loss_ctc=45.660, loss_att=31.290, acc=0.802, loss=35.601, backward_time=0.235, grad_norm=32.817, clip=100.000, loss_scale=3.808e+33, optim_step_time=0.075, optim0_lr0=2.392e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-02-29 23:01:08,127 (trainer:762) INFO: 18epoch:train:7501-8000batch: iter_time=1.490e-04, forward_time=0.209, loss_ctc=44.630, loss_att=30.403, acc=0.804, loss=34.671, backward_time=0.238, grad_norm=33.680, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.389e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-02-29 23:08:38,263 (trainer:762) INFO: 18epoch:train:8001-8500batch: iter_time=1.513e-04, forward_time=0.209, loss_ctc=42.376, loss_att=28.937, acc=0.811, loss=32.969, backward_time=0.236, grad_norm=31.290, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.387e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-02-29 23:16:09,443 (trainer:762) INFO: 18epoch:train:8501-9000batch: iter_time=1.453e-04, forward_time=0.209, loss_ctc=43.470, loss_att=29.291, acc=0.813, loss=33.545, backward_time=0.230, grad_norm=30.921, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.385e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-02-29 23:23:45,484 (trainer:762) INFO: 18epoch:train:9001-9500batch: iter_time=1.424e-04, forward_time=0.208, loss_ctc=44.365, loss_att=31.367, acc=0.803, loss=35.267, backward_time=0.231, grad_norm=34.324, clip=100.000, loss_scale=3.977e+33, optim_step_time=0.075, optim0_lr0=2.383e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-29 23:31:21,850 (trainer:762) INFO: 18epoch:train:9501-10000batch: iter_time=1.427e-04, forward_time=0.207, loss_ctc=44.449, loss_att=30.581, acc=0.807, loss=34.741, backward_time=0.234, grad_norm=30.266, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.380e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-02-29 23:31:26,963 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-02-29 23:32:04,772 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-02-29 23:32:12,757 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f03adc6d0>)
+[ip-10-0-216-33:0/16] 2024-02-29 23:32:12,757 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-02-29 23:32:12,762 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-02-29 23:46:40,245 (trainer:762) INFO: 18epoch:train:10001-10500batch: iter_time=0.875, forward_time=0.210, loss_ctc=46.471, loss_att=31.565, acc=0.816, loss=36.037, backward_time=0.239, grad_norm=32.170, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.378e-04, train_time=1.837
+[ip-10-0-216-33:0/16] 2024-02-29 23:54:18,540 (trainer:762) INFO: 18epoch:train:10501-11000batch: iter_time=1.478e-04, forward_time=0.209, loss_ctc=45.830, loss_att=31.864, acc=0.804, loss=36.054, backward_time=0.237, grad_norm=34.804, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.376e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-01 00:01:57,221 (trainer:762) INFO: 18epoch:train:11001-11500batch: iter_time=1.538e-04, forward_time=0.208, loss_ctc=43.288, loss_att=29.656, acc=0.806, loss=33.746, backward_time=0.231, grad_norm=31.101, clip=100.000, loss_scale=7.955e+33, optim_step_time=0.075, optim0_lr0=2.374e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-01 00:04:43,175 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 00:09:35,989 (trainer:762) INFO: 18epoch:train:11501-12000batch: iter_time=1.462e-04, forward_time=0.208, loss_ctc=43.130, loss_att=29.614, acc=0.812, loss=33.669, backward_time=0.233, grad_norm=30.614, clip=100.000, loss_scale=7.055e+33, optim_step_time=0.075, optim0_lr0=2.371e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-01 00:17:06,427 (trainer:762) INFO: 18epoch:train:12001-12500batch: iter_time=1.475e-04, forward_time=0.210, loss_ctc=45.291, loss_att=31.066, acc=0.803, loss=35.334, backward_time=0.233, grad_norm=33.412, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.369e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-01 00:24:37,867 (trainer:762) INFO: 18epoch:train:12501-13000batch: iter_time=1.498e-04, forward_time=0.210, loss_ctc=44.527, loss_att=30.416, acc=0.805, loss=34.649, backward_time=0.232, grad_norm=33.792, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.367e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-01 00:32:13,113 (trainer:762) INFO: 18epoch:train:13001-13500batch: iter_time=1.503e-04, forward_time=0.209, loss_ctc=42.184, loss_att=28.849, acc=0.812, loss=32.849, backward_time=0.231, grad_norm=33.365, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.365e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-01 00:35:23,133 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 00:35:34,701 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 00:39:45,189 (trainer:762) INFO: 18epoch:train:13501-14000batch: iter_time=1.507e-04, forward_time=0.208, loss_ctc=43.061, loss_att=29.144, acc=0.814, loss=33.319, backward_time=0.232, grad_norm=30.780, clip=100.000, loss_scale=4.019e+33, optim_step_time=0.075, optim0_lr0=2.363e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-01 00:47:16,528 (trainer:762) INFO: 18epoch:train:14001-14500batch: iter_time=1.496e-04, forward_time=0.210, loss_ctc=44.281, loss_att=31.251, acc=0.803, loss=35.160, backward_time=0.233, grad_norm=34.667, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.360e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-01 00:54:53,359 (trainer:762) INFO: 18epoch:train:14501-15000batch: iter_time=1.496e-04, forward_time=0.207, loss_ctc=44.169, loss_att=30.375, acc=0.808, loss=34.513, backward_time=0.236, grad_norm=31.594, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.358e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-01 01:08:57,988 (trainer:361) INFO: 18epoch results: [train] iter_time=0.087, forward_time=0.209, loss_ctc=44.580, loss_att=30.582, acc=0.807, loss=34.781, backward_time=0.234, grad_norm=32.708, clip=100.000, loss_scale=3.796e+33, optim_step_time=0.075, optim0_lr0=2.391e-04, train_time=0.999, time=4 hours, 9 minutes and 57.68 seconds, total_count=270000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=42.581, cer_ctc=0.236, loss_att=31.416, acc=0.742, cer=0.229, wer=0.993, loss=34.765, time=13 minutes and 48.91 seconds, total_count=38394, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-01 01:09:07,824 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-01 01:09:07,850 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/13epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-01 01:09:07,851 (trainer:290) INFO: 19/45epoch started. Estimated time to finish: 4 days, 23 hours and 14 minutes
+[ip-10-0-216-33:0/16] 2024-03-01 01:09:07,858 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 01:09:43,763 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 01:09:52,075 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfb5aa00>)
+[ip-10-0-216-33:0/16] 2024-03-01 01:09:52,075 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-01 01:09:52,080 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 01:23:47,886 (trainer:762) INFO: 19epoch:train:1-500batch: iter_time=0.842, forward_time=0.209, loss_ctc=45.516, loss_att=31.230, acc=0.804, loss=35.516, backward_time=0.227, grad_norm=34.222, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.356e-04, train_time=1.760
+[ip-10-0-216-33:0/16] 2024-03-01 01:31:22,351 (trainer:762) INFO: 19epoch:train:501-1000batch: iter_time=1.464e-04, forward_time=0.206, loss_ctc=41.128, loss_att=28.366, acc=0.804, loss=32.195, backward_time=0.229, grad_norm=32.023, clip=100.000, loss_scale=4.045e+33, optim_step_time=0.075, optim0_lr0=2.354e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 01:38:55,477 (trainer:762) INFO: 19epoch:train:1001-1500batch: iter_time=1.471e-04, forward_time=0.210, loss_ctc=47.559, loss_att=32.597, acc=0.806, loss=37.085, backward_time=0.236, grad_norm=34.930, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.352e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-01 01:40:23,792 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 01:46:30,260 (trainer:762) INFO: 19epoch:train:1501-2000batch: iter_time=1.432e-04, forward_time=0.210, loss_ctc=45.119, loss_att=30.828, acc=0.810, loss=35.116, backward_time=0.235, grad_norm=33.248, clip=100.000, loss_scale=3.090e+33, optim_step_time=0.075, optim0_lr0=2.350e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 01:54:04,625 (trainer:762) INFO: 19epoch:train:2001-2500batch: iter_time=1.472e-04, forward_time=0.209, loss_ctc=45.629, loss_att=31.403, acc=0.801, loss=35.671, backward_time=0.233, grad_norm=35.751, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.347e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 02:01:42,312 (trainer:762) INFO: 19epoch:train:2501-3000batch: iter_time=1.449e-04, forward_time=0.209, loss_ctc=45.530, loss_att=30.895, acc=0.812, loss=35.285, backward_time=0.230, grad_norm=32.773, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.345e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 02:09:16,643 (trainer:762) INFO: 19epoch:train:3001-3500batch: iter_time=1.450e-04, forward_time=0.209, loss_ctc=44.924, loss_att=30.688, acc=0.805, loss=34.959, backward_time=0.230, grad_norm=33.373, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.343e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 02:16:47,546 (trainer:762) INFO: 19epoch:train:3501-4000batch: iter_time=1.490e-04, forward_time=0.208, loss_ctc=44.148, loss_att=30.167, acc=0.814, loss=34.361, backward_time=0.237, grad_norm=32.178, clip=100.000, loss_scale=4.694e+33, optim_step_time=0.075, optim0_lr0=2.341e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-01 02:24:25,604 (trainer:762) INFO: 19epoch:train:4001-4500batch: iter_time=1.466e-04, forward_time=0.208, loss_ctc=44.755, loss_att=30.444, acc=0.809, loss=34.738, backward_time=0.234, grad_norm=31.797, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.339e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-01 02:26:19,873 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 02:32:06,375 (trainer:762) INFO: 19epoch:train:4501-5000batch: iter_time=1.476e-04, forward_time=0.207, loss_ctc=43.818, loss_att=30.469, acc=0.812, loss=34.474, backward_time=0.234, grad_norm=33.045, clip=100.000, loss_scale=3.236e+33, optim_step_time=0.075, optim0_lr0=2.337e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-01 02:32:11,031 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 02:32:47,607 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 02:32:56,335 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf8fcdc0>)
+[ip-10-0-216-33:0/16] 2024-03-01 02:32:56,335 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-01 02:32:56,340 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 02:47:11,995 (trainer:762) INFO: 19epoch:train:5001-5500batch: iter_time=0.882, forward_time=0.207, loss_ctc=44.935, loss_att=30.840, acc=0.805, loss=35.068, backward_time=0.233, grad_norm=33.424, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.335e-04, train_time=1.811
+[ip-10-0-216-33:0/16] 2024-03-01 02:54:49,564 (trainer:762) INFO: 19epoch:train:5501-6000batch: iter_time=1.460e-04, forward_time=0.208, loss_ctc=41.009, loss_att=28.261, acc=0.805, loss=32.086, backward_time=0.233, grad_norm=32.371, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.332e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 03:02:20,378 (trainer:762) INFO: 19epoch:train:6001-6500batch: iter_time=1.452e-04, forward_time=0.210, loss_ctc=47.030, loss_att=32.305, acc=0.808, loss=36.723, backward_time=0.237, grad_norm=35.404, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.330e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-01 03:09:57,591 (trainer:762) INFO: 19epoch:train:6501-7000batch: iter_time=1.411e-04, forward_time=0.207, loss_ctc=44.571, loss_att=30.493, acc=0.812, loss=34.716, backward_time=0.233, grad_norm=33.316, clip=100.000, loss_scale=4.548e+33, optim_step_time=0.075, optim0_lr0=2.328e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-01 03:17:34,556 (trainer:762) INFO: 19epoch:train:7001-7500batch: iter_time=1.407e-04, forward_time=0.207, loss_ctc=45.102, loss_att=31.048, acc=0.803, loss=35.265, backward_time=0.239, grad_norm=34.342, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.326e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-01 03:20:50,921 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 03:25:12,113 (trainer:762) INFO: 19epoch:train:7501-8000batch: iter_time=1.402e-04, forward_time=0.207, loss_ctc=45.073, loss_att=30.689, acc=0.814, loss=35.004, backward_time=0.235, grad_norm=33.272, clip=100.000, loss_scale=3.694e+33, optim_step_time=0.075, optim0_lr0=2.324e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 03:32:49,444 (trainer:762) INFO: 19epoch:train:8001-8500batch: iter_time=1.460e-04, forward_time=0.207, loss_ctc=44.638, loss_att=30.382, acc=0.807, loss=34.659, backward_time=0.231, grad_norm=34.361, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.322e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-01 03:40:24,069 (trainer:762) INFO: 19epoch:train:8501-9000batch: iter_time=1.428e-04, forward_time=0.207, loss_ctc=43.750, loss_att=29.870, acc=0.815, loss=34.034, backward_time=0.234, grad_norm=33.019, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.320e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 03:48:01,315 (trainer:762) INFO: 19epoch:train:9001-9500batch: iter_time=1.433e-04, forward_time=0.209, loss_ctc=44.477, loss_att=30.292, acc=0.811, loss=34.547, backward_time=0.229, grad_norm=32.102, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.318e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-01 03:51:21,183 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 03:55:33,770 (trainer:762) INFO: 19epoch:train:9501-10000batch: iter_time=1.448e-04, forward_time=0.207, loss_ctc=43.573, loss_att=30.305, acc=0.813, loss=34.285, backward_time=0.234, grad_norm=31.924, clip=100.000, loss_scale=2.627e+33, optim_step_time=0.075, optim0_lr0=2.316e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-01 03:55:38,835 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 03:56:15,813 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 03:56:24,417 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f09347280>)
+[ip-10-0-216-33:0/16] 2024-03-01 03:56:24,417 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-01 03:56:24,422 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 04:10:49,143 (trainer:762) INFO: 19epoch:train:10001-10500batch: iter_time=0.905, forward_time=0.208, loss_ctc=44.730, loss_att=30.763, acc=0.807, loss=34.953, backward_time=0.232, grad_norm=33.258, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.314e-04, train_time=1.831
+[ip-10-0-216-33:0/16] 2024-03-01 04:18:19,204 (trainer:762) INFO: 19epoch:train:10501-11000batch: iter_time=1.492e-04, forward_time=0.210, loss_ctc=40.639, loss_att=28.032, acc=0.807, loss=31.814, backward_time=0.232, grad_norm=31.855, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.312e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-01 04:25:53,125 (trainer:762) INFO: 19epoch:train:11001-11500batch: iter_time=1.487e-04, forward_time=0.208, loss_ctc=46.971, loss_att=32.277, acc=0.808, loss=36.685, backward_time=0.230, grad_norm=35.247, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.310e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 04:33:31,572 (trainer:762) INFO: 19epoch:train:11501-12000batch: iter_time=1.567e-04, forward_time=0.208, loss_ctc=44.411, loss_att=30.300, acc=0.813, loss=34.533, backward_time=0.241, grad_norm=37.587, clip=100.000, loss_scale=4.055e+33, optim_step_time=0.075, optim0_lr0=2.307e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-01 04:41:06,242 (trainer:762) INFO: 19epoch:train:12001-12500batch: iter_time=1.536e-04, forward_time=0.209, loss_ctc=44.763, loss_att=30.920, acc=0.804, loss=35.073, backward_time=0.236, grad_norm=34.622, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.305e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 04:42:57,320 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 04:48:37,451 (trainer:762) INFO: 19epoch:train:12501-13000batch: iter_time=1.520e-04, forward_time=0.210, loss_ctc=44.891, loss_att=30.617, acc=0.814, loss=34.899, backward_time=0.228, grad_norm=67.651, clip=100.000, loss_scale=3.231e+33, optim_step_time=0.075, optim0_lr0=2.303e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-01 04:56:21,407 (trainer:762) INFO: 19epoch:train:13001-13500batch: iter_time=1.511e-04, forward_time=0.205, loss_ctc=44.232, loss_att=30.187, acc=0.808, loss=34.400, backward_time=0.234, grad_norm=33.744, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.301e-04, train_time=0.928
+[ip-10-0-216-33:0/16] 2024-03-01 05:04:02,640 (trainer:762) INFO: 19epoch:train:13501-14000batch: iter_time=1.510e-04, forward_time=0.209, loss_ctc=43.423, loss_att=29.780, acc=0.816, loss=33.873, backward_time=0.233, grad_norm=32.423, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.299e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-01 05:11:43,749 (trainer:762) INFO: 19epoch:train:14001-14500batch: iter_time=1.478e-04, forward_time=0.207, loss_ctc=44.195, loss_att=30.163, acc=0.811, loss=34.373, backward_time=0.230, grad_norm=31.813, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.297e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-01 05:16:13,277 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 05:19:23,989 (trainer:762) INFO: 19epoch:train:14501-15000batch: iter_time=1.444e-04, forward_time=0.209, loss_ctc=43.482, loss_att=30.229, acc=0.814, loss=34.205, backward_time=0.230, grad_norm=31.818, clip=100.000, loss_scale=3.481e+33, optim_step_time=0.075, optim0_lr0=2.295e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-01 05:33:24,715 (trainer:361) INFO: 19epoch results: [train] iter_time=0.088, forward_time=0.208, loss_ctc=44.467, loss_att=30.495, acc=0.809, loss=34.686, backward_time=0.233, grad_norm=34.561, clip=100.000, loss_scale=3.300e+33, optim_step_time=0.075, optim0_lr0=2.325e-04, train_time=1.001, time=4 hours, 10 minutes and 30.84 seconds, total_count=285000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=42.117, cer_ctc=0.231, loss_att=30.659, acc=0.752, cer=0.216, wer=0.993, loss=34.096, time=13 minutes and 45.74 seconds, total_count=40527, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-01 05:33:34,377 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-01 05:33:34,403 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/14epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-01 05:33:34,403 (trainer:290) INFO: 20/45epoch started. Estimated time to finish: 4 days, 18 hours and 49 minutes
+[ip-10-0-216-33:0/16] 2024-03-01 05:33:34,411 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 05:34:10,647 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 05:34:18,490 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed03d66a0>)
+[ip-10-0-216-33:0/16] 2024-03-01 05:34:18,490 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-01 05:34:18,495 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 05:48:17,526 (trainer:762) INFO: 20epoch:train:1-500batch: iter_time=0.842, forward_time=0.210, loss_ctc=43.136, loss_att=29.619, acc=0.815, loss=33.674, backward_time=0.229, grad_norm=47.671, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.293e-04, train_time=1.766
+[ip-10-0-216-33:0/16] 2024-03-01 05:55:55,272 (trainer:762) INFO: 20epoch:train:501-1000batch: iter_time=1.499e-04, forward_time=0.209, loss_ctc=46.054, loss_att=31.667, acc=0.803, loss=35.983, backward_time=0.234, grad_norm=37.233, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.291e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 06:03:30,938 (trainer:762) INFO: 20epoch:train:1001-1500batch: iter_time=1.510e-04, forward_time=0.208, loss_ctc=41.999, loss_att=28.686, acc=0.813, loss=32.679, backward_time=0.232, grad_norm=32.327, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.289e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-01 06:11:05,643 (trainer:762) INFO: 20epoch:train:1501-2000batch: iter_time=1.475e-04, forward_time=0.211, loss_ctc=45.496, loss_att=31.746, acc=0.813, loss=35.871, backward_time=0.236, grad_norm=34.218, clip=100.000, loss_scale=3.666e+33, optim_step_time=0.075, optim0_lr0=2.287e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 06:18:44,244 (trainer:762) INFO: 20epoch:train:2001-2500batch: iter_time=1.474e-04, forward_time=0.208, loss_ctc=44.122, loss_att=30.763, acc=0.803, loss=34.771, backward_time=0.232, grad_norm=35.002, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.285e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-01 06:26:23,149 (trainer:762) INFO: 20epoch:train:2501-3000batch: iter_time=1.468e-04, forward_time=0.208, loss_ctc=45.401, loss_att=31.259, acc=0.809, loss=35.502, backward_time=0.236, grad_norm=33.877, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.283e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-01 06:33:58,045 (trainer:762) INFO: 20epoch:train:3001-3500batch: iter_time=1.475e-04, forward_time=0.209, loss_ctc=41.596, loss_att=28.716, acc=0.810, loss=32.580, backward_time=0.233, grad_norm=32.857, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.281e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-01 06:38:48,294 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 06:41:30,376 (trainer:762) INFO: 20epoch:train:3501-4000batch: iter_time=1.495e-04, forward_time=0.209, loss_ctc=44.551, loss_att=31.107, acc=0.813, loss=35.140, backward_time=0.238, grad_norm=34.681, clip=100.000, loss_scale=5.452e+33, optim_step_time=0.075, optim0_lr0=2.279e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-01 06:49:09,535 (trainer:762) INFO: 20epoch:train:4001-4500batch: iter_time=1.459e-04, forward_time=0.210, loss_ctc=47.142, loss_att=32.305, acc=0.812, loss=36.756, backward_time=0.231, grad_norm=34.142, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.277e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-01 06:56:47,057 (trainer:762) INFO: 20epoch:train:4501-5000batch: iter_time=1.469e-04, forward_time=0.209, loss_ctc=45.155, loss_att=30.767, acc=0.812, loss=35.083, backward_time=0.234, grad_norm=34.070, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.275e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 06:56:53,045 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 06:57:30,657 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 06:57:38,619 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed6056be0>)
+[ip-10-0-216-33:0/16] 2024-03-01 06:57:38,620 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-01 06:57:38,624 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 07:12:22,412 (trainer:762) INFO: 20epoch:train:5001-5500batch: iter_time=0.905, forward_time=0.210, loss_ctc=42.843, loss_att=29.381, acc=0.816, loss=33.420, backward_time=0.237, grad_norm=33.705, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.273e-04, train_time=1.870
+[ip-10-0-216-33:0/16] 2024-03-01 07:18:48,221 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 07:19:56,410 (trainer:762) INFO: 20epoch:train:5501-6000batch: iter_time=1.569e-04, forward_time=0.209, loss_ctc=45.671, loss_att=31.476, acc=0.804, loss=35.735, backward_time=0.237, grad_norm=36.264, clip=100.000, loss_scale=6.274e+33, optim_step_time=0.075, optim0_lr0=2.271e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 07:27:28,906 (trainer:762) INFO: 20epoch:train:6001-6500batch: iter_time=1.502e-04, forward_time=0.209, loss_ctc=41.765, loss_att=28.529, acc=0.814, loss=32.500, backward_time=0.241, grad_norm=33.145, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.270e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-01 07:35:09,364 (trainer:762) INFO: 20epoch:train:6501-7000batch: iter_time=1.505e-04, forward_time=0.209, loss_ctc=45.150, loss_att=31.472, acc=0.814, loss=35.575, backward_time=0.228, grad_norm=33.955, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.268e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-01 07:42:45,615 (trainer:762) INFO: 20epoch:train:7001-7500batch: iter_time=1.442e-04, forward_time=0.209, loss_ctc=43.691, loss_att=30.489, acc=0.804, loss=34.450, backward_time=0.233, grad_norm=41.300, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.266e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-01 07:47:39,422 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 07:50:21,581 (trainer:762) INFO: 20epoch:train:7501-8000batch: iter_time=1.486e-04, forward_time=0.209, loss_ctc=45.045, loss_att=31.053, acc=0.810, loss=35.251, backward_time=0.231, grad_norm=35.404, clip=100.000, loss_scale=4.261e+33, optim_step_time=0.075, optim0_lr0=2.264e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-01 07:57:56,309 (trainer:762) INFO: 20epoch:train:8001-8500batch: iter_time=1.497e-04, forward_time=0.210, loss_ctc=41.123, loss_att=28.471, acc=0.811, loss=32.266, backward_time=0.237, grad_norm=32.847, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.262e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 08:05:33,841 (trainer:762) INFO: 20epoch:train:8501-9000batch: iter_time=1.457e-04, forward_time=0.210, loss_ctc=44.002, loss_att=30.813, acc=0.815, loss=34.769, backward_time=0.234, grad_norm=35.165, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.260e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 08:13:13,104 (trainer:762) INFO: 20epoch:train:9001-9500batch: iter_time=1.493e-04, forward_time=0.209, loss_ctc=46.841, loss_att=32.142, acc=0.813, loss=36.552, backward_time=0.235, grad_norm=33.449, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.258e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-01 08:20:47,927 (trainer:762) INFO: 20epoch:train:9501-10000batch: iter_time=1.437e-04, forward_time=0.210, loss_ctc=44.438, loss_att=30.458, acc=0.814, loss=34.652, backward_time=0.234, grad_norm=32.589, clip=100.000, loss_scale=3.526e+33, optim_step_time=0.075, optim0_lr0=2.256e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 08:20:53,473 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 08:21:32,705 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 08:21:41,330 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f0857edf0>)
+[ip-10-0-216-33:0/16] 2024-03-01 08:21:41,330 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-01 08:21:41,335 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 08:29:56,357 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 08:35:58,946 (trainer:762) INFO: 20epoch:train:10001-10500batch: iter_time=0.853, forward_time=0.210, loss_ctc=42.489, loss_att=29.146, acc=0.816, loss=33.149, backward_time=0.235, grad_norm=32.878, clip=100.000, loss_scale=3.106e+33, optim_step_time=0.075, optim0_lr0=2.254e-04, train_time=1.822
+[ip-10-0-216-33:0/16] 2024-03-01 08:43:32,344 (trainer:762) INFO: 20epoch:train:10501-11000batch: iter_time=1.473e-04, forward_time=0.210, loss_ctc=45.596, loss_att=31.414, acc=0.805, loss=35.669, backward_time=0.235, grad_norm=35.892, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.252e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-01 08:51:02,059 (trainer:762) INFO: 20epoch:train:11001-11500batch: iter_time=1.497e-04, forward_time=0.211, loss_ctc=41.460, loss_att=28.263, acc=0.816, loss=32.222, backward_time=0.239, grad_norm=32.967, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.250e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-01 08:55:00,304 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 08:58:36,433 (trainer:762) INFO: 20epoch:train:11501-12000batch: iter_time=1.480e-04, forward_time=0.211, loss_ctc=44.589, loss_att=31.173, acc=0.816, loss=35.198, backward_time=0.233, grad_norm=34.535, clip=100.000, loss_scale=1.974e+33, optim_step_time=0.075, optim0_lr0=2.248e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 09:06:06,913 (trainer:762) INFO: 20epoch:train:12001-12500batch: iter_time=1.512e-04, forward_time=0.211, loss_ctc=43.329, loss_att=30.401, acc=0.805, loss=34.279, backward_time=0.234, grad_norm=34.105, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.247e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-01 09:13:38,967 (trainer:762) INFO: 20epoch:train:12501-13000batch: iter_time=1.506e-04, forward_time=0.210, loss_ctc=44.759, loss_att=30.949, acc=0.810, loss=35.092, backward_time=0.232, grad_norm=33.217, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.245e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-01 09:21:17,320 (trainer:762) INFO: 20epoch:train:13001-13500batch: iter_time=1.501e-04, forward_time=0.210, loss_ctc=40.985, loss_att=28.365, acc=0.812, loss=32.151, backward_time=0.236, grad_norm=32.906, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.243e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-01 09:28:54,779 (trainer:762) INFO: 20epoch:train:13501-14000batch: iter_time=1.467e-04, forward_time=0.210, loss_ctc=43.672, loss_att=30.676, acc=0.815, loss=34.575, backward_time=0.235, grad_norm=34.559, clip=100.000, loss_scale=1.919e+33, optim_step_time=0.075, optim0_lr0=2.241e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 09:36:26,460 (trainer:762) INFO: 20epoch:train:14001-14500batch: iter_time=1.483e-04, forward_time=0.210, loss_ctc=46.543, loss_att=31.893, acc=0.814, loss=36.288, backward_time=0.234, grad_norm=33.485, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.239e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-01 09:44:00,696 (trainer:762) INFO: 20epoch:train:14501-15000batch: iter_time=1.483e-04, forward_time=0.211, loss_ctc=44.275, loss_att=30.353, acc=0.814, loss=34.529, backward_time=0.236, grad_norm=33.140, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.237e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 09:58:11,012 (trainer:361) INFO: 20epoch results: [train] iter_time=0.087, forward_time=0.210, loss_ctc=44.097, loss_att=30.452, acc=0.811, loss=34.545, backward_time=0.234, grad_norm=34.719, clip=100.000, loss_scale=3.559e+33, optim_step_time=0.075, optim0_lr0=2.265e-04, train_time=1.002, time=4 hours, 10 minutes and 41.83 seconds, total_count=300000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=41.410, cer_ctc=0.230, loss_att=32.884, acc=0.719, cer=0.306, wer=1.000, loss=35.442, time=13 minutes and 54.47 seconds, total_count=42660, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-01 09:58:20,918 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-01 09:58:20,942 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/15epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-01 09:58:20,943 (trainer:290) INFO: 21/45epoch started. Estimated time to finish: 4 days, 14 hours and 23 minutes
+[ip-10-0-216-33:0/16] 2024-03-01 09:58:20,950 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 09:58:56,977 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 09:59:04,942 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5eceba1d60>)
+[ip-10-0-216-33:0/16] 2024-03-01 09:59:04,942 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-01 09:59:04,947 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 10:13:06,397 (trainer:762) INFO: 21epoch:train:1-500batch: iter_time=0.852, forward_time=0.208, loss_ctc=43.374, loss_att=29.769, acc=0.808, loss=33.850, backward_time=0.231, grad_norm=34.850, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.235e-04, train_time=1.771
+[ip-10-0-216-33:0/16] 2024-03-01 10:20:40,991 (trainer:762) INFO: 21epoch:train:501-1000batch: iter_time=1.468e-04, forward_time=0.209, loss_ctc=44.792, loss_att=30.507, acc=0.812, loss=34.793, backward_time=0.232, grad_norm=33.114, clip=100.000, loss_scale=3.837e+33, optim_step_time=0.075, optim0_lr0=2.233e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 10:28:15,060 (trainer:762) INFO: 21epoch:train:1001-1500batch: iter_time=1.505e-04, forward_time=0.208, loss_ctc=44.538, loss_att=30.690, acc=0.807, loss=34.844, backward_time=0.233, grad_norm=37.574, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.232e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 10:33:57,055 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 10:35:47,931 (trainer:762) INFO: 21epoch:train:1501-2000batch: iter_time=1.484e-04, forward_time=0.209, loss_ctc=42.962, loss_att=29.423, acc=0.810, loss=33.485, backward_time=0.235, grad_norm=35.017, clip=100.000, loss_scale=4.547e+33, optim_step_time=0.075, optim0_lr0=2.230e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-01 10:43:25,255 (trainer:762) INFO: 21epoch:train:2001-2500batch: iter_time=1.476e-04, forward_time=0.210, loss_ctc=45.490, loss_att=31.792, acc=0.813, loss=35.901, backward_time=0.232, grad_norm=35.118, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.228e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-01 10:51:01,788 (trainer:762) INFO: 21epoch:train:2501-3000batch: iter_time=1.495e-04, forward_time=0.207, loss_ctc=41.313, loss_att=28.305, acc=0.811, loss=32.207, backward_time=0.233, grad_norm=33.014, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.226e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-01 10:58:39,050 (trainer:762) INFO: 21epoch:train:3001-3500batch: iter_time=1.493e-04, forward_time=0.209, loss_ctc=42.651, loss_att=29.777, acc=0.812, loss=33.639, backward_time=0.235, grad_norm=34.566, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.224e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-01 11:06:19,663 (trainer:762) INFO: 21epoch:train:3501-4000batch: iter_time=1.497e-04, forward_time=0.208, loss_ctc=41.539, loss_att=28.515, acc=0.813, loss=32.422, backward_time=0.240, grad_norm=33.685, clip=100.000, loss_scale=3.240e+33, optim_step_time=0.075, optim0_lr0=2.222e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-01 11:14:00,053 (trainer:762) INFO: 21epoch:train:4001-4500batch: iter_time=1.511e-04, forward_time=0.208, loss_ctc=42.932, loss_att=29.791, acc=0.814, loss=33.734, backward_time=0.236, grad_norm=34.703, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.221e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-01 11:20:17,641 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 11:21:40,272 (trainer:762) INFO: 21epoch:train:4501-5000batch: iter_time=1.461e-04, forward_time=0.208, loss_ctc=47.452, loss_att=32.643, acc=0.806, loss=37.086, backward_time=0.234, grad_norm=36.686, clip=100.000, loss_scale=4.729e+33, optim_step_time=0.075, optim0_lr0=2.219e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-01 11:21:44,814 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 11:22:21,444 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 11:22:29,490 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf1c8370>)
+[ip-10-0-216-33:0/16] 2024-03-01 11:22:29,490 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-01 11:22:29,495 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 11:36:48,003 (trainer:762) INFO: 21epoch:train:5001-5500batch: iter_time=0.900, forward_time=0.210, loss_ctc=42.944, loss_att=29.513, acc=0.810, loss=33.542, backward_time=0.237, grad_norm=34.851, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.217e-04, train_time=1.815
+[ip-10-0-216-33:0/16] 2024-03-01 11:44:20,415 (trainer:762) INFO: 21epoch:train:5501-6000batch: iter_time=1.689e-04, forward_time=0.209, loss_ctc=44.215, loss_att=30.142, acc=0.815, loss=34.364, backward_time=0.235, grad_norm=32.395, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.215e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-01 11:51:53,692 (trainer:762) INFO: 21epoch:train:6001-6500batch: iter_time=1.709e-04, forward_time=0.208, loss_ctc=44.024, loss_att=30.397, acc=0.808, loss=34.485, backward_time=0.229, grad_norm=35.663, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.213e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-01 11:59:32,900 (trainer:762) INFO: 21epoch:train:6501-7000batch: iter_time=1.686e-04, forward_time=0.208, loss_ctc=42.760, loss_att=29.361, acc=0.810, loss=33.381, backward_time=0.231, grad_norm=35.378, clip=100.000, loss_scale=3.058e+33, optim_step_time=0.075, optim0_lr0=2.211e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-01 12:07:09,380 (trainer:762) INFO: 21epoch:train:7001-7500batch: iter_time=1.648e-04, forward_time=0.208, loss_ctc=45.031, loss_att=31.541, acc=0.814, loss=35.588, backward_time=0.232, grad_norm=35.932, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.210e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-01 12:14:42,262 (trainer:762) INFO: 21epoch:train:7501-8000batch: iter_time=1.535e-04, forward_time=0.209, loss_ctc=41.122, loss_att=28.104, acc=0.812, loss=32.009, backward_time=0.231, grad_norm=33.456, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.208e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-01 12:15:59,722 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 12:19:21,043 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 12:22:18,653 (trainer:762) INFO: 21epoch:train:8001-8500batch: iter_time=1.590e-04, forward_time=0.207, loss_ctc=42.036, loss_att=29.544, acc=0.813, loss=33.292, backward_time=0.232, grad_norm=37.626, clip=100.000, loss_scale=2.526e+33, optim_step_time=0.075, optim0_lr0=2.206e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-01 12:29:52,488 (trainer:762) INFO: 21epoch:train:8501-9000batch: iter_time=1.536e-04, forward_time=0.208, loss_ctc=41.196, loss_att=28.304, acc=0.814, loss=32.171, backward_time=0.234, grad_norm=34.740, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.204e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-01 12:37:28,374 (trainer:762) INFO: 21epoch:train:9001-9500batch: iter_time=1.606e-04, forward_time=0.208, loss_ctc=42.819, loss_att=29.846, acc=0.814, loss=33.738, backward_time=0.237, grad_norm=34.222, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.203e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-01 12:45:03,095 (trainer:762) INFO: 21epoch:train:9501-10000batch: iter_time=1.522e-04, forward_time=0.208, loss_ctc=47.111, loss_att=32.457, acc=0.806, loss=36.853, backward_time=0.236, grad_norm=36.486, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=2.201e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 12:45:08,361 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 12:45:45,990 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 12:45:53,915 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5edcdfd220>)
+[ip-10-0-216-33:0/16] 2024-03-01 12:45:53,916 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-01 12:45:53,920 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 13:00:23,429 (trainer:762) INFO: 21epoch:train:10001-10500batch: iter_time=0.884, forward_time=0.208, loss_ctc=42.686, loss_att=29.332, acc=0.811, loss=33.338, backward_time=0.237, grad_norm=34.913, clip=100.000, loss_scale=1.804e+33, optim_step_time=0.075, optim0_lr0=2.199e-04, train_time=1.840
+[ip-10-0-216-33:0/16] 2024-03-01 13:08:03,259 (trainer:762) INFO: 21epoch:train:10501-11000batch: iter_time=1.450e-04, forward_time=0.208, loss_ctc=43.923, loss_att=30.056, acc=0.815, loss=34.216, backward_time=0.235, grad_norm=33.003, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.197e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-01 13:15:36,398 (trainer:762) INFO: 21epoch:train:11001-11500batch: iter_time=1.499e-04, forward_time=0.209, loss_ctc=43.854, loss_att=30.370, acc=0.809, loss=34.415, backward_time=0.229, grad_norm=36.607, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.195e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-01 13:23:11,048 (trainer:762) INFO: 21epoch:train:11501-12000batch: iter_time=1.477e-04, forward_time=0.208, loss_ctc=42.297, loss_att=29.128, acc=0.812, loss=33.079, backward_time=0.241, grad_norm=34.725, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.194e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 13:30:45,072 (trainer:762) INFO: 21epoch:train:12001-12500batch: iter_time=1.494e-04, forward_time=0.210, loss_ctc=44.558, loss_att=31.237, acc=0.815, loss=35.233, backward_time=0.232, grad_norm=35.211, clip=100.000, loss_scale=3.609e+33, optim_step_time=0.075, optim0_lr0=2.192e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 13:38:19,315 (trainer:762) INFO: 21epoch:train:12501-13000batch: iter_time=1.457e-04, forward_time=0.208, loss_ctc=40.880, loss_att=27.994, acc=0.813, loss=31.860, backward_time=0.234, grad_norm=32.758, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.190e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 13:45:58,014 (trainer:762) INFO: 21epoch:train:13001-13500batch: iter_time=1.530e-04, forward_time=0.210, loss_ctc=41.936, loss_att=29.436, acc=0.813, loss=33.186, backward_time=0.235, grad_norm=35.512, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.188e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-01 13:53:33,177 (trainer:762) INFO: 21epoch:train:13501-14000batch: iter_time=1.533e-04, forward_time=0.208, loss_ctc=40.829, loss_att=28.129, acc=0.816, loss=31.939, backward_time=0.236, grad_norm=33.601, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.187e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-01 13:59:50,859 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 14:01:08,752 (trainer:762) INFO: 21epoch:train:14001-14500batch: iter_time=1.460e-04, forward_time=0.211, loss_ctc=42.510, loss_att=29.578, acc=0.815, loss=33.458, backward_time=0.237, grad_norm=34.850, clip=100.000, loss_scale=6.306e+33, optim_step_time=0.075, optim0_lr0=2.185e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-01 14:08:41,344 (trainer:762) INFO: 21epoch:train:14501-15000batch: iter_time=1.459e-04, forward_time=0.208, loss_ctc=46.613, loss_att=32.154, acc=0.807, loss=36.492, backward_time=0.231, grad_norm=35.899, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.183e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-01 14:23:02,570 (trainer:361) INFO: 21epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=43.346, loss_att=29.928, acc=0.812, loss=33.953, backward_time=0.234, grad_norm=34.871, clip=100.000, loss_scale=3.501e+33, optim_step_time=0.075, optim0_lr0=2.209e-04, train_time=1.001, time=4 hours, 10 minutes and 35.41 seconds, total_count=315000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=40.808, cer_ctc=0.223, loss_att=31.526, acc=0.731, cer=0.295, wer=0.997, loss=34.311, time=14 minutes and 5.9 seconds, total_count=44793, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-01 14:23:12,310 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-01 14:23:12,328 (trainer:290) INFO: 22/45epoch started. Estimated time to finish: 4 days, 9 hours and 58 minutes
+[ip-10-0-216-33:0/16] 2024-03-01 14:23:12,335 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 14:23:48,213 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 14:23:56,587 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf4e4970>)
+[ip-10-0-216-33:0/16] 2024-03-01 14:23:56,587 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-01 14:23:56,592 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 14:34:12,124 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 14:37:50,143 (trainer:762) INFO: 22epoch:train:1-500batch: iter_time=0.838, forward_time=0.209, loss_ctc=45.001, loss_att=30.605, acc=0.809, loss=34.924, backward_time=0.235, grad_norm=35.681, clip=100.000, loss_scale=3.959e+33, optim_step_time=0.075, optim0_lr0=2.181e-04, train_time=1.755
+[ip-10-0-216-33:0/16] 2024-03-01 14:45:24,982 (trainer:762) INFO: 22epoch:train:501-1000batch: iter_time=1.462e-04, forward_time=0.209, loss_ctc=43.417, loss_att=29.453, acc=0.817, loss=33.642, backward_time=0.236, grad_norm=34.314, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.180e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 14:53:04,706 (trainer:762) INFO: 22epoch:train:1001-1500batch: iter_time=1.468e-04, forward_time=0.209, loss_ctc=44.080, loss_att=30.687, acc=0.809, loss=34.705, backward_time=0.234, grad_norm=36.154, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.178e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-01 15:00:41,013 (trainer:762) INFO: 22epoch:train:1501-2000batch: iter_time=1.516e-04, forward_time=0.209, loss_ctc=45.417, loss_att=31.232, acc=0.807, loss=35.487, backward_time=0.240, grad_norm=35.705, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.176e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-01 15:08:14,877 (trainer:762) INFO: 22epoch:train:2001-2500batch: iter_time=1.486e-04, forward_time=0.208, loss_ctc=42.517, loss_att=28.884, acc=0.808, loss=32.974, backward_time=0.230, grad_norm=34.454, clip=100.000, loss_scale=3.827e+33, optim_step_time=0.075, optim0_lr0=2.175e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-01 15:15:50,322 (trainer:762) INFO: 22epoch:train:2501-3000batch: iter_time=1.427e-04, forward_time=0.209, loss_ctc=42.102, loss_att=28.834, acc=0.820, loss=32.814, backward_time=0.232, grad_norm=34.805, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.173e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-01 15:23:25,578 (trainer:762) INFO: 22epoch:train:3001-3500batch: iter_time=1.445e-04, forward_time=0.209, loss_ctc=42.410, loss_att=29.268, acc=0.811, loss=33.211, backward_time=0.237, grad_norm=34.025, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.171e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-01 15:30:56,034 (trainer:762) INFO: 22epoch:train:3501-4000batch: iter_time=1.471e-04, forward_time=0.209, loss_ctc=45.141, loss_att=31.336, acc=0.807, loss=35.478, backward_time=0.236, grad_norm=36.174, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.169e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-01 15:37:05,663 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 15:38:34,667 (trainer:762) INFO: 22epoch:train:4001-4500batch: iter_time=1.474e-04, forward_time=0.210, loss_ctc=43.619, loss_att=30.381, acc=0.811, loss=34.352, backward_time=0.234, grad_norm=36.427, clip=100.000, loss_scale=6.628e+33, optim_step_time=0.075, optim0_lr0=2.168e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-01 15:46:08,665 (trainer:762) INFO: 22epoch:train:4501-5000batch: iter_time=1.492e-04, forward_time=0.209, loss_ctc=42.232, loss_att=28.798, acc=0.813, loss=32.828, backward_time=0.232, grad_norm=33.771, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.166e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 15:46:13,147 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 15:46:49,934 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 15:46:58,744 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf881a90>)
+[ip-10-0-216-33:0/16] 2024-03-01 15:46:58,745 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-01 15:46:58,749 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 16:00:21,522 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 16:01:24,768 (trainer:762) INFO: 22epoch:train:5001-5500batch: iter_time=0.894, forward_time=0.208, loss_ctc=44.607, loss_att=30.292, acc=0.810, loss=34.587, backward_time=0.228, grad_norm=35.440, clip=100.000, loss_scale=4.828e+33, optim_step_time=0.075, optim0_lr0=2.164e-04, train_time=1.832
+[ip-10-0-216-33:0/16] 2024-03-01 16:09:00,335 (trainer:762) INFO: 22epoch:train:5501-6000batch: iter_time=1.452e-04, forward_time=0.209, loss_ctc=42.978, loss_att=29.192, acc=0.818, loss=33.328, backward_time=0.235, grad_norm=33.522, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.163e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-01 16:16:33,822 (trainer:762) INFO: 22epoch:train:6001-6500batch: iter_time=1.460e-04, forward_time=0.208, loss_ctc=43.873, loss_att=30.516, acc=0.810, loss=34.523, backward_time=0.229, grad_norm=35.857, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.161e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-01 16:24:12,974 (trainer:762) INFO: 22epoch:train:6501-7000batch: iter_time=1.452e-04, forward_time=0.206, loss_ctc=44.818, loss_att=30.893, acc=0.808, loss=35.070, backward_time=0.235, grad_norm=35.469, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.159e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-01 16:31:43,641 (trainer:762) INFO: 22epoch:train:7001-7500batch: iter_time=1.472e-04, forward_time=0.210, loss_ctc=42.333, loss_att=28.701, acc=0.809, loss=32.791, backward_time=0.234, grad_norm=35.400, clip=100.000, loss_scale=2.960e+33, optim_step_time=0.075, optim0_lr0=2.158e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-01 16:39:13,856 (trainer:762) INFO: 22epoch:train:7501-8000batch: iter_time=1.488e-04, forward_time=0.209, loss_ctc=41.850, loss_att=28.705, acc=0.821, loss=32.648, backward_time=0.235, grad_norm=35.777, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.156e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-01 16:46:47,108 (trainer:762) INFO: 22epoch:train:8001-8500batch: iter_time=1.471e-04, forward_time=0.210, loss_ctc=41.521, loss_att=28.767, acc=0.814, loss=32.593, backward_time=0.238, grad_norm=33.931, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.154e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-01 16:54:15,051 (trainer:762) INFO: 22epoch:train:8501-9000batch: iter_time=1.464e-04, forward_time=0.209, loss_ctc=44.667, loss_att=31.124, acc=0.809, loss=35.187, backward_time=0.236, grad_norm=36.039, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.153e-04, train_time=0.896
+[ip-10-0-216-33:0/16] 2024-03-01 17:01:42,168 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 17:01:49,412 (trainer:762) INFO: 22epoch:train:9001-9500batch: iter_time=1.523e-04, forward_time=0.210, loss_ctc=43.351, loss_att=30.362, acc=0.812, loss=34.258, backward_time=0.233, grad_norm=37.006, clip=100.000, loss_scale=5.827e+33, optim_step_time=0.075, optim0_lr0=2.151e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 17:09:18,638 (trainer:762) INFO: 22epoch:train:9501-10000batch: iter_time=1.482e-04, forward_time=0.209, loss_ctc=41.897, loss_att=28.623, acc=0.814, loss=32.605, backward_time=0.236, grad_norm=33.740, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.149e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-03-01 17:09:24,021 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 17:10:02,098 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 17:10:10,852 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f07e75100>)
+[ip-10-0-216-33:0/16] 2024-03-01 17:10:10,852 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-01 17:10:10,857 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 17:24:31,549 (trainer:762) INFO: 22epoch:train:10001-10500batch: iter_time=0.866, forward_time=0.209, loss_ctc=44.346, loss_att=30.124, acc=0.811, loss=34.391, backward_time=0.239, grad_norm=35.458, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.148e-04, train_time=1.826
+[ip-10-0-216-33:0/16] 2024-03-01 17:29:52,576 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 17:32:11,589 (trainer:762) INFO: 22epoch:train:10501-11000batch: iter_time=1.473e-04, forward_time=0.209, loss_ctc=42.521, loss_att=28.982, acc=0.819, loss=33.044, backward_time=0.230, grad_norm=34.098, clip=100.000, loss_scale=4.396e+33, optim_step_time=0.075, optim0_lr0=2.146e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-01 17:39:50,123 (trainer:762) INFO: 22epoch:train:11001-11500batch: iter_time=1.549e-04, forward_time=0.207, loss_ctc=43.566, loss_att=30.319, acc=0.811, loss=34.293, backward_time=0.235, grad_norm=36.184, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.144e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-01 17:47:20,569 (trainer:762) INFO: 22epoch:train:11501-12000batch: iter_time=1.529e-04, forward_time=0.211, loss_ctc=44.283, loss_att=30.541, acc=0.810, loss=34.663, backward_time=0.234, grad_norm=36.515, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.143e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-01 17:54:54,476 (trainer:762) INFO: 22epoch:train:12001-12500batch: iter_time=1.537e-04, forward_time=0.209, loss_ctc=42.024, loss_att=28.574, acc=0.810, loss=32.609, backward_time=0.229, grad_norm=35.484, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.141e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 18:02:24,464 (trainer:762) INFO: 22epoch:train:12501-13000batch: iter_time=1.480e-04, forward_time=0.211, loss_ctc=41.468, loss_att=28.546, acc=0.822, loss=32.423, backward_time=0.238, grad_norm=36.815, clip=100.000, loss_scale=3.391e+33, optim_step_time=0.075, optim0_lr0=2.139e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-01 18:09:58,574 (trainer:762) INFO: 22epoch:train:13001-13500batch: iter_time=1.494e-04, forward_time=0.209, loss_ctc=41.543, loss_att=28.859, acc=0.813, loss=32.664, backward_time=0.238, grad_norm=34.534, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.138e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 18:17:37,850 (trainer:762) INFO: 22epoch:train:13501-14000batch: iter_time=1.462e-04, forward_time=0.207, loss_ctc=44.414, loss_att=30.943, acc=0.809, loss=34.984, backward_time=0.234, grad_norm=36.757, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.136e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-01 18:25:19,351 (trainer:762) INFO: 22epoch:train:14001-14500batch: iter_time=1.516e-04, forward_time=0.207, loss_ctc=43.254, loss_att=30.011, acc=0.813, loss=33.984, backward_time=0.230, grad_norm=36.419, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.135e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-01 18:32:52,872 (trainer:762) INFO: 22epoch:train:14501-15000batch: iter_time=1.492e-04, forward_time=0.208, loss_ctc=41.711, loss_att=28.558, acc=0.814, loss=32.504, backward_time=0.230, grad_norm=33.574, clip=100.000, loss_scale=6.781e+33, optim_step_time=0.075, optim0_lr0=2.133e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-01 18:47:24,682 (trainer:361) INFO: 22epoch results: [train] iter_time=0.087, forward_time=0.209, loss_ctc=43.232, loss_att=29.737, acc=0.812, loss=33.786, backward_time=0.234, grad_norm=35.318, clip=100.000, loss_scale=4.275e+33, optim_step_time=0.075, optim0_lr0=2.157e-04, train_time=0.998, time=4 hours, 9 minutes and 57.4 seconds, total_count=330000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=40.064, cer_ctc=0.219, loss_att=30.057, acc=0.756, cer=0.218, wer=0.993, loss=33.059, time=14 minutes and 14.64 seconds, total_count=46926, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-01 18:47:34,259 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-01 18:47:34,284 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/17epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-01 18:47:34,284 (trainer:290) INFO: 23/45epoch started. Estimated time to finish: 4 days, 5 hours and 33 minutes
+[ip-10-0-216-33:0/16] 2024-03-01 18:47:34,291 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 18:48:10,000 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 18:48:17,849 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed05bd070>)
+[ip-10-0-216-33:0/16] 2024-03-01 18:48:17,849 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-01 18:48:17,854 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 18:57:56,100 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 19:02:23,464 (trainer:762) INFO: 23epoch:train:1-500batch: iter_time=0.839, forward_time=0.211, loss_ctc=44.181, loss_att=30.158, acc=0.819, loss=34.365, backward_time=0.233, grad_norm=35.276, clip=100.000, loss_scale=7.346e+33, optim_step_time=0.075, optim0_lr0=2.131e-04, train_time=1.778
+[ip-10-0-216-33:0/16] 2024-03-01 19:10:02,321 (trainer:762) INFO: 23epoch:train:501-1000batch: iter_time=1.487e-04, forward_time=0.208, loss_ctc=43.363, loss_att=29.459, acc=0.821, loss=33.630, backward_time=0.233, grad_norm=33.634, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.130e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-01 19:17:40,099 (trainer:762) INFO: 23epoch:train:1001-1500batch: iter_time=1.511e-04, forward_time=0.209, loss_ctc=44.458, loss_att=30.425, acc=0.810, loss=34.635, backward_time=0.235, grad_norm=35.138, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.128e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 19:25:19,497 (trainer:762) INFO: 23epoch:train:1501-2000batch: iter_time=1.479e-04, forward_time=0.209, loss_ctc=43.793, loss_att=30.212, acc=0.814, loss=34.286, backward_time=0.239, grad_norm=34.466, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.127e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-01 19:32:56,508 (trainer:762) INFO: 23epoch:train:2001-2500batch: iter_time=1.482e-04, forward_time=0.208, loss_ctc=40.844, loss_att=28.145, acc=0.819, loss=31.955, backward_time=0.237, grad_norm=35.066, clip=100.000, loss_scale=8.225e+33, optim_step_time=0.075, optim0_lr0=2.125e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-01 19:36:04,036 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 19:40:36,786 (trainer:762) INFO: 23epoch:train:2501-3000batch: iter_time=1.498e-04, forward_time=0.210, loss_ctc=44.833, loss_att=30.675, acc=0.819, loss=34.922, backward_time=0.234, grad_norm=35.044, clip=100.000, loss_scale=7.315e+33, optim_step_time=0.075, optim0_lr0=2.123e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-01 19:48:11,480 (trainer:762) INFO: 23epoch:train:3001-3500batch: iter_time=1.512e-04, forward_time=0.209, loss_ctc=41.355, loss_att=28.605, acc=0.813, loss=32.430, backward_time=0.238, grad_norm=34.551, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.122e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 19:55:44,713 (trainer:762) INFO: 23epoch:train:3501-4000batch: iter_time=1.512e-04, forward_time=0.209, loss_ctc=44.607, loss_att=30.263, acc=0.816, loss=34.566, backward_time=0.234, grad_norm=35.242, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.120e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-01 20:03:24,188 (trainer:762) INFO: 23epoch:train:4001-4500batch: iter_time=1.517e-04, forward_time=0.209, loss_ctc=43.262, loss_att=29.765, acc=0.814, loss=33.814, backward_time=0.235, grad_norm=36.185, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.119e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-01 20:10:11,933 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 20:11:01,795 (trainer:762) INFO: 23epoch:train:4501-5000batch: iter_time=1.508e-04, forward_time=0.209, loss_ctc=42.728, loss_att=29.255, acc=0.818, loss=33.297, backward_time=0.234, grad_norm=33.813, clip=100.000, loss_scale=7.690e+33, optim_step_time=0.075, optim0_lr0=2.117e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 20:11:06,582 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 20:11:43,205 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 20:11:51,207 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecffce490>)
+[ip-10-0-216-33:0/16] 2024-03-01 20:11:51,208 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-01 20:11:51,212 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 20:26:15,675 (trainer:762) INFO: 23epoch:train:5001-5500batch: iter_time=0.897, forward_time=0.209, loss_ctc=43.249, loss_att=29.564, acc=0.822, loss=33.669, backward_time=0.229, grad_norm=35.007, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.115e-04, train_time=1.828
+[ip-10-0-216-33:0/16] 2024-03-01 20:33:53,178 (trainer:762) INFO: 23epoch:train:5501-6000batch: iter_time=1.479e-04, forward_time=0.208, loss_ctc=42.800, loss_att=29.084, acc=0.822, loss=33.199, backward_time=0.236, grad_norm=32.996, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.114e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 20:41:27,561 (trainer:762) INFO: 23epoch:train:6001-6500batch: iter_time=1.589e-04, forward_time=0.209, loss_ctc=43.952, loss_att=30.038, acc=0.812, loss=34.212, backward_time=0.237, grad_norm=34.124, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.112e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 20:49:01,042 (trainer:762) INFO: 23epoch:train:6501-7000batch: iter_time=1.465e-04, forward_time=0.208, loss_ctc=43.513, loss_att=29.946, acc=0.814, loss=34.016, backward_time=0.229, grad_norm=35.219, clip=100.000, loss_scale=5.753e+33, optim_step_time=0.075, optim0_lr0=2.111e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-01 20:50:40,437 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 20:56:38,344 (trainer:762) INFO: 23epoch:train:7001-7500batch: iter_time=1.451e-04, forward_time=0.207, loss_ctc=40.742, loss_att=28.011, acc=0.820, loss=31.830, backward_time=0.233, grad_norm=56.709, clip=100.000, loss_scale=6.306e+33, optim_step_time=0.075, optim0_lr0=2.109e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-01 21:04:10,823 (trainer:762) INFO: 23epoch:train:7501-8000batch: iter_time=1.431e-04, forward_time=0.209, loss_ctc=44.597, loss_att=30.549, acc=0.819, loss=34.764, backward_time=0.235, grad_norm=35.454, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.108e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-01 21:11:45,360 (trainer:762) INFO: 23epoch:train:8001-8500batch: iter_time=1.423e-04, forward_time=0.208, loss_ctc=41.032, loss_att=28.346, acc=0.814, loss=32.152, backward_time=0.233, grad_norm=35.337, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.106e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-01 21:19:19,642 (trainer:762) INFO: 23epoch:train:8501-9000batch: iter_time=1.445e-04, forward_time=0.210, loss_ctc=44.194, loss_att=29.960, acc=0.817, loss=34.230, backward_time=0.232, grad_norm=34.246, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.104e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-01 21:26:50,787 (trainer:762) INFO: 23epoch:train:9001-9500batch: iter_time=1.486e-04, forward_time=0.209, loss_ctc=43.179, loss_att=29.743, acc=0.814, loss=33.774, backward_time=0.230, grad_norm=35.402, clip=100.000, loss_scale=9.263e+33, optim_step_time=0.075, optim0_lr0=2.103e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-01 21:27:02,724 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 21:33:02,094 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 21:34:22,062 (trainer:762) INFO: 23epoch:train:9501-10000batch: iter_time=1.479e-04, forward_time=0.210, loss_ctc=42.367, loss_att=28.998, acc=0.819, loss=33.009, backward_time=0.233, grad_norm=33.506, clip=100.000, loss_scale=4.859e+33, optim_step_time=0.075, optim0_lr0=2.101e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-01 21:34:26,726 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 21:35:04,092 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 21:35:12,102 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f0989cbe0>)
+[ip-10-0-216-33:0/16] 2024-03-01 21:35:12,103 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-01 21:35:12,107 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 21:49:41,024 (trainer:762) INFO: 23epoch:train:10001-10500batch: iter_time=0.898, forward_time=0.209, loss_ctc=43.192, loss_att=29.519, acc=0.822, loss=33.621, backward_time=0.231, grad_norm=34.880, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.100e-04, train_time=1.838
+[ip-10-0-216-33:0/16] 2024-03-01 21:57:17,908 (trainer:762) INFO: 23epoch:train:10501-11000batch: iter_time=1.530e-04, forward_time=0.209, loss_ctc=42.607, loss_att=29.037, acc=0.823, loss=33.108, backward_time=0.232, grad_norm=33.000, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.098e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-01 22:04:58,933 (trainer:762) INFO: 23epoch:train:11001-11500batch: iter_time=1.657e-04, forward_time=0.208, loss_ctc=43.632, loss_att=29.842, acc=0.813, loss=33.979, backward_time=0.232, grad_norm=34.437, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.097e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-01 22:12:36,998 (trainer:762) INFO: 23epoch:train:11501-12000batch: iter_time=1.536e-04, forward_time=0.209, loss_ctc=43.184, loss_att=29.847, acc=0.815, loss=33.848, backward_time=0.235, grad_norm=35.094, clip=100.000, loss_scale=3.053e+33, optim_step_time=0.075, optim0_lr0=2.095e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-01 22:20:06,973 (trainer:762) INFO: 23epoch:train:12001-12500batch: iter_time=1.457e-04, forward_time=0.210, loss_ctc=40.497, loss_att=27.851, acc=0.820, loss=31.645, backward_time=0.227, grad_norm=33.437, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.094e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-01 22:27:42,894 (trainer:762) INFO: 23epoch:train:12501-13000batch: iter_time=1.471e-04, forward_time=0.209, loss_ctc=44.674, loss_att=30.570, acc=0.819, loss=34.801, backward_time=0.234, grad_norm=35.020, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.092e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-01 22:35:19,722 (trainer:762) INFO: 23epoch:train:13001-13500batch: iter_time=1.508e-04, forward_time=0.207, loss_ctc=40.803, loss_att=28.248, acc=0.815, loss=32.015, backward_time=0.236, grad_norm=34.732, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.091e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-01 22:42:52,842 (trainer:762) INFO: 23epoch:train:13501-14000batch: iter_time=1.520e-04, forward_time=0.211, loss_ctc=43.912, loss_att=29.885, acc=0.818, loss=34.093, backward_time=0.231, grad_norm=34.886, clip=100.000, loss_scale=6.106e+33, optim_step_time=0.075, optim0_lr0=2.089e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-01 22:43:52,184 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 22:50:24,740 (trainer:762) INFO: 23epoch:train:14001-14500batch: iter_time=1.509e-04, forward_time=0.209, loss_ctc=42.713, loss_att=29.374, acc=0.815, loss=33.376, backward_time=0.230, grad_norm=35.018, clip=100.000, loss_scale=5.858e+33, optim_step_time=0.075, optim0_lr0=2.088e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-01 22:51:37,062 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 22:57:57,190 (trainer:762) INFO: 23epoch:train:14501-15000batch: iter_time=1.487e-04, forward_time=0.211, loss_ctc=42.386, loss_att=28.975, acc=0.820, loss=32.999, backward_time=0.233, grad_norm=38.102, clip=100.000, loss_scale=3.002e+33, optim_step_time=0.075, optim0_lr0=2.086e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-01 23:12:27,584 (trainer:361) INFO: 23epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=43.022, loss_att=29.478, acc=0.817, loss=33.541, backward_time=0.233, grad_norm=35.500, clip=100.000, loss_scale=5.348e+33, optim_step_time=0.075, optim0_lr0=2.108e-04, train_time=1.001, time=4 hours, 10 minutes and 39.33 seconds, total_count=345000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=40.329, cer_ctc=0.224, loss_att=30.798, acc=0.736, cer=0.276, wer=0.999, loss=33.657, time=14 minutes and 13.66 seconds, total_count=49059, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-01 23:12:37,368 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-01 23:12:37,398 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/16epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-01 23:12:37,398 (trainer:290) INFO: 24/45epoch started. Estimated time to finish: 4 days, 1 hour and 8 minutes
+[ip-10-0-216-33:0/16] 2024-03-01 23:12:37,405 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-01 23:13:13,087 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-01 23:13:20,925 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf3a89a0>)
+[ip-10-0-216-33:0/16] 2024-03-01 23:13:20,926 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-01 23:13:20,930 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-01 23:27:15,722 (trainer:762) INFO: 24epoch:train:1-500batch: iter_time=0.845, forward_time=0.210, loss_ctc=41.223, loss_att=28.462, acc=0.819, loss=32.290, backward_time=0.233, grad_norm=34.700, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.085e-04, train_time=1.756
+[ip-10-0-216-33:0/16] 2024-03-01 23:34:53,811 (trainer:762) INFO: 24epoch:train:501-1000batch: iter_time=1.538e-04, forward_time=0.209, loss_ctc=43.082, loss_att=29.547, acc=0.814, loss=33.607, backward_time=0.233, grad_norm=38.783, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.083e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-01 23:42:31,650 (trainer:762) INFO: 24epoch:train:1001-1500batch: iter_time=1.526e-04, forward_time=0.210, loss_ctc=41.477, loss_att=28.825, acc=0.817, loss=32.621, backward_time=0.233, grad_norm=35.025, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.082e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-01 23:50:07,661 (trainer:762) INFO: 24epoch:train:1501-2000batch: iter_time=1.537e-04, forward_time=0.209, loss_ctc=42.428, loss_att=29.581, acc=0.814, loss=33.435, backward_time=0.233, grad_norm=35.639, clip=100.000, loss_scale=4.782e+33, optim_step_time=0.075, optim0_lr0=2.080e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-01 23:53:57,976 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-01 23:57:45,949 (trainer:762) INFO: 24epoch:train:2001-2500batch: iter_time=1.519e-04, forward_time=0.210, loss_ctc=42.458, loss_att=29.215, acc=0.816, loss=33.188, backward_time=0.234, grad_norm=34.283, clip=100.000, loss_scale=3.897e+33, optim_step_time=0.075, optim0_lr0=2.079e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 00:05:31,638 (trainer:762) INFO: 24epoch:train:2501-3000batch: iter_time=1.543e-04, forward_time=0.208, loss_ctc=42.655, loss_att=29.231, acc=0.814, loss=33.259, backward_time=0.234, grad_norm=34.957, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.077e-04, train_time=0.931
+[ip-10-0-216-33:0/16] 2024-03-02 00:13:08,235 (trainer:762) INFO: 24epoch:train:3001-3500batch: iter_time=1.534e-04, forward_time=0.210, loss_ctc=41.740, loss_att=28.402, acc=0.816, loss=32.403, backward_time=0.235, grad_norm=35.128, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.076e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-02 00:20:45,072 (trainer:762) INFO: 24epoch:train:3501-4000batch: iter_time=1.492e-04, forward_time=0.211, loss_ctc=43.602, loss_att=30.107, acc=0.815, loss=34.156, backward_time=0.238, grad_norm=35.209, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.074e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-02 00:28:23,311 (trainer:762) INFO: 24epoch:train:4001-4500batch: iter_time=1.505e-04, forward_time=0.208, loss_ctc=41.162, loss_att=28.078, acc=0.824, loss=32.003, backward_time=0.232, grad_norm=33.399, clip=100.000, loss_scale=3.889e+33, optim_step_time=0.075, optim0_lr0=2.073e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 00:36:07,182 (trainer:762) INFO: 24epoch:train:4501-5000batch: iter_time=1.511e-04, forward_time=0.207, loss_ctc=43.105, loss_att=30.297, acc=0.812, loss=34.140, backward_time=0.233, grad_norm=36.459, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.071e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-03-02 00:36:11,913 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 00:36:48,514 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 00:36:56,533 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e9a201700>)
+[ip-10-0-216-33:0/16] 2024-03-02 00:36:56,534 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-02 00:36:56,538 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 00:47:35,911 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 00:51:24,267 (trainer:762) INFO: 24epoch:train:5001-5500batch: iter_time=0.904, forward_time=0.209, loss_ctc=40.911, loss_att=28.427, acc=0.819, loss=32.172, backward_time=0.235, grad_norm=35.218, clip=100.000, loss_scale=3.907e+33, optim_step_time=0.075, optim0_lr0=2.070e-04, train_time=1.834
+[ip-10-0-216-33:0/16] 2024-03-02 00:59:03,027 (trainer:762) INFO: 24epoch:train:5501-6000batch: iter_time=1.468e-04, forward_time=0.209, loss_ctc=42.563, loss_att=29.058, acc=0.815, loss=33.110, backward_time=0.236, grad_norm=35.821, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.068e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-02 01:06:34,218 (trainer:762) INFO: 24epoch:train:6001-6500batch: iter_time=1.495e-04, forward_time=0.209, loss_ctc=41.363, loss_att=28.729, acc=0.819, loss=32.519, backward_time=0.230, grad_norm=34.484, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.067e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-02 01:14:11,798 (trainer:762) INFO: 24epoch:train:6501-7000batch: iter_time=1.455e-04, forward_time=0.209, loss_ctc=42.188, loss_att=29.455, acc=0.815, loss=33.275, backward_time=0.232, grad_norm=35.513, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.065e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-02 01:21:50,655 (trainer:762) INFO: 24epoch:train:7001-7500batch: iter_time=1.600e-04, forward_time=0.207, loss_ctc=42.347, loss_att=29.154, acc=0.817, loss=33.112, backward_time=0.233, grad_norm=34.605, clip=100.000, loss_scale=3.879e+33, optim_step_time=0.075, optim0_lr0=2.064e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-02 01:29:25,713 (trainer:762) INFO: 24epoch:train:7501-8000batch: iter_time=1.458e-04, forward_time=0.209, loss_ctc=42.514, loss_att=29.203, acc=0.815, loss=33.197, backward_time=0.234, grad_norm=35.301, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.062e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-02 01:36:58,351 (trainer:762) INFO: 24epoch:train:8001-8500batch: iter_time=1.527e-04, forward_time=0.208, loss_ctc=41.544, loss_att=28.296, acc=0.816, loss=32.271, backward_time=0.233, grad_norm=34.374, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.061e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-02 01:44:29,525 (trainer:762) INFO: 24epoch:train:8501-9000batch: iter_time=1.484e-04, forward_time=0.211, loss_ctc=43.132, loss_att=29.911, acc=0.816, loss=33.877, backward_time=0.228, grad_norm=34.803, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.059e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-02 01:48:37,816 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 01:52:07,446 (trainer:762) INFO: 24epoch:train:9001-9500batch: iter_time=1.453e-04, forward_time=0.207, loss_ctc=40.759, loss_att=27.848, acc=0.825, loss=31.721, backward_time=0.232, grad_norm=32.131, clip=100.000, loss_scale=5.369e+33, optim_step_time=0.075, optim0_lr0=2.058e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 01:59:50,176 (trainer:762) INFO: 24epoch:train:9501-10000batch: iter_time=1.487e-04, forward_time=0.208, loss_ctc=42.871, loss_att=30.096, acc=0.813, loss=33.929, backward_time=0.234, grad_norm=35.983, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.056e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-02 01:59:54,833 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 02:00:32,303 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 02:00:40,326 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e9a20e7c0>)
+[ip-10-0-216-33:0/16] 2024-03-02 02:00:40,326 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-02 02:00:40,331 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 02:18:10,212 (trainer:762) INFO: 24epoch:train:10001-10500batch: iter_time=1.246, forward_time=0.210, loss_ctc=40.723, loss_att=28.347, acc=0.819, loss=32.060, backward_time=0.235, grad_norm=34.448, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.055e-04, train_time=2.200
+[ip-10-0-216-33:0/16] 2024-03-02 02:25:43,968 (trainer:762) INFO: 24epoch:train:10501-11000batch: iter_time=1.523e-04, forward_time=0.210, loss_ctc=42.420, loss_att=29.076, acc=0.816, loss=33.079, backward_time=0.234, grad_norm=35.618, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.054e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-02 02:33:22,362 (trainer:762) INFO: 24epoch:train:11001-11500batch: iter_time=1.480e-04, forward_time=0.210, loss_ctc=41.117, loss_att=28.501, acc=0.819, loss=32.286, backward_time=0.236, grad_norm=35.322, clip=100.000, loss_scale=7.570e+33, optim_step_time=0.075, optim0_lr0=2.052e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-02 02:41:05,608 (trainer:762) INFO: 24epoch:train:11501-12000batch: iter_time=1.529e-04, forward_time=0.208, loss_ctc=42.206, loss_att=29.416, acc=0.815, loss=33.253, backward_time=0.234, grad_norm=35.275, clip=100.000, loss_scale=1.038e+34, optim_step_time=0.075, optim0_lr0=2.051e-04, train_time=0.926
+[ip-10-0-216-33:0/16] 2024-03-02 02:42:26,244 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 02:48:46,456 (trainer:762) INFO: 24epoch:train:12001-12500batch: iter_time=1.520e-04, forward_time=0.210, loss_ctc=41.903, loss_att=29.012, acc=0.818, loss=32.879, backward_time=0.233, grad_norm=34.657, clip=100.000, loss_scale=6.087e+33, optim_step_time=0.075, optim0_lr0=2.049e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-02 02:56:22,007 (trainer:762) INFO: 24epoch:train:12501-13000batch: iter_time=1.507e-04, forward_time=0.208, loss_ctc=42.406, loss_att=29.091, acc=0.816, loss=33.085, backward_time=0.231, grad_norm=35.100, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.048e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-02 03:04:01,502 (trainer:762) INFO: 24epoch:train:13001-13500batch: iter_time=1.489e-04, forward_time=0.208, loss_ctc=41.379, loss_att=28.189, acc=0.817, loss=32.146, backward_time=0.239, grad_norm=34.781, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.046e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-02 03:11:42,808 (trainer:762) INFO: 24epoch:train:13501-14000batch: iter_time=1.517e-04, forward_time=0.208, loss_ctc=43.262, loss_att=30.009, acc=0.816, loss=33.985, backward_time=0.235, grad_norm=34.088, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.045e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-02 03:15:21,539 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 03:18:46,196 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 03:19:24,277 (trainer:762) INFO: 24epoch:train:14001-14500batch: iter_time=1.596e-04, forward_time=0.209, loss_ctc=40.528, loss_att=27.716, acc=0.826, loss=31.560, backward_time=0.237, grad_norm=32.885, clip=100.000, loss_scale=6.522e+33, optim_step_time=0.075, optim0_lr0=2.044e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-02 03:27:05,945 (trainer:762) INFO: 24epoch:train:14501-15000batch: iter_time=1.507e-04, forward_time=0.207, loss_ctc=42.478, loss_att=29.827, acc=0.814, loss=33.622, backward_time=0.235, grad_norm=36.510, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.042e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-02 03:41:30,117 (trainer:361) INFO: 24epoch results: [train] iter_time=0.100, forward_time=0.209, loss_ctc=42.051, loss_att=29.037, acc=0.817, loss=32.941, backward_time=0.234, grad_norm=35.017, clip=100.000, loss_scale=4.472e+33, optim_step_time=0.075, optim0_lr0=2.063e-04, train_time=1.018, time=4 hours, 14 minutes and 44.2 seconds, total_count=360000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=39.640, cer_ctc=0.217, loss_att=31.657, acc=0.728, cer=0.313, wer=1.000, loss=34.052, time=14 minutes and 8.21 seconds, total_count=51192, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-02 03:41:39,958 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-02 03:41:39,979 (trainer:290) INFO: 25/45epoch started. Estimated time to finish: 3 days, 20 hours and 47 minutes
+[ip-10-0-216-33:0/16] 2024-03-02 03:41:39,986 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 03:42:16,359 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 03:42:24,653 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed028b550>)
+[ip-10-0-216-33:0/16] 2024-03-02 03:42:24,654 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-02 03:42:24,658 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 03:56:16,256 (trainer:762) INFO: 25epoch:train:1-500batch: iter_time=0.841, forward_time=0.210, loss_ctc=42.746, loss_att=29.057, acc=0.817, loss=33.164, backward_time=0.231, grad_norm=37.279, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.041e-04, train_time=1.752
+[ip-10-0-216-33:0/16] 2024-03-02 04:03:54,854 (trainer:762) INFO: 25epoch:train:501-1000batch: iter_time=1.530e-04, forward_time=0.209, loss_ctc=41.092, loss_att=28.391, acc=0.812, loss=32.202, backward_time=0.233, grad_norm=36.933, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.039e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-02 04:11:33,832 (trainer:762) INFO: 25epoch:train:1001-1500batch: iter_time=1.514e-04, forward_time=0.208, loss_ctc=42.296, loss_att=28.685, acc=0.814, loss=32.769, backward_time=0.237, grad_norm=35.672, clip=100.000, loss_scale=2.809e+33, optim_step_time=0.075, optim0_lr0=2.038e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-02 04:19:07,664 (trainer:762) INFO: 25epoch:train:1501-2000batch: iter_time=1.496e-04, forward_time=0.210, loss_ctc=43.309, loss_att=29.958, acc=0.815, loss=33.964, backward_time=0.241, grad_norm=38.219, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.036e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-02 04:26:43,671 (trainer:762) INFO: 25epoch:train:2001-2500batch: iter_time=1.477e-04, forward_time=0.208, loss_ctc=42.798, loss_att=29.283, acc=0.816, loss=33.338, backward_time=0.233, grad_norm=36.222, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.035e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-02 04:27:42,055 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 04:34:25,162 (trainer:762) INFO: 25epoch:train:2501-3000batch: iter_time=1.497e-04, forward_time=0.209, loss_ctc=46.561, loss_att=31.675, acc=0.817, loss=36.141, backward_time=0.232, grad_norm=38.708, clip=100.000, loss_scale=2.924e+33, optim_step_time=0.075, optim0_lr0=2.034e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-02 04:41:59,302 (trainer:762) INFO: 25epoch:train:3001-3500batch: iter_time=1.514e-04, forward_time=0.209, loss_ctc=42.037, loss_att=28.308, acc=0.818, loss=32.427, backward_time=0.235, grad_norm=34.461, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.032e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 04:49:37,773 (trainer:762) INFO: 25epoch:train:3501-4000batch: iter_time=1.484e-04, forward_time=0.207, loss_ctc=42.857, loss_att=29.159, acc=0.810, loss=33.269, backward_time=0.235, grad_norm=36.050, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.031e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-02 04:57:14,107 (trainer:762) INFO: 25epoch:train:4001-4500batch: iter_time=1.527e-04, forward_time=0.210, loss_ctc=45.147, loss_att=30.658, acc=0.821, loss=35.005, backward_time=0.238, grad_norm=36.313, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.029e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-02 05:04:56,011 (trainer:762) INFO: 25epoch:train:4501-5000batch: iter_time=1.472e-04, forward_time=0.206, loss_ctc=42.938, loss_att=29.189, acc=0.808, loss=33.314, backward_time=0.230, grad_norm=36.517, clip=100.000, loss_scale=4.860e+33, optim_step_time=0.075, optim0_lr0=2.028e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-02 05:05:00,731 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 05:05:37,625 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 05:05:45,795 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecede4e50>)
+[ip-10-0-216-33:0/16] 2024-03-02 05:05:45,795 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-02 05:05:45,799 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 05:20:15,490 (trainer:762) INFO: 25epoch:train:5001-5500batch: iter_time=0.893, forward_time=0.207, loss_ctc=42.282, loss_att=28.683, acc=0.820, loss=32.763, backward_time=0.232, grad_norm=37.224, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.027e-04, train_time=1.839
+[ip-10-0-216-33:0/16] 2024-03-02 05:21:38,656 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 05:27:55,901 (trainer:762) INFO: 25epoch:train:5501-6000batch: iter_time=1.465e-04, forward_time=0.206, loss_ctc=40.653, loss_att=28.206, acc=0.813, loss=31.940, backward_time=0.231, grad_norm=36.338, clip=100.000, loss_scale=3.064e+33, optim_step_time=0.075, optim0_lr0=2.025e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-02 05:35:30,396 (trainer:762) INFO: 25epoch:train:6001-6500batch: iter_time=1.482e-04, forward_time=0.207, loss_ctc=41.812, loss_att=28.395, acc=0.815, loss=32.420, backward_time=0.239, grad_norm=35.894, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.024e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-02 05:43:07,089 (trainer:762) INFO: 25epoch:train:6501-7000batch: iter_time=1.450e-04, forward_time=0.207, loss_ctc=43.082, loss_att=29.813, acc=0.816, loss=33.794, backward_time=0.233, grad_norm=37.756, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.023e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-02 05:50:46,119 (trainer:762) INFO: 25epoch:train:7001-7500batch: iter_time=1.468e-04, forward_time=0.208, loss_ctc=41.928, loss_att=28.923, acc=0.817, loss=32.825, backward_time=0.236, grad_norm=34.661, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.021e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-02 05:58:21,186 (trainer:762) INFO: 25epoch:train:7501-8000batch: iter_time=1.475e-04, forward_time=0.209, loss_ctc=46.083, loss_att=31.396, acc=0.818, loss=35.803, backward_time=0.237, grad_norm=38.502, clip=100.000, loss_scale=4.720e+33, optim_step_time=0.075, optim0_lr0=2.020e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-02 06:05:56,153 (trainer:762) INFO: 25epoch:train:8001-8500batch: iter_time=1.469e-04, forward_time=0.209, loss_ctc=41.792, loss_att=28.192, acc=0.820, loss=32.272, backward_time=0.232, grad_norm=47.345, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.018e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-02 06:13:27,910 (trainer:762) INFO: 25epoch:train:8501-9000batch: iter_time=1.497e-04, forward_time=0.209, loss_ctc=42.461, loss_att=28.965, acc=0.811, loss=33.014, backward_time=0.234, grad_norm=35.801, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.017e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-02 06:21:02,125 (trainer:762) INFO: 25epoch:train:9001-9500batch: iter_time=1.488e-04, forward_time=0.209, loss_ctc=44.873, loss_att=30.484, acc=0.822, loss=34.800, backward_time=0.233, grad_norm=36.406, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.016e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 06:28:33,482 (trainer:762) INFO: 25epoch:train:9501-10000batch: iter_time=1.502e-04, forward_time=0.210, loss_ctc=42.623, loss_att=29.009, acc=0.809, loss=33.094, backward_time=0.229, grad_norm=36.064, clip=100.000, loss_scale=9.440e+33, optim_step_time=0.075, optim0_lr0=2.014e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-02 06:28:38,108 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 06:29:15,606 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 06:29:23,577 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f0dde5fa0>)
+[ip-10-0-216-33:0/16] 2024-03-02 06:29:23,577 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-02 06:29:23,582 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 06:40:19,855 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 06:43:45,994 (trainer:762) INFO: 25epoch:train:10001-10500batch: iter_time=0.897, forward_time=0.211, loss_ctc=42.013, loss_att=28.557, acc=0.820, loss=32.594, backward_time=0.230, grad_norm=35.737, clip=100.000, loss_scale=8.033e+33, optim_step_time=0.075, optim0_lr0=2.013e-04, train_time=1.825
+[ip-10-0-216-33:0/16] 2024-03-02 06:51:21,167 (trainer:762) INFO: 25epoch:train:10501-11000batch: iter_time=1.531e-04, forward_time=0.210, loss_ctc=40.231, loss_att=27.954, acc=0.814, loss=31.637, backward_time=0.235, grad_norm=36.584, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.012e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-02 06:58:54,083 (trainer:762) INFO: 25epoch:train:11001-11500batch: iter_time=1.529e-04, forward_time=0.211, loss_ctc=41.600, loss_att=28.229, acc=0.815, loss=32.241, backward_time=0.238, grad_norm=35.730, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.010e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-02 07:06:30,835 (trainer:762) INFO: 25epoch:train:11501-12000batch: iter_time=1.535e-04, forward_time=0.208, loss_ctc=42.881, loss_att=29.755, acc=0.817, loss=33.693, backward_time=0.232, grad_norm=36.970, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.009e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-02 07:10:53,554 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 07:14:08,329 (trainer:762) INFO: 25epoch:train:12001-12500batch: iter_time=1.484e-04, forward_time=0.210, loss_ctc=41.874, loss_att=28.794, acc=0.818, loss=32.718, backward_time=0.236, grad_norm=35.070, clip=100.000, loss_scale=5.328e+33, optim_step_time=0.075, optim0_lr0=2.008e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-02 07:21:45,553 (trainer:762) INFO: 25epoch:train:12501-13000batch: iter_time=1.494e-04, forward_time=0.209, loss_ctc=46.130, loss_att=31.400, acc=0.819, loss=35.819, backward_time=0.232, grad_norm=38.817, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=2.006e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-02 07:26:51,614 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 07:29:22,889 (trainer:762) INFO: 25epoch:train:13001-13500batch: iter_time=1.529e-04, forward_time=0.208, loss_ctc=41.715, loss_att=28.139, acc=0.820, loss=32.212, backward_time=0.230, grad_norm=50.256, clip=100.000, loss_scale=4.339e+33, optim_step_time=0.075, optim0_lr0=2.005e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-02 07:37:02,429 (trainer:762) INFO: 25epoch:train:13501-14000batch: iter_time=1.494e-04, forward_time=0.206, loss_ctc=42.183, loss_att=28.853, acc=0.812, loss=32.852, backward_time=0.226, grad_norm=35.612, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.004e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-02 07:44:40,249 (trainer:762) INFO: 25epoch:train:14001-14500batch: iter_time=1.465e-04, forward_time=0.209, loss_ctc=44.722, loss_att=30.339, acc=0.823, loss=34.654, backward_time=0.236, grad_norm=36.695, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.002e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-02 07:52:14,045 (trainer:762) INFO: 25epoch:train:14501-15000batch: iter_time=1.503e-04, forward_time=0.210, loss_ctc=42.443, loss_att=28.928, acc=0.809, loss=32.983, backward_time=0.238, grad_norm=36.293, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=2.001e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-02 08:09:40,900 (trainer:361) INFO: 25epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=42.839, loss_att=29.246, acc=0.816, loss=33.324, backward_time=0.234, grad_norm=37.337, clip=100.000, loss_scale=4.200e+33, optim_step_time=0.075, optim0_lr0=2.021e-04, train_time=1.002, time=4 hours, 10 minutes and 51.85 seconds, total_count=375000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=38.967, cer_ctc=0.216, loss_att=29.761, acc=0.749, cer=0.242, wer=0.996, loss=32.522, time=17 minutes and 8.74 seconds, total_count=53325, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-02 08:09:50,801 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-02 08:09:50,839 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/20epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-02 08:09:50,840 (trainer:290) INFO: 26/45epoch started. Estimated time to finish: 3 days, 16 hours and 24 minutes
+[ip-10-0-216-33:0/16] 2024-03-02 08:09:50,847 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 08:10:27,220 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 08:10:35,261 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed02bc640>)
+[ip-10-0-216-33:0/16] 2024-03-02 08:10:35,261 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-02 08:10:35,265 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 08:24:38,774 (trainer:762) INFO: 26epoch:train:1-500batch: iter_time=0.843, forward_time=0.209, loss_ctc=41.124, loss_att=28.048, acc=0.823, loss=31.971, backward_time=0.236, grad_norm=35.770, clip=100.000, loss_scale=3.448e+33, optim_step_time=0.075, optim0_lr0=2.000e-04, train_time=1.776
+[ip-10-0-216-33:0/16] 2024-03-02 08:32:14,535 (trainer:762) INFO: 26epoch:train:501-1000batch: iter_time=1.490e-04, forward_time=0.210, loss_ctc=44.775, loss_att=30.626, acc=0.820, loss=34.871, backward_time=0.234, grad_norm=37.231, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.998e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-02 08:39:49,814 (trainer:762) INFO: 26epoch:train:1001-1500batch: iter_time=1.487e-04, forward_time=0.209, loss_ctc=40.701, loss_att=28.244, acc=0.822, loss=31.981, backward_time=0.238, grad_norm=35.664, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.997e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-02 08:47:27,086 (trainer:762) INFO: 26epoch:train:1501-2000batch: iter_time=1.488e-04, forward_time=0.209, loss_ctc=43.070, loss_att=29.471, acc=0.827, loss=33.551, backward_time=0.236, grad_norm=35.857, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.996e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-02 08:53:21,534 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 08:54:19,539 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 08:55:01,013 (trainer:762) INFO: 26epoch:train:2001-2500batch: iter_time=1.508e-04, forward_time=0.210, loss_ctc=42.708, loss_att=29.606, acc=0.822, loss=33.537, backward_time=0.235, grad_norm=36.909, clip=100.000, loss_scale=5.505e+33, optim_step_time=0.075, optim0_lr0=1.994e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 09:02:35,605 (trainer:762) INFO: 26epoch:train:2501-3000batch: iter_time=1.506e-04, forward_time=0.209, loss_ctc=42.218, loss_att=29.131, acc=0.820, loss=33.058, backward_time=0.233, grad_norm=36.891, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.993e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-02 09:10:15,631 (trainer:762) INFO: 26epoch:train:3001-3500batch: iter_time=1.523e-04, forward_time=0.209, loss_ctc=40.255, loss_att=27.361, acc=0.823, loss=31.229, backward_time=0.236, grad_norm=35.531, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.992e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-02 09:17:58,883 (trainer:762) INFO: 26epoch:train:3501-4000batch: iter_time=1.477e-04, forward_time=0.209, loss_ctc=42.607, loss_att=29.586, acc=0.821, loss=33.492, backward_time=0.238, grad_norm=36.318, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.990e-04, train_time=0.926
+[ip-10-0-216-33:0/16] 2024-03-02 09:25:40,883 (trainer:762) INFO: 26epoch:train:4001-4500batch: iter_time=1.468e-04, forward_time=0.209, loss_ctc=42.280, loss_att=28.969, acc=0.823, loss=32.963, backward_time=0.232, grad_norm=34.822, clip=100.000, loss_scale=2.835e+33, optim_step_time=0.075, optim0_lr0=1.989e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-02 09:32:27,954 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 09:33:15,146 (trainer:762) INFO: 26epoch:train:4501-5000batch: iter_time=1.476e-04, forward_time=0.211, loss_ctc=43.168, loss_att=29.874, acc=0.817, loss=33.862, backward_time=0.235, grad_norm=37.870, clip=100.000, loss_scale=4.927e+33, optim_step_time=0.075, optim0_lr0=1.988e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 09:33:20,635 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 09:33:56,937 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 09:34:04,853 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed3a04be0>)
+[ip-10-0-216-33:0/16] 2024-03-02 09:34:04,853 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-02 09:34:04,858 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 09:48:25,853 (trainer:762) INFO: 26epoch:train:5001-5500batch: iter_time=0.887, forward_time=0.210, loss_ctc=40.932, loss_att=27.995, acc=0.823, loss=31.876, backward_time=0.235, grad_norm=35.834, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.986e-04, train_time=1.821
+[ip-10-0-216-33:0/16] 2024-03-02 09:55:59,670 (trainer:762) INFO: 26epoch:train:5501-6000batch: iter_time=1.434e-04, forward_time=0.210, loss_ctc=44.515, loss_att=30.665, acc=0.820, loss=34.820, backward_time=0.234, grad_norm=37.408, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.985e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-02 10:03:30,483 (trainer:762) INFO: 26epoch:train:6001-6500batch: iter_time=1.448e-04, forward_time=0.208, loss_ctc=40.413, loss_att=28.182, acc=0.822, loss=31.851, backward_time=0.235, grad_norm=35.045, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.984e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-02 10:11:07,890 (trainer:762) INFO: 26epoch:train:6501-7000batch: iter_time=1.409e-04, forward_time=0.210, loss_ctc=42.700, loss_att=29.266, acc=0.828, loss=33.297, backward_time=0.232, grad_norm=36.461, clip=100.000, loss_scale=2.861e+33, optim_step_time=0.075, optim0_lr0=1.982e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-02 10:18:42,176 (trainer:762) INFO: 26epoch:train:7001-7500batch: iter_time=1.415e-04, forward_time=0.209, loss_ctc=42.451, loss_att=29.495, acc=0.822, loss=33.382, backward_time=0.229, grad_norm=36.201, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.981e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 10:26:21,553 (trainer:762) INFO: 26epoch:train:7501-8000batch: iter_time=1.465e-04, forward_time=0.209, loss_ctc=41.989, loss_att=28.950, acc=0.821, loss=32.862, backward_time=0.233, grad_norm=36.293, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.980e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-02 10:33:58,001 (trainer:762) INFO: 26epoch:train:8001-8500batch: iter_time=1.453e-04, forward_time=0.207, loss_ctc=39.941, loss_att=27.146, acc=0.823, loss=30.985, backward_time=0.230, grad_norm=35.467, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.979e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-02 10:41:29,105 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 10:41:32,604 (trainer:762) INFO: 26epoch:train:8501-9000batch: iter_time=1.395e-04, forward_time=0.209, loss_ctc=42.319, loss_att=29.752, acc=0.821, loss=33.522, backward_time=0.234, grad_norm=36.901, clip=100.000, loss_scale=5.671e+33, optim_step_time=0.075, optim0_lr0=1.977e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-02 10:49:10,053 (trainer:762) INFO: 26epoch:train:9001-9500batch: iter_time=1.510e-04, forward_time=0.209, loss_ctc=41.794, loss_att=28.682, acc=0.824, loss=32.616, backward_time=0.237, grad_norm=34.051, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.976e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-02 10:54:34,077 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 10:56:50,141 (trainer:762) INFO: 26epoch:train:9501-10000batch: iter_time=1.426e-04, forward_time=0.206, loss_ctc=43.157, loss_att=29.922, acc=0.817, loss=33.892, backward_time=0.234, grad_norm=36.970, clip=100.000, loss_scale=4.417e+33, optim_step_time=0.075, optim0_lr0=1.975e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-02 10:56:55,297 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 10:57:32,349 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 10:57:40,457 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ee6cf9430>)
+[ip-10-0-216-33:0/16] 2024-03-02 10:57:40,458 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-02 10:57:40,462 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 11:12:10,763 (trainer:762) INFO: 26epoch:train:10001-10500batch: iter_time=0.912, forward_time=0.209, loss_ctc=40.718, loss_att=27.859, acc=0.824, loss=31.717, backward_time=0.231, grad_norm=35.258, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.973e-04, train_time=1.841
+[ip-10-0-216-33:0/16] 2024-03-02 11:19:44,519 (trainer:762) INFO: 26epoch:train:10501-11000batch: iter_time=1.460e-04, forward_time=0.210, loss_ctc=44.215, loss_att=30.528, acc=0.821, loss=34.634, backward_time=0.231, grad_norm=36.889, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.972e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-02 11:23:44,020 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 11:27:21,486 (trainer:762) INFO: 26epoch:train:11001-11500batch: iter_time=1.532e-04, forward_time=0.209, loss_ctc=40.267, loss_att=28.107, acc=0.822, loss=31.755, backward_time=0.229, grad_norm=63.949, clip=100.000, loss_scale=1.987e+33, optim_step_time=0.075, optim0_lr0=1.971e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-02 11:35:04,406 (trainer:762) INFO: 26epoch:train:11501-12000batch: iter_time=1.538e-04, forward_time=0.207, loss_ctc=42.530, loss_att=29.190, acc=0.828, loss=33.192, backward_time=0.233, grad_norm=35.424, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.970e-04, train_time=0.926
+[ip-10-0-216-33:0/16] 2024-03-02 11:42:45,178 (trainer:762) INFO: 26epoch:train:12001-12500batch: iter_time=1.497e-04, forward_time=0.209, loss_ctc=42.062, loss_att=29.299, acc=0.823, loss=33.128, backward_time=0.232, grad_norm=35.527, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.968e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-02 11:50:23,322 (trainer:762) INFO: 26epoch:train:12501-13000batch: iter_time=1.473e-04, forward_time=0.208, loss_ctc=41.781, loss_att=28.867, acc=0.821, loss=32.741, backward_time=0.231, grad_norm=36.030, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.967e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 11:57:58,918 (trainer:762) INFO: 26epoch:train:13001-13500batch: iter_time=1.444e-04, forward_time=0.208, loss_ctc=39.734, loss_att=26.988, acc=0.824, loss=30.812, backward_time=0.232, grad_norm=34.628, clip=100.000, loss_scale=1.906e+33, optim_step_time=0.075, optim0_lr0=1.966e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-02 12:05:36,968 (trainer:762) INFO: 26epoch:train:13501-14000batch: iter_time=1.490e-04, forward_time=0.208, loss_ctc=42.177, loss_att=29.494, acc=0.822, loss=33.299, backward_time=0.233, grad_norm=39.726, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.964e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 12:13:14,593 (trainer:762) INFO: 26epoch:train:14001-14500batch: iter_time=1.507e-04, forward_time=0.209, loss_ctc=41.762, loss_att=28.630, acc=0.825, loss=32.569, backward_time=0.237, grad_norm=33.585, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.963e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-02 12:20:51,919 (trainer:762) INFO: 26epoch:train:14501-15000batch: iter_time=1.514e-04, forward_time=0.208, loss_ctc=42.747, loss_att=29.624, acc=0.819, loss=33.561, backward_time=0.232, grad_norm=37.135, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.962e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-02 12:35:02,489 (trainer:361) INFO: 26epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=42.037, loss_att=28.985, acc=0.822, loss=32.901, backward_time=0.234, grad_norm=37.053, clip=100.000, loss_scale=3.411e+33, optim_step_time=0.075, optim0_lr0=1.981e-04, train_time=1.004, time=4 hours, 11 minutes and 19.56 seconds, total_count=390000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=38.746, cer_ctc=0.211, loss_att=30.664, acc=0.736, cer=0.312, wer=1.000, loss=33.089, time=13 minutes and 51.8 seconds, total_count=55458, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-02 12:35:12,234 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-02 12:35:12,270 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/21epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-02 12:35:12,270 (trainer:290) INFO: 27/45epoch started. Estimated time to finish: 3 days, 11 hours and 59 minutes
+[ip-10-0-216-33:0/16] 2024-03-02 12:35:12,277 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 12:35:48,293 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 12:35:56,574 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5eceb98b20>)
+[ip-10-0-216-33:0/16] 2024-03-02 12:35:56,574 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-02 12:35:56,579 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 12:49:51,500 (trainer:762) INFO: 27epoch:train:1-500batch: iter_time=0.843, forward_time=0.210, loss_ctc=43.093, loss_att=29.722, acc=0.820, loss=33.733, backward_time=0.236, grad_norm=38.367, clip=100.000, loss_scale=3.811e+33, optim_step_time=0.075, optim0_lr0=1.961e-04, train_time=1.758
+[ip-10-0-216-33:0/16] 2024-03-02 12:57:26,096 (trainer:762) INFO: 27epoch:train:501-1000batch: iter_time=1.500e-04, forward_time=0.209, loss_ctc=42.408, loss_att=29.334, acc=0.819, loss=33.256, backward_time=0.230, grad_norm=38.161, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.959e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-02 13:05:09,378 (trainer:762) INFO: 27epoch:train:1001-1500batch: iter_time=1.515e-04, forward_time=0.208, loss_ctc=41.784, loss_att=28.400, acc=0.821, loss=32.415, backward_time=0.236, grad_norm=37.123, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.958e-04, train_time=0.926
+[ip-10-0-216-33:0/16] 2024-03-02 13:12:53,125 (trainer:762) INFO: 27epoch:train:1501-2000batch: iter_time=1.487e-04, forward_time=0.208, loss_ctc=44.717, loss_att=30.929, acc=0.815, loss=35.065, backward_time=0.234, grad_norm=40.473, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.957e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-03-02 13:18:02,548 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 13:20:23,989 (trainer:762) INFO: 27epoch:train:2001-2500batch: iter_time=1.489e-04, forward_time=0.209, loss_ctc=40.644, loss_att=27.983, acc=0.816, loss=31.781, backward_time=0.231, grad_norm=37.460, clip=100.000, loss_scale=5.973e+33, optim_step_time=0.075, optim0_lr0=1.956e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-02 13:27:57,926 (trainer:762) INFO: 27epoch:train:2501-3000batch: iter_time=1.474e-04, forward_time=0.209, loss_ctc=44.070, loss_att=29.672, acc=0.817, loss=33.992, backward_time=0.232, grad_norm=38.546, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.954e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 13:35:32,185 (trainer:762) INFO: 27epoch:train:3001-3500batch: iter_time=1.496e-04, forward_time=0.210, loss_ctc=40.563, loss_att=27.688, acc=0.823, loss=31.551, backward_time=0.230, grad_norm=35.124, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.953e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 13:43:06,244 (trainer:762) INFO: 27epoch:train:3501-4000batch: iter_time=1.542e-04, forward_time=0.209, loss_ctc=39.792, loss_att=27.553, acc=0.827, loss=31.224, backward_time=0.236, grad_norm=34.627, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.952e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 13:50:43,498 (trainer:762) INFO: 27epoch:train:4001-4500batch: iter_time=1.472e-04, forward_time=0.209, loss_ctc=41.092, loss_att=29.742, acc=0.819, loss=33.147, backward_time=0.234, grad_norm=38.546, clip=100.000, loss_scale=6.833e+33, optim_step_time=0.075, optim0_lr0=1.951e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-02 13:55:17,364 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 13:58:17,185 (trainer:762) INFO: 27epoch:train:4501-5000batch: iter_time=1.535e-04, forward_time=0.209, loss_ctc=42.827, loss_att=29.389, acc=0.818, loss=33.421, backward_time=0.236, grad_norm=37.872, clip=100.000, loss_scale=8.356e+33, optim_step_time=0.075, optim0_lr0=1.950e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-02 13:58:21,529 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 13:58:58,264 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 13:59:06,982 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed046c700>)
+[ip-10-0-216-33:0/16] 2024-03-02 13:59:06,982 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-02 13:59:06,987 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 14:13:27,005 (trainer:762) INFO: 27epoch:train:5001-5500batch: iter_time=0.897, forward_time=0.210, loss_ctc=42.641, loss_att=29.337, acc=0.821, loss=33.328, backward_time=0.231, grad_norm=38.185, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.948e-04, train_time=1.819
+[ip-10-0-216-33:0/16] 2024-03-02 14:20:57,776 (trainer:762) INFO: 27epoch:train:5501-6000batch: iter_time=1.505e-04, forward_time=0.209, loss_ctc=42.016, loss_att=29.087, acc=0.820, loss=32.966, backward_time=0.227, grad_norm=37.145, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.947e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-02 14:28:35,196 (trainer:762) INFO: 27epoch:train:6001-6500batch: iter_time=1.480e-04, forward_time=0.209, loss_ctc=41.554, loss_att=28.230, acc=0.822, loss=32.228, backward_time=0.234, grad_norm=39.478, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.946e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-02 14:34:52,034 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 14:36:05,967 (trainer:762) INFO: 27epoch:train:6501-7000batch: iter_time=1.454e-04, forward_time=0.208, loss_ctc=44.197, loss_att=30.548, acc=0.816, loss=34.643, backward_time=0.223, grad_norm=39.443, clip=100.000, loss_scale=6.358e+33, optim_step_time=0.075, optim0_lr0=1.945e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-02 14:43:18,958 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 14:43:33,194 (trainer:762) INFO: 27epoch:train:7001-7500batch: iter_time=1.483e-04, forward_time=0.209, loss_ctc=40.145, loss_att=27.563, acc=0.817, loss=31.337, backward_time=0.236, grad_norm=36.682, clip=100.000, loss_scale=5.109e+33, optim_step_time=0.075, optim0_lr0=1.943e-04, train_time=0.894
+[ip-10-0-216-33:0/16] 2024-03-02 14:51:07,900 (trainer:762) INFO: 27epoch:train:7501-8000batch: iter_time=1.467e-04, forward_time=0.209, loss_ctc=43.477, loss_att=29.485, acc=0.817, loss=33.683, backward_time=0.238, grad_norm=38.434, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.942e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-02 14:58:41,305 (trainer:762) INFO: 27epoch:train:8001-8500batch: iter_time=1.465e-04, forward_time=0.210, loss_ctc=40.241, loss_att=27.514, acc=0.824, loss=31.332, backward_time=0.236, grad_norm=36.393, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.941e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-02 15:06:14,053 (trainer:762) INFO: 27epoch:train:8501-9000batch: iter_time=1.459e-04, forward_time=0.208, loss_ctc=39.474, loss_att=27.383, acc=0.828, loss=31.010, backward_time=0.234, grad_norm=36.086, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.940e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-02 15:13:48,741 (trainer:762) INFO: 27epoch:train:9001-9500batch: iter_time=1.477e-04, forward_time=0.210, loss_ctc=41.084, loss_att=29.807, acc=0.819, loss=33.190, backward_time=0.229, grad_norm=36.699, clip=100.000, loss_scale=2.679e+33, optim_step_time=0.075, optim0_lr0=1.938e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-02 15:21:18,471 (trainer:762) INFO: 27epoch:train:9501-10000batch: iter_time=1.457e-04, forward_time=0.208, loss_ctc=42.629, loss_att=29.202, acc=0.820, loss=33.230, backward_time=0.237, grad_norm=37.755, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.937e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-02 15:21:23,018 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 15:22:00,373 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 15:22:09,339 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f01af5c40>)
+[ip-10-0-216-33:0/16] 2024-03-02 15:22:09,339 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-02 15:22:09,343 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 15:36:37,748 (trainer:762) INFO: 27epoch:train:10001-10500batch: iter_time=0.892, forward_time=0.208, loss_ctc=42.488, loss_att=29.301, acc=0.821, loss=33.257, backward_time=0.233, grad_norm=38.236, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.936e-04, train_time=1.838
+[ip-10-0-216-33:0/16] 2024-03-02 15:44:18,750 (trainer:762) INFO: 27epoch:train:10501-11000batch: iter_time=1.524e-04, forward_time=0.208, loss_ctc=42.162, loss_att=29.195, acc=0.820, loss=33.085, backward_time=0.233, grad_norm=37.698, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.935e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-02 15:44:47,438 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 15:51:57,464 (trainer:762) INFO: 27epoch:train:11001-11500batch: iter_time=1.499e-04, forward_time=0.207, loss_ctc=41.366, loss_att=28.130, acc=0.823, loss=32.101, backward_time=0.236, grad_norm=36.613, clip=100.000, loss_scale=2.752e+33, optim_step_time=0.075, optim0_lr0=1.934e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-02 15:59:36,699 (trainer:762) INFO: 27epoch:train:11501-12000batch: iter_time=1.528e-04, forward_time=0.208, loss_ctc=44.216, loss_att=30.707, acc=0.815, loss=34.759, backward_time=0.231, grad_norm=40.094, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.932e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-02 16:07:09,453 (trainer:762) INFO: 27epoch:train:12001-12500batch: iter_time=1.558e-04, forward_time=0.209, loss_ctc=40.100, loss_att=27.590, acc=0.817, loss=31.343, backward_time=0.232, grad_norm=37.320, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.931e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-02 16:14:43,711 (trainer:762) INFO: 27epoch:train:12501-13000batch: iter_time=1.472e-04, forward_time=0.209, loss_ctc=43.342, loss_att=29.399, acc=0.818, loss=33.582, backward_time=0.231, grad_norm=37.640, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.930e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 16:22:03,768 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 16:22:17,998 (trainer:762) INFO: 27epoch:train:13001-13500batch: iter_time=1.522e-04, forward_time=0.209, loss_ctc=40.210, loss_att=27.445, acc=0.825, loss=31.274, backward_time=0.230, grad_norm=36.009, clip=100.000, loss_scale=4.948e+33, optim_step_time=0.075, optim0_lr0=1.929e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 16:29:50,651 (trainer:762) INFO: 27epoch:train:13501-14000batch: iter_time=1.548e-04, forward_time=0.210, loss_ctc=39.398, loss_att=27.314, acc=0.828, loss=30.939, backward_time=0.231, grad_norm=35.230, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.928e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-02 16:37:24,742 (trainer:762) INFO: 27epoch:train:14001-14500batch: iter_time=1.526e-04, forward_time=0.209, loss_ctc=40.980, loss_att=29.646, acc=0.819, loss=33.046, backward_time=0.232, grad_norm=35.930, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.926e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 16:45:02,268 (trainer:762) INFO: 27epoch:train:14501-15000batch: iter_time=1.511e-04, forward_time=0.208, loss_ctc=42.378, loss_att=29.149, acc=0.819, loss=33.118, backward_time=0.227, grad_norm=39.063, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.925e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-02 16:59:21,887 (trainer:361) INFO: 27epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=41.836, loss_att=28.881, acc=0.820, loss=32.768, backward_time=0.233, grad_norm=37.548, clip=100.000, loss_scale=4.416e+33, optim_step_time=0.075, optim0_lr0=1.943e-04, train_time=0.999, time=4 hours, 10 minutes and 5.44 seconds, total_count=405000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=38.325, cer_ctc=0.210, loss_att=29.635, acc=0.746, cer=0.279, wer=0.997, loss=32.242, time=14 minutes and 3.86 seconds, total_count=57591, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-02 16:59:31,962 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-02 16:59:31,985 (trainer:290) INFO: 28/45epoch started. Estimated time to finish: 3 days, 7 hours and 33 minutes
+[ip-10-0-216-33:0/16] 2024-03-02 16:59:31,992 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 17:00:07,955 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 17:00:15,824 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed06dd790>)
+[ip-10-0-216-33:0/16] 2024-03-02 17:00:15,824 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-02 17:00:15,829 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 17:14:16,462 (trainer:762) INFO: 28epoch:train:1-500batch: iter_time=0.815, forward_time=0.210, loss_ctc=41.010, loss_att=28.048, acc=0.819, loss=31.937, backward_time=0.231, grad_norm=39.636, clip=100.000, loss_scale=2.679e+33, optim_step_time=0.075, optim0_lr0=1.924e-04, train_time=1.769
+[ip-10-0-216-33:0/16] 2024-03-02 17:21:57,966 (trainer:762) INFO: 28epoch:train:501-1000batch: iter_time=1.532e-04, forward_time=0.209, loss_ctc=42.447, loss_att=28.749, acc=0.819, loss=32.858, backward_time=0.233, grad_norm=39.133, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.923e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-02 17:29:35,058 (trainer:762) INFO: 28epoch:train:1001-1500batch: iter_time=1.564e-04, forward_time=0.209, loss_ctc=40.563, loss_att=28.451, acc=0.822, loss=32.085, backward_time=0.233, grad_norm=38.689, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.076, optim0_lr0=1.922e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-02 17:37:08,174 (trainer:762) INFO: 28epoch:train:1501-2000batch: iter_time=1.515e-04, forward_time=0.209, loss_ctc=40.257, loss_att=27.946, acc=0.819, loss=31.640, backward_time=0.235, grad_norm=37.286, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.921e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-02 17:44:46,106 (trainer:762) INFO: 28epoch:train:2001-2500batch: iter_time=1.506e-04, forward_time=0.209, loss_ctc=44.270, loss_att=30.216, acc=0.811, loss=34.432, backward_time=0.233, grad_norm=39.508, clip=100.000, loss_scale=5.358e+33, optim_step_time=0.075, optim0_lr0=1.919e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 17:48:00,331 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 17:51:30,985 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 17:52:19,866 (trainer:762) INFO: 28epoch:train:2501-3000batch: iter_time=1.574e-04, forward_time=0.208, loss_ctc=41.356, loss_att=28.431, acc=0.817, loss=32.308, backward_time=0.230, grad_norm=38.191, clip=100.000, loss_scale=7.121e+33, optim_step_time=0.075, optim0_lr0=1.918e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-02 17:59:54,803 (trainer:762) INFO: 28epoch:train:3001-3500batch: iter_time=1.486e-04, forward_time=0.209, loss_ctc=40.606, loss_att=28.211, acc=0.816, loss=31.930, backward_time=0.232, grad_norm=38.288, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.917e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-02 18:07:32,896 (trainer:762) INFO: 28epoch:train:3501-4000batch: iter_time=1.497e-04, forward_time=0.208, loss_ctc=42.258, loss_att=29.132, acc=0.816, loss=33.069, backward_time=0.230, grad_norm=39.642, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.916e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 18:15:07,943 (trainer:762) INFO: 28epoch:train:4001-4500batch: iter_time=1.474e-04, forward_time=0.209, loss_ctc=42.033, loss_att=28.977, acc=0.822, loss=32.894, backward_time=0.237, grad_norm=39.158, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.915e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-02 18:22:44,537 (trainer:762) INFO: 28epoch:train:4501-5000batch: iter_time=1.473e-04, forward_time=0.210, loss_ctc=40.594, loss_att=28.007, acc=0.825, loss=31.783, backward_time=0.234, grad_norm=35.749, clip=100.000, loss_scale=2.877e+33, optim_step_time=0.075, optim0_lr0=1.914e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-02 18:22:49,171 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 18:23:25,724 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 18:23:33,695 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ea4ff19a0>)
+[ip-10-0-216-33:0/16] 2024-03-02 18:23:33,695 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-02 18:23:33,700 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 18:37:58,111 (trainer:762) INFO: 28epoch:train:5001-5500batch: iter_time=0.907, forward_time=0.210, loss_ctc=40.574, loss_att=27.889, acc=0.821, loss=31.694, backward_time=0.234, grad_norm=38.630, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.912e-04, train_time=1.827
+[ip-10-0-216-33:0/16] 2024-03-02 18:45:25,631 (trainer:762) INFO: 28epoch:train:5501-6000batch: iter_time=1.475e-04, forward_time=0.209, loss_ctc=42.180, loss_att=28.559, acc=0.820, loss=32.645, backward_time=0.232, grad_norm=39.034, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.911e-04, train_time=0.895
+[ip-10-0-216-33:0/16] 2024-03-02 18:52:54,639 (trainer:762) INFO: 28epoch:train:6001-6500batch: iter_time=1.472e-04, forward_time=0.210, loss_ctc=40.108, loss_att=28.072, acc=0.824, loss=31.683, backward_time=0.234, grad_norm=38.502, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.910e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-03-02 19:00:23,796 (trainer:762) INFO: 28epoch:train:6501-7000batch: iter_time=1.423e-04, forward_time=0.209, loss_ctc=39.958, loss_att=27.671, acc=0.820, loss=31.357, backward_time=0.229, grad_norm=37.414, clip=100.000, loss_scale=5.753e+33, optim_step_time=0.075, optim0_lr0=1.909e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-03-02 19:03:41,740 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 19:07:57,411 (trainer:762) INFO: 28epoch:train:7001-7500batch: iter_time=1.457e-04, forward_time=0.208, loss_ctc=44.022, loss_att=30.053, acc=0.812, loss=34.244, backward_time=0.233, grad_norm=39.863, clip=100.000, loss_scale=7.450e+33, optim_step_time=0.075, optim0_lr0=1.908e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-02 19:15:28,088 (trainer:762) INFO: 28epoch:train:7501-8000batch: iter_time=1.469e-04, forward_time=0.210, loss_ctc=41.047, loss_att=28.174, acc=0.819, loss=32.036, backward_time=0.237, grad_norm=38.304, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.907e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-02 19:22:58,369 (trainer:762) INFO: 28epoch:train:8001-8500batch: iter_time=1.453e-04, forward_time=0.209, loss_ctc=40.154, loss_att=27.981, acc=0.816, loss=31.633, backward_time=0.233, grad_norm=36.969, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.905e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-02 19:30:27,336 (trainer:762) INFO: 28epoch:train:8501-9000batch: iter_time=1.439e-04, forward_time=0.210, loss_ctc=42.132, loss_att=29.001, acc=0.817, loss=32.941, backward_time=0.231, grad_norm=39.239, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.904e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-03-02 19:37:57,311 (trainer:762) INFO: 28epoch:train:9001-9500batch: iter_time=1.434e-04, forward_time=0.209, loss_ctc=41.687, loss_att=28.791, acc=0.823, loss=32.660, backward_time=0.233, grad_norm=37.382, clip=100.000, loss_scale=8.121e+33, optim_step_time=0.075, optim0_lr0=1.903e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-02 19:40:33,368 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 19:45:26,943 (trainer:762) INFO: 28epoch:train:9501-10000batch: iter_time=1.421e-04, forward_time=0.210, loss_ctc=40.471, loss_att=28.000, acc=0.826, loss=31.741, backward_time=0.233, grad_norm=36.619, clip=100.000, loss_scale=6.972e+33, optim_step_time=0.075, optim0_lr0=1.902e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-02 19:45:31,724 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 19:46:08,833 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 19:46:16,852 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f033fb1c0>)
+[ip-10-0-216-33:0/16] 2024-03-02 19:46:16,852 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-02 19:46:16,856 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 20:00:44,645 (trainer:762) INFO: 28epoch:train:10001-10500batch: iter_time=0.883, forward_time=0.208, loss_ctc=40.403, loss_att=27.616, acc=0.822, loss=31.452, backward_time=0.233, grad_norm=38.118, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.901e-04, train_time=1.835
+[ip-10-0-216-33:0/16] 2024-03-02 20:08:24,185 (trainer:762) INFO: 28epoch:train:10501-11000batch: iter_time=1.480e-04, forward_time=0.207, loss_ctc=41.943, loss_att=28.497, acc=0.820, loss=32.531, backward_time=0.231, grad_norm=39.578, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.900e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-02 20:16:06,097 (trainer:762) INFO: 28epoch:train:11001-11500batch: iter_time=1.489e-04, forward_time=0.205, loss_ctc=40.069, loss_att=28.066, acc=0.824, loss=31.667, backward_time=0.227, grad_norm=39.250, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.899e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-02 20:23:43,572 (trainer:762) INFO: 28epoch:train:11501-12000batch: iter_time=1.534e-04, forward_time=0.206, loss_ctc=39.679, loss_att=27.641, acc=0.821, loss=31.253, backward_time=0.234, grad_norm=51.599, clip=100.000, loss_scale=8.598e+33, optim_step_time=0.075, optim0_lr0=1.897e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-02 20:30:40,852 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 20:31:16,440 (trainer:762) INFO: 28epoch:train:12001-12500batch: iter_time=1.459e-04, forward_time=0.211, loss_ctc=43.932, loss_att=30.003, acc=0.813, loss=34.181, backward_time=0.237, grad_norm=38.954, clip=100.000, loss_scale=9.989e+33, optim_step_time=0.075, optim0_lr0=1.896e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-02 20:38:55,430 (trainer:762) INFO: 28epoch:train:12501-13000batch: iter_time=1.472e-04, forward_time=0.206, loss_ctc=40.507, loss_att=27.901, acc=0.820, loss=31.682, backward_time=0.233, grad_norm=37.499, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.076, optim0_lr0=1.895e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-02 20:46:33,535 (trainer:762) INFO: 28epoch:train:13001-13500batch: iter_time=1.464e-04, forward_time=0.208, loss_ctc=39.956, loss_att=27.822, acc=0.818, loss=31.462, backward_time=0.233, grad_norm=38.028, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.894e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 20:54:05,727 (trainer:762) INFO: 28epoch:train:13501-14000batch: iter_time=1.476e-04, forward_time=0.209, loss_ctc=41.893, loss_att=28.805, acc=0.818, loss=32.731, backward_time=0.233, grad_norm=37.900, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.893e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-02 21:01:37,594 (trainer:762) INFO: 28epoch:train:14001-14500batch: iter_time=1.463e-04, forward_time=0.210, loss_ctc=41.586, loss_att=28.754, acc=0.823, loss=32.604, backward_time=0.231, grad_norm=37.912, clip=100.000, loss_scale=5.587e+33, optim_step_time=0.076, optim0_lr0=1.892e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-02 21:04:07,295 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 21:09:12,444 (trainer:762) INFO: 28epoch:train:14501-15000batch: iter_time=1.467e-04, forward_time=0.209, loss_ctc=40.252, loss_att=27.920, acc=0.826, loss=31.620, backward_time=0.233, grad_norm=35.870, clip=100.000, loss_scale=6.878e+33, optim_step_time=0.076, optim0_lr0=1.891e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-02 21:23:25,328 (trainer:361) INFO: 28epoch results: [train] iter_time=0.087, forward_time=0.209, loss_ctc=41.265, loss_att=28.446, acc=0.820, loss=32.292, backward_time=0.233, grad_norm=38.732, clip=100.000, loss_scale=5.434e+33, optim_step_time=0.075, optim0_lr0=1.907e-04, train_time=0.998, time=4 hours, 9 minutes and 57.56 seconds, total_count=420000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=38.202, cer_ctc=0.207, loss_att=30.050, acc=0.759, cer=0.222, wer=0.989, loss=32.496, time=13 minutes and 55.47 seconds, total_count=59724, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-02 21:23:35,268 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-02 21:23:35,304 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/18epoch.pth, exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/23epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-02 21:23:35,305 (trainer:290) INFO: 29/45epoch started. Estimated time to finish: 3 days, 3 hours and 7 minutes
+[ip-10-0-216-33:0/16] 2024-03-02 21:23:35,312 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 21:24:11,252 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 21:24:19,092 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecdf36190>)
+[ip-10-0-216-33:0/16] 2024-03-02 21:24:19,093 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-02 21:24:19,097 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 21:38:09,667 (trainer:762) INFO: 29epoch:train:1-500batch: iter_time=0.826, forward_time=0.209, loss_ctc=40.974, loss_att=28.423, acc=0.820, loss=32.188, backward_time=0.232, grad_norm=43.213, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.076, optim0_lr0=1.889e-04, train_time=1.748
+[ip-10-0-216-33:0/16] 2024-03-02 21:45:47,572 (trainer:762) INFO: 29epoch:train:501-1000batch: iter_time=1.500e-04, forward_time=0.209, loss_ctc=41.458, loss_att=28.301, acc=0.824, loss=32.248, backward_time=0.231, grad_norm=38.420, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.888e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 21:53:22,331 (trainer:762) INFO: 29epoch:train:1001-1500batch: iter_time=1.558e-04, forward_time=0.210, loss_ctc=43.009, loss_att=29.966, acc=0.820, loss=33.879, backward_time=0.232, grad_norm=40.358, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.887e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-02 21:56:50,327 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 22:00:57,296 (trainer:762) INFO: 29epoch:train:1501-2000batch: iter_time=1.490e-04, forward_time=0.209, loss_ctc=40.310, loss_att=27.663, acc=0.823, loss=31.457, backward_time=0.230, grad_norm=37.586, clip=100.000, loss_scale=5.858e+33, optim_step_time=0.075, optim0_lr0=1.886e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-02 22:08:32,638 (trainer:762) INFO: 29epoch:train:2001-2500batch: iter_time=1.471e-04, forward_time=0.210, loss_ctc=41.161, loss_att=29.108, acc=0.821, loss=32.724, backward_time=0.232, grad_norm=39.255, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.885e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-02 22:16:07,574 (trainer:762) INFO: 29epoch:train:2501-3000batch: iter_time=1.487e-04, forward_time=0.209, loss_ctc=41.001, loss_att=27.796, acc=0.819, loss=31.758, backward_time=0.235, grad_norm=38.371, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.884e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-02 22:18:19,514 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 22:23:42,169 (trainer:762) INFO: 29epoch:train:3001-3500batch: iter_time=1.495e-04, forward_time=0.209, loss_ctc=42.031, loss_att=28.616, acc=0.831, loss=32.640, backward_time=0.232, grad_norm=34.893, clip=100.000, loss_scale=3.345e+33, optim_step_time=0.075, optim0_lr0=1.883e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-02 22:31:16,747 (trainer:762) INFO: 29epoch:train:3501-4000batch: iter_time=1.482e-04, forward_time=0.209, loss_ctc=39.694, loss_att=27.773, acc=0.822, loss=31.350, backward_time=0.232, grad_norm=42.499, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.882e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-02 22:38:57,542 (trainer:762) INFO: 29epoch:train:4001-4500batch: iter_time=1.464e-04, forward_time=0.208, loss_ctc=40.521, loss_att=27.829, acc=0.827, loss=31.637, backward_time=0.237, grad_norm=37.836, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.881e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-02 22:46:36,020 (trainer:762) INFO: 29epoch:train:4501-5000batch: iter_time=1.491e-04, forward_time=0.208, loss_ctc=39.527, loss_att=27.274, acc=0.826, loss=30.950, backward_time=0.233, grad_norm=36.904, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.879e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-02 22:46:40,313 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-02 22:47:17,312 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-02 22:47:25,346 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed42adb50>)
+[ip-10-0-216-33:0/16] 2024-03-02 22:47:25,347 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-02 22:47:25,351 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-02 22:56:04,638 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-02 23:01:47,524 (trainer:762) INFO: 29epoch:train:5001-5500batch: iter_time=0.898, forward_time=0.207, loss_ctc=40.649, loss_att=28.026, acc=0.822, loss=31.813, backward_time=0.235, grad_norm=38.983, clip=100.000, loss_scale=1.631e+33, optim_step_time=0.075, optim0_lr0=1.878e-04, train_time=1.823
+[ip-10-0-216-33:0/16] 2024-03-02 23:09:25,628 (trainer:762) INFO: 29epoch:train:5501-6000batch: iter_time=1.456e-04, forward_time=0.208, loss_ctc=41.222, loss_att=28.163, acc=0.824, loss=32.081, backward_time=0.235, grad_norm=39.282, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.877e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 23:16:59,621 (trainer:762) INFO: 29epoch:train:6001-6500batch: iter_time=1.480e-04, forward_time=0.207, loss_ctc=42.537, loss_att=29.615, acc=0.821, loss=33.492, backward_time=0.234, grad_norm=40.508, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.876e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-02 23:24:45,176 (trainer:762) INFO: 29epoch:train:6501-7000batch: iter_time=1.468e-04, forward_time=0.205, loss_ctc=40.122, loss_att=27.564, acc=0.824, loss=31.331, backward_time=0.236, grad_norm=37.654, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.875e-04, train_time=0.931
+[ip-10-0-216-33:0/16] 2024-03-02 23:32:22,580 (trainer:762) INFO: 29epoch:train:7001-7500batch: iter_time=1.474e-04, forward_time=0.207, loss_ctc=40.996, loss_att=28.898, acc=0.822, loss=32.527, backward_time=0.235, grad_norm=38.903, clip=100.000, loss_scale=2.261e+33, optim_step_time=0.075, optim0_lr0=1.874e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-02 23:40:00,894 (trainer:762) INFO: 29epoch:train:7501-8000batch: iter_time=1.454e-04, forward_time=0.207, loss_ctc=40.607, loss_att=27.589, acc=0.819, loss=31.494, backward_time=0.232, grad_norm=37.849, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.873e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-02 23:47:43,286 (trainer:762) INFO: 29epoch:train:8001-8500batch: iter_time=1.472e-04, forward_time=0.208, loss_ctc=41.710, loss_att=28.442, acc=0.832, loss=32.422, backward_time=0.230, grad_norm=35.289, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.872e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-02 23:55:22,500 (trainer:762) INFO: 29epoch:train:8501-9000batch: iter_time=1.529e-04, forward_time=0.206, loss_ctc=39.161, loss_att=27.539, acc=0.822, loss=31.026, backward_time=0.232, grad_norm=36.382, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.871e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 00:03:02,936 (trainer:762) INFO: 29epoch:train:9001-9500batch: iter_time=1.500e-04, forward_time=0.208, loss_ctc=40.175, loss_att=27.621, acc=0.828, loss=31.388, backward_time=0.229, grad_norm=37.610, clip=100.000, loss_scale=4.522e+33, optim_step_time=0.075, optim0_lr0=1.870e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-03 00:10:34,501 (trainer:762) INFO: 29epoch:train:9501-10000batch: iter_time=1.460e-04, forward_time=0.208, loss_ctc=39.213, loss_att=27.067, acc=0.828, loss=30.711, backward_time=0.231, grad_norm=36.409, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.868e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-03 00:10:39,116 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 00:11:16,395 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 00:11:24,637 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f0be31b20>)
+[ip-10-0-216-33:0/16] 2024-03-03 00:11:24,638 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-03 00:11:24,642 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 00:25:40,141 (trainer:762) INFO: 29epoch:train:10001-10500batch: iter_time=0.885, forward_time=0.210, loss_ctc=40.219, loss_att=27.736, acc=0.823, loss=31.481, backward_time=0.231, grad_norm=39.406, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.867e-04, train_time=1.811
+[ip-10-0-216-33:0/16] 2024-03-03 00:31:40,997 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 00:33:19,353 (trainer:762) INFO: 29epoch:train:10501-11000batch: iter_time=1.499e-04, forward_time=0.210, loss_ctc=41.129, loss_att=28.100, acc=0.825, loss=32.009, backward_time=0.228, grad_norm=37.372, clip=100.000, loss_scale=4.641e+33, optim_step_time=0.075, optim0_lr0=1.866e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 00:40:59,353 (trainer:762) INFO: 29epoch:train:11001-11500batch: iter_time=1.514e-04, forward_time=0.207, loss_ctc=42.686, loss_att=29.652, acc=0.821, loss=33.562, backward_time=0.235, grad_norm=40.965, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.865e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-03 00:48:38,280 (trainer:762) INFO: 29epoch:train:11501-12000batch: iter_time=1.502e-04, forward_time=0.209, loss_ctc=39.732, loss_att=27.309, acc=0.825, loss=31.036, backward_time=0.234, grad_norm=37.600, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.864e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 00:56:16,470 (trainer:762) INFO: 29epoch:train:12001-12500batch: iter_time=1.579e-04, forward_time=0.208, loss_ctc=41.125, loss_att=28.886, acc=0.822, loss=32.557, backward_time=0.234, grad_norm=40.483, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.863e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-03 01:03:55,722 (trainer:762) INFO: 29epoch:train:12501-13000batch: iter_time=1.534e-04, forward_time=0.206, loss_ctc=40.345, loss_att=27.467, acc=0.820, loss=31.330, backward_time=0.232, grad_norm=38.036, clip=100.000, loss_scale=3.147e+33, optim_step_time=0.075, optim0_lr0=1.862e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 01:11:38,138 (trainer:762) INFO: 29epoch:train:13001-13500batch: iter_time=1.518e-04, forward_time=0.209, loss_ctc=41.714, loss_att=28.391, acc=0.832, loss=32.388, backward_time=0.234, grad_norm=35.500, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.861e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-03 01:13:13,653 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 01:13:39,082 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 01:19:18,154 (trainer:762) INFO: 29epoch:train:13501-14000batch: iter_time=1.484e-04, forward_time=0.206, loss_ctc=39.168, loss_att=27.504, acc=0.822, loss=31.003, backward_time=0.233, grad_norm=38.452, clip=100.000, loss_scale=2.182e+33, optim_step_time=0.075, optim0_lr0=1.860e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-03 01:26:50,530 (trainer:762) INFO: 29epoch:train:14001-14500batch: iter_time=1.475e-04, forward_time=0.211, loss_ctc=40.002, loss_att=27.558, acc=0.828, loss=31.291, backward_time=0.231, grad_norm=37.301, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.859e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-03 01:34:21,570 (trainer:762) INFO: 29epoch:train:14501-15000batch: iter_time=1.469e-04, forward_time=0.209, loss_ctc=39.152, loss_att=27.052, acc=0.828, loss=30.682, backward_time=0.229, grad_norm=36.543, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.858e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-03 01:48:24,698 (trainer:361) INFO: 29epoch results: [train] iter_time=0.087, forward_time=0.208, loss_ctc=40.712, loss_att=28.098, acc=0.824, loss=31.882, backward_time=0.233, grad_norm=38.329, clip=100.000, loss_scale=3.299e+33, optim_step_time=0.075, optim0_lr0=1.873e-04, train_time=1.003, time=4 hours, 11 minutes and 0.99 seconds, total_count=435000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=38.279, cer_ctc=0.207, loss_att=29.394, acc=0.764, cer=0.216, wer=0.990, loss=32.059, time=13 minutes and 48.1 seconds, total_count=61857, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-03 01:48:34,651 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-03 01:48:34,686 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/24epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-03 01:48:34,687 (trainer:290) INFO: 30/45epoch started. Estimated time to finish: 2 days, 22 hours and 42 minutes
+[ip-10-0-216-33:0/16] 2024-03-03 01:48:34,694 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 01:49:10,398 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 01:49:18,177 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed06f2c10>)
+[ip-10-0-216-33:0/16] 2024-03-03 01:49:18,177 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-03 01:49:18,182 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 02:03:17,206 (trainer:762) INFO: 30epoch:train:1-500batch: iter_time=0.840, forward_time=0.210, loss_ctc=41.313, loss_att=28.340, acc=0.826, loss=32.232, backward_time=0.235, grad_norm=38.196, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.857e-04, train_time=1.765
+[ip-10-0-216-33:0/16] 2024-03-03 02:10:56,516 (trainer:762) INFO: 30epoch:train:501-1000batch: iter_time=1.529e-04, forward_time=0.208, loss_ctc=41.256, loss_att=28.139, acc=0.819, loss=32.074, backward_time=0.240, grad_norm=39.072, clip=100.000, loss_scale=2.251e+33, optim_step_time=0.075, optim0_lr0=1.856e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 02:18:30,786 (trainer:762) INFO: 30epoch:train:1001-1500batch: iter_time=1.506e-04, forward_time=0.209, loss_ctc=40.954, loss_att=28.134, acc=0.818, loss=31.980, backward_time=0.232, grad_norm=41.803, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.854e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-03 02:26:08,659 (trainer:762) INFO: 30epoch:train:1501-2000batch: iter_time=1.486e-04, forward_time=0.208, loss_ctc=38.754, loss_att=26.550, acc=0.824, loss=30.211, backward_time=0.231, grad_norm=38.183, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.853e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-03 02:33:50,310 (trainer:762) INFO: 30epoch:train:2001-2500batch: iter_time=1.532e-04, forward_time=0.207, loss_ctc=42.048, loss_att=28.539, acc=0.826, loss=32.592, backward_time=0.234, grad_norm=38.210, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.852e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-03 02:41:27,286 (trainer:762) INFO: 30epoch:train:2501-3000batch: iter_time=1.530e-04, forward_time=0.208, loss_ctc=40.212, loss_att=28.100, acc=0.817, loss=31.734, backward_time=0.235, grad_norm=38.410, clip=100.000, loss_scale=4.502e+33, optim_step_time=0.075, optim0_lr0=1.851e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 02:43:19,218 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 02:49:04,265 (trainer:762) INFO: 30epoch:train:3001-3500batch: iter_time=1.494e-04, forward_time=0.209, loss_ctc=41.643, loss_att=28.824, acc=0.821, loss=32.670, backward_time=0.237, grad_norm=38.326, clip=100.000, loss_scale=3.220e+33, optim_step_time=0.075, optim0_lr0=1.850e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 02:56:49,170 (trainer:762) INFO: 30epoch:train:3501-4000batch: iter_time=1.493e-04, forward_time=0.206, loss_ctc=42.214, loss_att=29.037, acc=0.826, loss=32.990, backward_time=0.230, grad_norm=38.155, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.849e-04, train_time=0.930
+[ip-10-0-216-33:0/16] 2024-03-03 03:04:28,718 (trainer:762) INFO: 30epoch:train:4001-4500batch: iter_time=1.518e-04, forward_time=0.209, loss_ctc=42.681, loss_att=29.105, acc=0.826, loss=33.178, backward_time=0.234, grad_norm=37.125, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.848e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-03 03:12:04,629 (trainer:762) INFO: 30epoch:train:4501-5000batch: iter_time=1.503e-04, forward_time=0.208, loss_ctc=42.075, loss_att=28.626, acc=0.820, loss=32.661, backward_time=0.233, grad_norm=39.715, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.847e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 03:12:10,027 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 03:12:46,380 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 03:12:54,355 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0327910>)
+[ip-10-0-216-33:0/16] 2024-03-03 03:12:54,355 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-03 03:12:54,360 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 03:27:18,459 (trainer:762) INFO: 30epoch:train:5001-5500batch: iter_time=0.904, forward_time=0.208, loss_ctc=40.600, loss_att=27.932, acc=0.828, loss=31.732, backward_time=0.232, grad_norm=37.070, clip=100.000, loss_scale=4.564e+33, optim_step_time=0.075, optim0_lr0=1.846e-04, train_time=1.827
+[ip-10-0-216-33:0/16] 2024-03-03 03:34:52,032 (trainer:762) INFO: 30epoch:train:5501-6000batch: iter_time=1.442e-04, forward_time=0.207, loss_ctc=40.817, loss_att=27.860, acc=0.821, loss=31.747, backward_time=0.227, grad_norm=38.221, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.845e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-03 03:42:31,550 (trainer:762) INFO: 30epoch:train:6001-6500batch: iter_time=1.447e-04, forward_time=0.208, loss_ctc=40.479, loss_att=27.864, acc=0.820, loss=31.648, backward_time=0.240, grad_norm=39.147, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.844e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-03 03:50:05,915 (trainer:762) INFO: 30epoch:train:6501-7000batch: iter_time=1.441e-04, forward_time=0.206, loss_ctc=38.380, loss_att=26.473, acc=0.824, loss=30.045, backward_time=0.234, grad_norm=38.637, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.843e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-03 03:57:42,845 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 03:57:42,875 (trainer:762) INFO: 30epoch:train:7001-7500batch: iter_time=1.459e-04, forward_time=0.208, loss_ctc=41.606, loss_att=28.350, acc=0.826, loss=32.327, backward_time=0.235, grad_norm=36.766, clip=100.000, loss_scale=9.126e+33, optim_step_time=0.075, optim0_lr0=1.842e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 04:05:13,048 (trainer:762) INFO: 30epoch:train:7501-8000batch: iter_time=1.450e-04, forward_time=0.208, loss_ctc=40.141, loss_att=28.020, acc=0.818, loss=31.656, backward_time=0.237, grad_norm=37.561, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.841e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-03 04:12:48,238 (trainer:762) INFO: 30epoch:train:8001-8500batch: iter_time=1.483e-04, forward_time=0.210, loss_ctc=41.549, loss_att=28.787, acc=0.821, loss=32.615, backward_time=0.229, grad_norm=37.902, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.840e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-03 04:18:22,486 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 04:20:21,605 (trainer:762) INFO: 30epoch:train:8501-9000batch: iter_time=1.457e-04, forward_time=0.211, loss_ctc=41.699, loss_att=28.811, acc=0.827, loss=32.678, backward_time=0.233, grad_norm=37.298, clip=100.000, loss_scale=4.516e+33, optim_step_time=0.075, optim0_lr0=1.839e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-03 04:27:54,525 (trainer:762) INFO: 30epoch:train:9001-9500batch: iter_time=1.440e-04, forward_time=0.209, loss_ctc=42.344, loss_att=28.830, acc=0.827, loss=32.885, backward_time=0.231, grad_norm=36.375, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.838e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-03 04:31:46,580 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 04:35:35,865 (trainer:762) INFO: 30epoch:train:9501-10000batch: iter_time=1.423e-04, forward_time=0.209, loss_ctc=41.706, loss_att=28.370, acc=0.822, loss=32.371, backward_time=0.239, grad_norm=39.958, clip=100.000, loss_scale=1.956e+33, optim_step_time=0.075, optim0_lr0=1.837e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-03 04:35:41,013 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 04:36:18,199 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 04:36:26,161 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5f0ce87070>)
+[ip-10-0-216-33:0/16] 2024-03-03 04:36:26,161 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-03 04:36:26,166 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 04:51:14,901 (trainer:762) INFO: 30epoch:train:10001-10500batch: iter_time=0.885, forward_time=0.211, loss_ctc=40.596, loss_att=27.862, acc=0.828, loss=31.682, backward_time=0.236, grad_norm=36.706, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.836e-04, train_time=1.878
+[ip-10-0-216-33:0/16] 2024-03-03 04:58:50,792 (trainer:762) INFO: 30epoch:train:10501-11000batch: iter_time=1.462e-04, forward_time=0.207, loss_ctc=40.700, loss_att=27.769, acc=0.822, loss=31.649, backward_time=0.238, grad_norm=38.218, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.835e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 05:06:23,474 (trainer:762) INFO: 30epoch:train:11001-11500batch: iter_time=1.472e-04, forward_time=0.209, loss_ctc=40.511, loss_att=27.944, acc=0.819, loss=31.714, backward_time=0.234, grad_norm=39.051, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.834e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-03 05:13:59,188 (trainer:762) INFO: 30epoch:train:11501-12000batch: iter_time=1.490e-04, forward_time=0.206, loss_ctc=38.356, loss_att=26.418, acc=0.824, loss=30.000, backward_time=0.236, grad_norm=38.283, clip=100.000, loss_scale=1.937e+33, optim_step_time=0.075, optim0_lr0=1.833e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-03 05:21:35,492 (trainer:762) INFO: 30epoch:train:12001-12500batch: iter_time=1.478e-04, forward_time=0.209, loss_ctc=41.527, loss_att=28.303, acc=0.827, loss=32.270, backward_time=0.226, grad_norm=37.355, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.832e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 05:29:11,466 (trainer:762) INFO: 30epoch:train:12501-13000batch: iter_time=1.483e-04, forward_time=0.208, loss_ctc=39.872, loss_att=27.901, acc=0.818, loss=31.492, backward_time=0.235, grad_norm=52.096, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.076, optim0_lr0=1.831e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 05:36:45,983 (trainer:762) INFO: 30epoch:train:13001-13500batch: iter_time=1.504e-04, forward_time=0.208, loss_ctc=41.480, loss_att=28.737, acc=0.821, loss=32.560, backward_time=0.238, grad_norm=39.214, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.829e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-03 05:44:23,314 (trainer:762) INFO: 30epoch:train:13501-14000batch: iter_time=1.469e-04, forward_time=0.210, loss_ctc=41.505, loss_att=28.654, acc=0.828, loss=32.509, backward_time=0.235, grad_norm=37.506, clip=100.000, loss_scale=3.873e+33, optim_step_time=0.075, optim0_lr0=1.828e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 05:51:54,841 (trainer:762) INFO: 30epoch:train:14001-14500batch: iter_time=1.485e-04, forward_time=0.209, loss_ctc=42.279, loss_att=28.886, acc=0.827, loss=32.904, backward_time=0.239, grad_norm=37.177, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.827e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-03 05:56:00,071 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 05:59:28,687 (trainer:762) INFO: 30epoch:train:14501-15000batch: iter_time=1.456e-04, forward_time=0.210, loss_ctc=41.300, loss_att=28.260, acc=0.823, loss=32.172, backward_time=0.231, grad_norm=39.646, clip=100.000, loss_scale=3.990e+33, optim_step_time=0.075, optim0_lr0=1.826e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-03 06:13:56,928 (trainer:361) INFO: 30epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=41.020, loss_att=28.181, acc=0.823, loss=32.033, backward_time=0.234, grad_norm=38.713, clip=100.000, loss_scale=3.408e+33, optim_step_time=0.075, optim0_lr0=1.841e-04, train_time=1.003, time=4 hours, 11 minutes and 11.24 seconds, total_count=450000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=37.104, cer_ctc=0.203, loss_att=28.175, acc=0.768, cer=0.234, wer=0.989, loss=30.854, time=14 minutes and 10.66 seconds, total_count=63990, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-03 06:14:06,744 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-03 06:14:06,793 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/25epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-03 06:14:06,794 (trainer:290) INFO: 31/45epoch started. Estimated time to finish: 2 days, 18 hours and 17 minutes
+[ip-10-0-216-33:0/16] 2024-03-03 06:14:06,802 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 06:14:42,717 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 06:14:50,521 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf23ce50>)
+[ip-10-0-216-33:0/16] 2024-03-03 06:14:50,522 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-03 06:14:50,526 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 06:28:56,913 (trainer:762) INFO: 31epoch:train:1-500batch: iter_time=0.845, forward_time=0.208, loss_ctc=42.269, loss_att=28.605, acc=0.820, loss=32.704, backward_time=0.232, grad_norm=39.446, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.825e-04, train_time=1.780
+[ip-10-0-216-33:0/16] 2024-03-03 06:36:39,722 (trainer:762) INFO: 31epoch:train:501-1000batch: iter_time=1.477e-04, forward_time=0.209, loss_ctc=42.524, loss_att=28.792, acc=0.823, loss=32.911, backward_time=0.240, grad_norm=38.575, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.824e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-03 06:44:15,604 (trainer:762) INFO: 31epoch:train:1001-1500batch: iter_time=1.466e-04, forward_time=0.209, loss_ctc=40.872, loss_att=28.563, acc=0.822, loss=32.256, backward_time=0.233, grad_norm=40.638, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.823e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 06:51:48,617 (trainer:762) INFO: 31epoch:train:1501-2000batch: iter_time=1.459e-04, forward_time=0.210, loss_ctc=43.299, loss_att=29.689, acc=0.823, loss=33.772, backward_time=0.229, grad_norm=57.345, clip=100.000, loss_scale=3.796e+33, optim_step_time=0.075, optim0_lr0=1.822e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-03 06:52:02,079 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 06:59:28,190 (trainer:762) INFO: 31epoch:train:2001-2500batch: iter_time=1.463e-04, forward_time=0.208, loss_ctc=41.613, loss_att=28.451, acc=0.823, loss=32.400, backward_time=0.238, grad_norm=41.343, clip=100.000, loss_scale=2.669e+33, optim_step_time=0.075, optim0_lr0=1.821e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-03 07:07:10,678 (trainer:762) INFO: 31epoch:train:2501-3000batch: iter_time=1.444e-04, forward_time=0.207, loss_ctc=41.194, loss_att=27.845, acc=0.826, loss=31.850, backward_time=0.235, grad_norm=38.151, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.820e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-03 07:14:48,052 (trainer:762) INFO: 31epoch:train:3001-3500batch: iter_time=1.481e-04, forward_time=0.208, loss_ctc=41.509, loss_att=27.986, acc=0.825, loss=32.042, backward_time=0.233, grad_norm=39.179, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.819e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 07:22:26,010 (trainer:762) INFO: 31epoch:train:3501-4000batch: iter_time=1.466e-04, forward_time=0.208, loss_ctc=40.262, loss_att=27.868, acc=0.826, loss=31.586, backward_time=0.233, grad_norm=37.011, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.818e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-03 07:30:04,325 (trainer:762) INFO: 31epoch:train:4001-4500batch: iter_time=1.450e-04, forward_time=0.207, loss_ctc=39.813, loss_att=27.888, acc=0.819, loss=31.466, backward_time=0.232, grad_norm=39.174, clip=100.000, loss_scale=5.114e+33, optim_step_time=0.075, optim0_lr0=1.817e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-03 07:37:42,331 (trainer:762) INFO: 31epoch:train:4501-5000batch: iter_time=1.475e-04, forward_time=0.208, loss_ctc=39.743, loss_att=27.032, acc=0.823, loss=30.845, backward_time=0.240, grad_norm=37.544, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.816e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-03 07:37:46,937 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 07:38:23,709 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 07:38:31,734 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ece1edfa0>)
+[ip-10-0-216-33:0/16] 2024-03-03 07:38:31,735 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-03 07:38:31,740 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 07:52:01,289 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 07:53:00,344 (trainer:762) INFO: 31epoch:train:5001-5500batch: iter_time=0.890, forward_time=0.207, loss_ctc=41.992, loss_att=28.857, acc=0.820, loss=32.798, backward_time=0.236, grad_norm=40.332, clip=100.000, loss_scale=4.854e+33, optim_step_time=0.075, optim0_lr0=1.815e-04, train_time=1.836
+[ip-10-0-216-33:0/16] 2024-03-03 08:00:40,626 (trainer:762) INFO: 31epoch:train:5501-6000batch: iter_time=1.461e-04, forward_time=0.207, loss_ctc=42.007, loss_att=28.562, acc=0.824, loss=32.595, backward_time=0.235, grad_norm=39.122, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.814e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-03 08:08:09,936 (trainer:762) INFO: 31epoch:train:6001-6500batch: iter_time=1.491e-04, forward_time=0.209, loss_ctc=40.649, loss_att=28.421, acc=0.823, loss=32.089, backward_time=0.232, grad_norm=40.138, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.813e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-03-03 08:15:50,522 (trainer:762) INFO: 31epoch:train:6501-7000batch: iter_time=1.441e-04, forward_time=0.208, loss_ctc=43.118, loss_att=29.529, acc=0.825, loss=33.606, backward_time=0.233, grad_norm=40.858, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.812e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-03 08:23:21,486 (trainer:762) INFO: 31epoch:train:7001-7500batch: iter_time=1.454e-04, forward_time=0.208, loss_ctc=41.495, loss_att=28.448, acc=0.823, loss=32.362, backward_time=0.231, grad_norm=39.681, clip=100.000, loss_scale=2.934e+33, optim_step_time=0.075, optim0_lr0=1.811e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-03 08:28:02,671 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 08:31:00,436 (trainer:762) INFO: 31epoch:train:7501-8000batch: iter_time=1.482e-04, forward_time=0.209, loss_ctc=41.003, loss_att=27.754, acc=0.827, loss=31.729, backward_time=0.238, grad_norm=37.941, clip=100.000, loss_scale=4.167e+33, optim_step_time=0.075, optim0_lr0=1.810e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 08:38:32,864 (trainer:762) INFO: 31epoch:train:8001-8500batch: iter_time=1.453e-04, forward_time=0.207, loss_ctc=40.954, loss_att=27.688, acc=0.826, loss=31.668, backward_time=0.232, grad_norm=39.204, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.809e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-03 08:46:06,807 (trainer:762) INFO: 31epoch:train:8501-9000batch: iter_time=1.439e-04, forward_time=0.208, loss_ctc=40.197, loss_att=27.861, acc=0.826, loss=31.561, backward_time=0.237, grad_norm=37.131, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.808e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-03 08:53:44,180 (trainer:762) INFO: 31epoch:train:9001-9500batch: iter_time=1.439e-04, forward_time=0.207, loss_ctc=39.771, loss_att=27.835, acc=0.819, loss=31.416, backward_time=0.233, grad_norm=40.043, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.807e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-03 08:59:00,218 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 09:01:16,990 (trainer:762) INFO: 31epoch:train:9501-10000batch: iter_time=1.428e-04, forward_time=0.208, loss_ctc=39.423, loss_att=26.780, acc=0.824, loss=30.573, backward_time=0.232, grad_norm=37.628, clip=100.000, loss_scale=2.841e+33, optim_step_time=0.075, optim0_lr0=1.806e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-03 09:01:21,436 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 09:01:58,734 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 09:02:06,745 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e77271d00>)
+[ip-10-0-216-33:0/16] 2024-03-03 09:02:06,745 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-03 09:02:06,750 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 09:16:29,909 (trainer:762) INFO: 31epoch:train:10001-10500batch: iter_time=0.904, forward_time=0.208, loss_ctc=41.689, loss_att=28.658, acc=0.821, loss=32.567, backward_time=0.232, grad_norm=41.050, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.805e-04, train_time=1.826
+[ip-10-0-216-33:0/16] 2024-03-03 09:24:13,338 (trainer:762) INFO: 31epoch:train:10501-11000batch: iter_time=1.475e-04, forward_time=0.208, loss_ctc=42.072, loss_att=28.620, acc=0.824, loss=32.655, backward_time=0.231, grad_norm=38.183, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.805e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-03-03 09:31:49,395 (trainer:762) INFO: 31epoch:train:11001-11500batch: iter_time=1.480e-04, forward_time=0.209, loss_ctc=40.724, loss_att=28.503, acc=0.823, loss=32.169, backward_time=0.231, grad_norm=41.403, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.804e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 09:39:12,637 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 09:39:29,992 (trainer:762) INFO: 31epoch:train:11501-12000batch: iter_time=1.444e-04, forward_time=0.208, loss_ctc=42.645, loss_att=29.309, acc=0.825, loss=33.310, backward_time=0.232, grad_norm=63.531, clip=100.000, loss_scale=3.267e+33, optim_step_time=0.075, optim0_lr0=1.803e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-03 09:47:08,690 (trainer:762) INFO: 31epoch:train:12001-12500batch: iter_time=1.459e-04, forward_time=0.209, loss_ctc=41.121, loss_att=28.245, acc=0.824, loss=32.107, backward_time=0.235, grad_norm=40.139, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.802e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-03 09:54:44,707 (trainer:762) INFO: 31epoch:train:12501-13000batch: iter_time=1.502e-04, forward_time=0.209, loss_ctc=40.833, loss_att=27.606, acc=0.828, loss=31.574, backward_time=0.237, grad_norm=37.213, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.076, optim0_lr0=1.801e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 10:02:17,027 (trainer:762) INFO: 31epoch:train:13001-13500batch: iter_time=1.505e-04, forward_time=0.210, loss_ctc=40.940, loss_att=27.658, acc=0.826, loss=31.642, backward_time=0.238, grad_norm=38.486, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.800e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-03 10:09:50,634 (trainer:762) INFO: 31epoch:train:13501-14000batch: iter_time=1.477e-04, forward_time=0.209, loss_ctc=39.922, loss_att=27.685, acc=0.827, loss=31.356, backward_time=0.235, grad_norm=36.730, clip=100.000, loss_scale=2.695e+33, optim_step_time=0.075, optim0_lr0=1.799e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-03 10:12:13,173 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 10:17:24,068 (trainer:762) INFO: 31epoch:train:14001-14500batch: iter_time=1.445e-04, forward_time=0.210, loss_ctc=39.287, loss_att=27.684, acc=0.820, loss=31.165, backward_time=0.238, grad_norm=39.681, clip=100.000, loss_scale=3.403e+33, optim_step_time=0.075, optim0_lr0=1.798e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-03 10:24:54,986 (trainer:762) INFO: 31epoch:train:14501-15000batch: iter_time=1.498e-04, forward_time=0.209, loss_ctc=39.494, loss_att=26.941, acc=0.823, loss=30.707, backward_time=0.231, grad_norm=37.742, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.797e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-03 10:39:01,975 (trainer:361) INFO: 31epoch results: [train] iter_time=0.088, forward_time=0.208, loss_ctc=41.081, loss_att=28.179, acc=0.824, loss=32.049, backward_time=0.234, grad_norm=40.487, clip=100.000, loss_scale=3.008e+33, optim_step_time=0.075, optim0_lr0=1.811e-04, train_time=1.003, time=4 hours, 11 minutes and 2.64 seconds, total_count=465000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=37.380, cer_ctc=0.204, loss_att=29.390, acc=0.762, cer=0.228, wer=0.990, loss=31.787, time=13 minutes and 52.22 seconds, total_count=66123, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-03 10:39:11,741 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-03 10:39:11,791 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/19epoch.pth, exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/26epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-03 10:39:11,791 (trainer:290) INFO: 32/45epoch started. Estimated time to finish: 2 days, 13 hours and 52 minutes
+[ip-10-0-216-33:0/16] 2024-03-03 10:39:11,798 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 10:39:47,827 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 10:39:56,124 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf8a3b80>)
+[ip-10-0-216-33:0/16] 2024-03-03 10:39:56,125 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-03 10:39:56,129 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 10:53:54,565 (trainer:762) INFO: 32epoch:train:1-500batch: iter_time=0.835, forward_time=0.210, loss_ctc=43.046, loss_att=29.726, acc=0.821, loss=33.722, backward_time=0.234, grad_norm=42.162, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.796e-04, train_time=1.765
+[ip-10-0-216-33:0/16] 2024-03-03 11:01:37,546 (trainer:762) INFO: 32epoch:train:501-1000batch: iter_time=1.461e-04, forward_time=0.206, loss_ctc=40.153, loss_att=27.299, acc=0.826, loss=31.156, backward_time=0.230, grad_norm=39.054, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.795e-04, train_time=0.926
+[ip-10-0-216-33:0/16] 2024-03-03 11:06:41,325 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 11:09:16,965 (trainer:762) INFO: 32epoch:train:1001-1500batch: iter_time=1.481e-04, forward_time=0.209, loss_ctc=41.503, loss_att=28.397, acc=0.820, loss=32.329, backward_time=0.230, grad_norm=38.902, clip=100.000, loss_scale=3.496e+33, optim_step_time=0.075, optim0_lr0=1.794e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-03 11:16:53,989 (trainer:762) INFO: 32epoch:train:1501-2000batch: iter_time=1.480e-04, forward_time=0.210, loss_ctc=41.385, loss_att=28.354, acc=0.827, loss=32.264, backward_time=0.230, grad_norm=39.036, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.793e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 11:24:31,549 (trainer:762) INFO: 32epoch:train:2001-2500batch: iter_time=1.474e-04, forward_time=0.207, loss_ctc=39.289, loss_att=26.751, acc=0.823, loss=30.512, backward_time=0.237, grad_norm=40.407, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.792e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-03 11:32:10,486 (trainer:762) INFO: 32epoch:train:2501-3000batch: iter_time=1.493e-04, forward_time=0.209, loss_ctc=40.938, loss_att=28.444, acc=0.826, loss=32.192, backward_time=0.234, grad_norm=38.480, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.791e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 11:39:21,179 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 11:39:51,326 (trainer:762) INFO: 32epoch:train:3001-3500batch: iter_time=1.475e-04, forward_time=0.209, loss_ctc=39.596, loss_att=27.670, acc=0.826, loss=31.248, backward_time=0.235, grad_norm=39.719, clip=100.000, loss_scale=3.309e+33, optim_step_time=0.075, optim0_lr0=1.790e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-03 11:47:26,913 (trainer:762) INFO: 32epoch:train:3501-4000batch: iter_time=1.508e-04, forward_time=0.210, loss_ctc=43.119, loss_att=29.599, acc=0.828, loss=33.655, backward_time=0.232, grad_norm=39.290, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.789e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-03 11:51:40,436 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 11:55:06,588 (trainer:762) INFO: 32epoch:train:4001-4500batch: iter_time=1.519e-04, forward_time=0.209, loss_ctc=42.989, loss_att=29.666, acc=0.825, loss=33.663, backward_time=0.234, grad_norm=41.599, clip=100.000, loss_scale=2.008e+33, optim_step_time=0.075, optim0_lr0=1.788e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-03 12:02:45,512 (trainer:762) INFO: 32epoch:train:4501-5000batch: iter_time=1.476e-04, forward_time=0.209, loss_ctc=39.931, loss_att=27.283, acc=0.825, loss=31.077, backward_time=0.232, grad_norm=40.842, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.787e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 12:02:50,553 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 12:03:26,820 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 12:03:35,464 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf730d00>)
+[ip-10-0-216-33:0/16] 2024-03-03 12:03:35,465 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-03 12:03:35,469 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 12:17:58,321 (trainer:762) INFO: 32epoch:train:5001-5500batch: iter_time=0.886, forward_time=0.207, loss_ctc=42.712, loss_att=29.325, acc=0.822, loss=33.341, backward_time=0.230, grad_norm=41.867, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.786e-04, train_time=1.825
+[ip-10-0-216-33:0/16] 2024-03-03 12:25:37,480 (trainer:762) INFO: 32epoch:train:5501-6000batch: iter_time=1.480e-04, forward_time=0.210, loss_ctc=40.032, loss_att=27.244, acc=0.826, loss=31.081, backward_time=0.237, grad_norm=38.595, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.785e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 12:31:33,431 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 12:33:10,116 (trainer:762) INFO: 32epoch:train:6001-6500batch: iter_time=1.485e-04, forward_time=0.209, loss_ctc=40.781, loss_att=28.041, acc=0.821, loss=31.863, backward_time=0.235, grad_norm=38.413, clip=100.000, loss_scale=1.608e+33, optim_step_time=0.075, optim0_lr0=1.784e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-03 12:40:50,954 (trainer:762) INFO: 32epoch:train:6501-7000batch: iter_time=1.465e-04, forward_time=0.207, loss_ctc=40.784, loss_att=27.995, acc=0.829, loss=31.832, backward_time=0.233, grad_norm=37.484, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.783e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-03 12:48:27,764 (trainer:762) INFO: 32epoch:train:7001-7500batch: iter_time=1.466e-04, forward_time=0.208, loss_ctc=39.155, loss_att=26.679, acc=0.823, loss=30.422, backward_time=0.232, grad_norm=40.399, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.782e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-03 12:56:10,348 (trainer:762) INFO: 32epoch:train:7501-8000batch: iter_time=1.490e-04, forward_time=0.206, loss_ctc=40.611, loss_att=28.273, acc=0.827, loss=31.974, backward_time=0.235, grad_norm=37.705, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.781e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-03 13:03:49,671 (trainer:762) INFO: 32epoch:train:8001-8500batch: iter_time=1.443e-04, forward_time=0.209, loss_ctc=39.202, loss_att=27.502, acc=0.827, loss=31.012, backward_time=0.232, grad_norm=39.641, clip=100.000, loss_scale=1.573e+33, optim_step_time=0.075, optim0_lr0=1.781e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 13:06:56,009 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 13:11:21,724 (trainer:762) INFO: 32epoch:train:8501-9000batch: iter_time=1.508e-04, forward_time=0.209, loss_ctc=42.714, loss_att=29.436, acc=0.828, loss=33.419, backward_time=0.234, grad_norm=39.996, clip=100.000, loss_scale=1.831e+33, optim_step_time=0.075, optim0_lr0=1.780e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-03 13:18:57,089 (trainer:762) INFO: 32epoch:train:9001-9500batch: iter_time=1.490e-04, forward_time=0.210, loss_ctc=42.577, loss_att=29.470, acc=0.825, loss=33.402, backward_time=0.230, grad_norm=40.813, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.076, optim0_lr0=1.779e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-03 13:26:28,155 (trainer:762) INFO: 32epoch:train:9501-10000batch: iter_time=1.430e-04, forward_time=0.209, loss_ctc=39.624, loss_att=27.019, acc=0.826, loss=30.800, backward_time=0.231, grad_norm=38.215, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.778e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-03 13:26:34,382 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 13:27:11,764 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 13:27:20,891 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e8b18bd60>)
+[ip-10-0-216-33:0/16] 2024-03-03 13:27:20,891 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-03 13:27:20,895 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 13:41:41,765 (trainer:762) INFO: 32epoch:train:10001-10500batch: iter_time=0.879, forward_time=0.210, loss_ctc=42.614, loss_att=29.274, acc=0.823, loss=33.276, backward_time=0.233, grad_norm=41.873, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.777e-04, train_time=1.827
+[ip-10-0-216-33:0/16] 2024-03-03 13:49:16,434 (trainer:762) INFO: 32epoch:train:10501-11000batch: iter_time=1.472e-04, forward_time=0.210, loss_ctc=39.900, loss_att=27.183, acc=0.826, loss=30.998, backward_time=0.234, grad_norm=39.653, clip=100.000, loss_scale=2.061e+33, optim_step_time=0.075, optim0_lr0=1.776e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-03 13:56:53,497 (trainer:762) INFO: 32epoch:train:11001-11500batch: iter_time=1.475e-04, forward_time=0.208, loss_ctc=40.802, loss_att=28.070, acc=0.822, loss=31.890, backward_time=0.233, grad_norm=38.655, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.775e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 14:04:29,364 (trainer:762) INFO: 32epoch:train:11501-12000batch: iter_time=1.463e-04, forward_time=0.209, loss_ctc=40.576, loss_att=27.884, acc=0.829, loss=31.691, backward_time=0.233, grad_norm=38.712, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.774e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-03 14:12:01,300 (trainer:762) INFO: 32epoch:train:12001-12500batch: iter_time=1.529e-04, forward_time=0.208, loss_ctc=38.882, loss_att=26.582, acc=0.824, loss=30.272, backward_time=0.232, grad_norm=39.376, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.773e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-03 14:15:21,580 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 14:19:32,904 (trainer:762) INFO: 32epoch:train:12501-13000batch: iter_time=1.456e-04, forward_time=0.211, loss_ctc=40.547, loss_att=28.250, acc=0.827, loss=31.939, backward_time=0.231, grad_norm=38.182, clip=100.000, loss_scale=2.664e+33, optim_step_time=0.075, optim0_lr0=1.772e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-03 14:27:06,243 (trainer:762) INFO: 32epoch:train:13001-13500batch: iter_time=1.464e-04, forward_time=0.210, loss_ctc=39.046, loss_att=27.422, acc=0.827, loss=30.909, backward_time=0.234, grad_norm=42.302, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.771e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-03 14:34:41,501 (trainer:762) INFO: 32epoch:train:13501-14000batch: iter_time=1.427e-04, forward_time=0.210, loss_ctc=42.734, loss_att=29.525, acc=0.828, loss=33.488, backward_time=0.233, grad_norm=39.680, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.770e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-03 14:39:49,305 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 14:42:23,554 (trainer:762) INFO: 32epoch:train:14001-14500batch: iter_time=1.445e-04, forward_time=0.207, loss_ctc=42.393, loss_att=29.285, acc=0.826, loss=33.218, backward_time=0.231, grad_norm=39.554, clip=100.000, loss_scale=2.157e+33, optim_step_time=0.075, optim0_lr0=1.769e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-03 14:49:53,274 (trainer:762) INFO: 32epoch:train:14501-15000batch: iter_time=1.436e-04, forward_time=0.210, loss_ctc=39.823, loss_att=27.112, acc=0.826, loss=30.925, backward_time=0.228, grad_norm=38.523, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.768e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-03 15:04:04,981 (trainer:361) INFO: 32epoch results: [train] iter_time=0.087, forward_time=0.209, loss_ctc=40.915, loss_att=28.159, acc=0.825, loss=31.986, backward_time=0.233, grad_norm=39.638, clip=100.000, loss_scale=2.075e+33, optim_step_time=0.075, optim0_lr0=1.782e-04, train_time=1.003, time=4 hours, 11 minutes and 0.77 seconds, total_count=480000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=36.626, cer_ctc=0.199, loss_att=28.425, acc=0.767, cer=0.223, wer=0.991, loss=30.886, time=13 minutes and 52.11 seconds, total_count=68256, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-03 15:04:14,921 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-03 15:04:14,975 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/22epoch.pth, exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/27epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-03 15:04:14,976 (trainer:290) INFO: 33/45epoch started. Estimated time to finish: 2 days, 9 hours and 27 minutes
+[ip-10-0-216-33:0/16] 2024-03-03 15:04:14,983 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 15:04:51,026 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 15:04:59,256 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfe39910>)
+[ip-10-0-216-33:0/16] 2024-03-03 15:04:59,256 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-03 15:04:59,260 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 15:18:53,719 (trainer:762) INFO: 33epoch:train:1-500batch: iter_time=0.844, forward_time=0.210, loss_ctc=42.744, loss_att=29.097, acc=0.822, loss=33.191, backward_time=0.235, grad_norm=41.769, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.768e-04, train_time=1.757
+[ip-10-0-216-33:0/16] 2024-03-03 15:26:33,036 (trainer:762) INFO: 33epoch:train:501-1000batch: iter_time=1.453e-04, forward_time=0.208, loss_ctc=42.111, loss_att=28.531, acc=0.822, loss=32.605, backward_time=0.237, grad_norm=40.171, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.767e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 15:34:10,132 (trainer:762) INFO: 33epoch:train:1001-1500batch: iter_time=1.492e-04, forward_time=0.210, loss_ctc=40.221, loss_att=27.917, acc=0.824, loss=31.608, backward_time=0.238, grad_norm=40.364, clip=100.000, loss_scale=1.737e+33, optim_step_time=0.075, optim0_lr0=1.766e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 15:36:13,830 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 15:41:53,533 (trainer:762) INFO: 33epoch:train:1501-2000batch: iter_time=1.484e-04, forward_time=0.209, loss_ctc=40.875, loss_att=28.060, acc=0.828, loss=31.904, backward_time=0.236, grad_norm=40.042, clip=100.000, loss_scale=1.644e+33, optim_step_time=0.075, optim0_lr0=1.765e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-03-03 15:49:28,341 (trainer:762) INFO: 33epoch:train:2001-2500batch: iter_time=1.473e-04, forward_time=0.209, loss_ctc=40.983, loss_att=28.325, acc=0.818, loss=32.123, backward_time=0.233, grad_norm=41.387, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.764e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-03 15:57:02,198 (trainer:762) INFO: 33epoch:train:2501-3000batch: iter_time=1.471e-04, forward_time=0.210, loss_ctc=39.172, loss_att=26.796, acc=0.822, loss=30.509, backward_time=0.234, grad_norm=38.852, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.763e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-03 16:04:40,870 (trainer:762) INFO: 33epoch:train:3001-3500batch: iter_time=1.474e-04, forward_time=0.210, loss_ctc=42.983, loss_att=29.196, acc=0.823, loss=33.332, backward_time=0.235, grad_norm=46.472, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.762e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-03 16:12:20,804 (trainer:762) INFO: 33epoch:train:3501-4000batch: iter_time=1.501e-04, forward_time=0.206, loss_ctc=39.218, loss_att=27.593, acc=0.824, loss=31.080, backward_time=0.235, grad_norm=39.254, clip=100.000, loss_scale=2.248e+33, optim_step_time=0.075, optim0_lr0=1.761e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-03 16:18:50,602 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 16:20:00,621 (trainer:762) INFO: 33epoch:train:4001-4500batch: iter_time=1.454e-04, forward_time=0.209, loss_ctc=40.459, loss_att=28.005, acc=0.830, loss=31.741, backward_time=0.232, grad_norm=40.473, clip=100.000, loss_scale=2.393e+33, optim_step_time=0.076, optim0_lr0=1.760e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-03 16:27:36,605 (trainer:762) INFO: 33epoch:train:4501-5000batch: iter_time=1.450e-04, forward_time=0.209, loss_ctc=40.105, loss_att=27.844, acc=0.814, loss=31.522, backward_time=0.228, grad_norm=42.147, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.759e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 16:27:42,631 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 16:28:18,758 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 16:28:27,619 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf378e80>)
+[ip-10-0-216-33:0/16] 2024-03-03 16:28:27,620 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-03 16:28:27,624 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 16:42:51,307 (trainer:762) INFO: 33epoch:train:5001-5500batch: iter_time=0.866, forward_time=0.210, loss_ctc=42.306, loss_att=29.067, acc=0.823, loss=33.038, backward_time=0.231, grad_norm=41.537, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.758e-04, train_time=1.829
+[ip-10-0-216-33:0/16] 2024-03-03 16:50:33,257 (trainer:762) INFO: 33epoch:train:5501-6000batch: iter_time=1.453e-04, forward_time=0.208, loss_ctc=41.630, loss_att=28.210, acc=0.824, loss=32.236, backward_time=0.239, grad_norm=40.480, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.757e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-03 16:58:06,170 (trainer:762) INFO: 33epoch:train:6001-6500batch: iter_time=1.433e-04, forward_time=0.209, loss_ctc=39.956, loss_att=27.959, acc=0.824, loss=31.558, backward_time=0.233, grad_norm=39.125, clip=100.000, loss_scale=1.501e+33, optim_step_time=0.075, optim0_lr0=1.757e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-03 17:05:39,566 (trainer:762) INFO: 33epoch:train:6501-7000batch: iter_time=1.436e-04, forward_time=0.209, loss_ctc=40.696, loss_att=27.964, acc=0.829, loss=31.784, backward_time=0.231, grad_norm=39.028, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.756e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-03 17:13:11,145 (trainer:762) INFO: 33epoch:train:7001-7500batch: iter_time=1.437e-04, forward_time=0.209, loss_ctc=40.793, loss_att=28.268, acc=0.819, loss=32.025, backward_time=0.233, grad_norm=41.215, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.755e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-03 17:20:48,205 (trainer:762) INFO: 33epoch:train:7501-8000batch: iter_time=1.446e-04, forward_time=0.207, loss_ctc=38.964, loss_att=26.679, acc=0.822, loss=30.364, backward_time=0.234, grad_norm=38.362, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.754e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 17:28:27,658 (trainer:762) INFO: 33epoch:train:8001-8500batch: iter_time=1.437e-04, forward_time=0.210, loss_ctc=42.675, loss_att=29.083, acc=0.824, loss=33.160, backward_time=0.235, grad_norm=41.421, clip=100.000, loss_scale=3.001e+33, optim_step_time=0.075, optim0_lr0=1.753e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-03 17:36:02,127 (trainer:762) INFO: 33epoch:train:8501-9000batch: iter_time=1.457e-04, forward_time=0.208, loss_ctc=38.832, loss_att=27.387, acc=0.825, loss=30.820, backward_time=0.231, grad_norm=39.025, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.752e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-03 17:43:38,475 (trainer:762) INFO: 33epoch:train:9001-9500batch: iter_time=1.424e-04, forward_time=0.209, loss_ctc=40.392, loss_att=28.013, acc=0.830, loss=31.727, backward_time=0.233, grad_norm=40.656, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.751e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 17:51:07,147 (trainer:762) INFO: 33epoch:train:9501-10000batch: iter_time=1.471e-04, forward_time=0.209, loss_ctc=39.695, loss_att=27.550, acc=0.815, loss=31.194, backward_time=0.233, grad_norm=40.774, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.750e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-03-03 17:51:13,165 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 17:51:50,720 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 17:51:59,381 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0696040>)
+[ip-10-0-216-33:0/16] 2024-03-03 17:51:59,381 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-03 17:51:59,385 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 18:06:08,509 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 18:06:29,278 (trainer:762) INFO: 33epoch:train:10001-10500batch: iter_time=0.900, forward_time=0.209, loss_ctc=41.984, loss_att=28.924, acc=0.823, loss=32.842, backward_time=0.231, grad_norm=42.229, clip=100.000, loss_scale=5.754e+33, optim_step_time=0.075, optim0_lr0=1.749e-04, train_time=1.844
+[ip-10-0-216-33:0/16] 2024-03-03 18:09:49,711 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 18:14:10,558 (trainer:762) INFO: 33epoch:train:10501-11000batch: iter_time=1.481e-04, forward_time=0.210, loss_ctc=41.094, loss_att=28.044, acc=0.824, loss=31.959, backward_time=0.237, grad_norm=40.703, clip=100.000, loss_scale=3.715e+33, optim_step_time=0.075, optim0_lr0=1.748e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-03 18:21:49,809 (trainer:762) INFO: 33epoch:train:11001-11500batch: iter_time=1.487e-04, forward_time=0.209, loss_ctc=39.867, loss_att=27.755, acc=0.825, loss=31.389, backward_time=0.231, grad_norm=39.928, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.748e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 18:29:23,754 (trainer:762) INFO: 33epoch:train:11501-12000batch: iter_time=1.470e-04, forward_time=0.211, loss_ctc=40.472, loss_att=27.818, acc=0.830, loss=31.614, backward_time=0.237, grad_norm=38.902, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.747e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-03 18:37:03,052 (trainer:762) INFO: 33epoch:train:12001-12500batch: iter_time=1.491e-04, forward_time=0.208, loss_ctc=40.536, loss_att=28.154, acc=0.820, loss=31.869, backward_time=0.235, grad_norm=40.508, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.746e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-03 18:44:46,829 (trainer:762) INFO: 33epoch:train:12501-13000batch: iter_time=1.539e-04, forward_time=0.209, loss_ctc=38.912, loss_att=26.595, acc=0.823, loss=30.290, backward_time=0.235, grad_norm=38.388, clip=100.000, loss_scale=4.071e+33, optim_step_time=0.075, optim0_lr0=1.745e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-03-03 18:52:22,756 (trainer:762) INFO: 33epoch:train:13001-13500batch: iter_time=1.522e-04, forward_time=0.210, loss_ctc=42.653, loss_att=28.924, acc=0.825, loss=33.043, backward_time=0.237, grad_norm=40.972, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.744e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 18:55:19,670 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 18:59:56,649 (trainer:762) INFO: 33epoch:train:13501-14000batch: iter_time=1.480e-04, forward_time=0.210, loss_ctc=38.737, loss_att=27.324, acc=0.825, loss=30.748, backward_time=0.232, grad_norm=39.641, clip=100.000, loss_scale=3.595e+33, optim_step_time=0.075, optim0_lr0=1.743e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-03 19:07:37,258 (trainer:762) INFO: 33epoch:train:14001-14500batch: iter_time=1.501e-04, forward_time=0.209, loss_ctc=39.862, loss_att=27.686, acc=0.831, loss=31.339, backward_time=0.236, grad_norm=39.844, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.742e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-03 19:15:09,547 (trainer:762) INFO: 33epoch:train:14501-15000batch: iter_time=1.457e-04, forward_time=0.210, loss_ctc=39.328, loss_att=27.377, acc=0.816, loss=30.962, backward_time=0.227, grad_norm=42.940, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.741e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-03 19:29:33,252 (trainer:361) INFO: 33epoch results: [train] iter_time=0.087, forward_time=0.209, loss_ctc=40.608, loss_att=28.005, acc=0.823, loss=31.786, backward_time=0.234, grad_norm=40.554, clip=100.000, loss_scale=2.719e+33, optim_step_time=0.075, optim0_lr0=1.754e-04, train_time=1.003, time=4 hours, 11 minutes and 9.93 seconds, total_count=495000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=36.512, cer_ctc=0.199, loss_att=30.865, acc=0.762, cer=0.210, wer=0.988, loss=32.559, time=14 minutes and 8 seconds, total_count=70389, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-03 19:29:43,186 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-03 19:29:43,221 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/28epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-03 19:29:43,222 (trainer:290) INFO: 34/45epoch started. Estimated time to finish: 2 days, 5 hours and 2 minutes
+[ip-10-0-216-33:0/16] 2024-03-03 19:29:43,229 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 19:30:18,952 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 19:30:26,786 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ece4089a0>)
+[ip-10-0-216-33:0/16] 2024-03-03 19:30:26,786 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-03 19:30:26,791 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 19:44:26,115 (trainer:762) INFO: 34epoch:train:1-500batch: iter_time=0.834, forward_time=0.211, loss_ctc=39.809, loss_att=27.547, acc=0.822, loss=31.226, backward_time=0.239, grad_norm=41.163, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.741e-04, train_time=1.765
+[ip-10-0-216-33:0/16] 2024-03-03 19:47:45,869 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 19:52:00,629 (trainer:762) INFO: 34epoch:train:501-1000batch: iter_time=1.477e-04, forward_time=0.211, loss_ctc=42.462, loss_att=28.830, acc=0.826, loss=32.920, backward_time=0.235, grad_norm=41.132, clip=100.000, loss_scale=2.731e+33, optim_step_time=0.075, optim0_lr0=1.740e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-03 19:59:35,872 (trainer:762) INFO: 34epoch:train:1001-1500batch: iter_time=1.468e-04, forward_time=0.209, loss_ctc=41.930, loss_att=28.627, acc=0.817, loss=32.618, backward_time=0.236, grad_norm=43.495, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.739e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-03 20:07:11,719 (trainer:762) INFO: 34epoch:train:1501-2000batch: iter_time=1.471e-04, forward_time=0.209, loss_ctc=40.045, loss_att=27.434, acc=0.826, loss=31.218, backward_time=0.239, grad_norm=40.295, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.738e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-03 20:14:45,248 (trainer:762) INFO: 34epoch:train:2001-2500batch: iter_time=1.451e-04, forward_time=0.211, loss_ctc=41.022, loss_att=27.741, acc=0.827, loss=31.725, backward_time=0.234, grad_norm=40.689, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.737e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-03 20:22:20,274 (trainer:762) INFO: 34epoch:train:2501-3000batch: iter_time=1.449e-04, forward_time=0.209, loss_ctc=39.688, loss_att=27.198, acc=0.829, loss=30.945, backward_time=0.231, grad_norm=38.777, clip=100.000, loss_scale=4.050e+33, optim_step_time=0.075, optim0_lr0=1.736e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-03 20:22:36,144 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 20:29:59,797 (trainer:762) INFO: 34epoch:train:3001-3500batch: iter_time=1.479e-04, forward_time=0.209, loss_ctc=40.179, loss_att=28.077, acc=0.823, loss=31.708, backward_time=0.235, grad_norm=39.084, clip=100.000, loss_scale=2.679e+33, optim_step_time=0.075, optim0_lr0=1.735e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-03 20:37:29,927 (trainer:762) INFO: 34epoch:train:3501-4000batch: iter_time=1.473e-04, forward_time=0.209, loss_ctc=41.488, loss_att=28.521, acc=0.825, loss=32.411, backward_time=0.232, grad_norm=41.500, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.734e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-03 20:45:08,439 (trainer:762) INFO: 34epoch:train:4001-4500batch: iter_time=1.438e-04, forward_time=0.209, loss_ctc=41.205, loss_att=28.263, acc=0.825, loss=32.145, backward_time=0.227, grad_norm=40.747, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.734e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-03 20:48:40,262 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 20:52:44,961 (trainer:762) INFO: 34epoch:train:4501-5000batch: iter_time=1.454e-04, forward_time=0.209, loss_ctc=40.669, loss_att=28.064, acc=0.825, loss=31.845, backward_time=0.232, grad_norm=39.550, clip=100.000, loss_scale=1.907e+33, optim_step_time=0.075, optim0_lr0=1.733e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-03 20:52:51,137 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 20:53:27,785 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 20:53:35,750 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfe08310>)
+[ip-10-0-216-33:0/16] 2024-03-03 20:53:35,751 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-03 20:53:35,755 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 21:10:22,937 (trainer:762) INFO: 34epoch:train:5001-5500batch: iter_time=1.177, forward_time=0.208, loss_ctc=39.390, loss_att=27.277, acc=0.824, loss=30.911, backward_time=0.232, grad_norm=41.589, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.732e-04, train_time=2.116
+[ip-10-0-216-33:0/16] 2024-03-03 21:18:05,209 (trainer:762) INFO: 34epoch:train:5501-6000batch: iter_time=1.462e-04, forward_time=0.207, loss_ctc=42.130, loss_att=28.527, acc=0.828, loss=32.608, backward_time=0.232, grad_norm=40.000, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.731e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-03 21:25:47,992 (trainer:762) INFO: 34epoch:train:6001-6500batch: iter_time=1.489e-04, forward_time=0.208, loss_ctc=41.798, loss_att=28.563, acc=0.818, loss=32.533, backward_time=0.233, grad_norm=42.901, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.730e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-03 21:33:29,953 (trainer:762) INFO: 34epoch:train:6501-7000batch: iter_time=1.467e-04, forward_time=0.207, loss_ctc=39.968, loss_att=27.370, acc=0.826, loss=31.149, backward_time=0.234, grad_norm=40.647, clip=100.000, loss_scale=1.986e+33, optim_step_time=0.075, optim0_lr0=1.729e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-03 21:37:05,032 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 21:41:08,146 (trainer:762) INFO: 34epoch:train:7001-7500batch: iter_time=1.479e-04, forward_time=0.209, loss_ctc=40.719, loss_att=27.555, acc=0.827, loss=31.504, backward_time=0.236, grad_norm=40.711, clip=100.000, loss_scale=1.902e+33, optim_step_time=0.075, optim0_lr0=1.728e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-03 21:48:44,465 (trainer:762) INFO: 34epoch:train:7501-8000batch: iter_time=1.524e-04, forward_time=0.210, loss_ctc=39.396, loss_att=27.039, acc=0.830, loss=30.747, backward_time=0.237, grad_norm=39.760, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.728e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-03 21:56:30,871 (trainer:762) INFO: 34epoch:train:8001-8500batch: iter_time=1.470e-04, forward_time=0.205, loss_ctc=39.853, loss_att=27.890, acc=0.824, loss=31.479, backward_time=0.236, grad_norm=38.518, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.727e-04, train_time=0.933
+[ip-10-0-216-33:0/16] 2024-03-03 22:04:08,051 (trainer:762) INFO: 34epoch:train:8501-9000batch: iter_time=1.473e-04, forward_time=0.208, loss_ctc=41.048, loss_att=28.232, acc=0.827, loss=32.077, backward_time=0.236, grad_norm=40.574, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.726e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 22:11:45,713 (trainer:762) INFO: 34epoch:train:9001-9500batch: iter_time=1.443e-04, forward_time=0.208, loss_ctc=41.150, loss_att=28.177, acc=0.825, loss=32.069, backward_time=0.237, grad_norm=40.668, clip=100.000, loss_scale=1.991e+33, optim_step_time=0.075, optim0_lr0=1.725e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-03 22:19:22,221 (trainer:762) INFO: 34epoch:train:9501-10000batch: iter_time=1.475e-04, forward_time=0.209, loss_ctc=40.449, loss_att=27.921, acc=0.826, loss=31.679, backward_time=0.231, grad_norm=39.666, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.724e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-03 22:19:28,103 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 22:20:05,418 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 22:20:13,404 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e89d719d0>)
+[ip-10-0-216-33:0/16] 2024-03-03 22:20:13,405 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-03 22:20:13,409 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-03 22:34:39,914 (trainer:762) INFO: 34epoch:train:10001-10500batch: iter_time=0.908, forward_time=0.213, loss_ctc=39.202, loss_att=27.253, acc=0.824, loss=30.837, backward_time=0.230, grad_norm=41.255, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.076, optim0_lr0=1.723e-04, train_time=1.835
+[ip-10-0-216-33:0/16] 2024-03-03 22:42:17,034 (trainer:762) INFO: 34epoch:train:10501-11000batch: iter_time=1.842e-04, forward_time=0.214, loss_ctc=41.753, loss_att=28.370, acc=0.828, loss=32.385, backward_time=0.223, grad_norm=40.902, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.076, optim0_lr0=1.722e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 22:49:47,385 (trainer:762) INFO: 34epoch:train:11001-11500batch: iter_time=1.863e-04, forward_time=0.214, loss_ctc=41.670, loss_att=28.497, acc=0.818, loss=32.449, backward_time=0.226, grad_norm=42.598, clip=100.000, loss_scale=3.982e+33, optim_step_time=0.073, optim0_lr0=1.722e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-03 22:57:25,915 (trainer:762) INFO: 34epoch:train:11501-12000batch: iter_time=1.947e-04, forward_time=0.215, loss_ctc=39.813, loss_att=27.312, acc=0.827, loss=31.062, backward_time=0.223, grad_norm=39.934, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.081, optim0_lr0=1.721e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-03 23:04:57,943 (trainer:762) INFO: 34epoch:train:12001-12500batch: iter_time=1.722e-04, forward_time=0.215, loss_ctc=40.411, loss_att=27.429, acc=0.828, loss=31.323, backward_time=0.222, grad_norm=40.086, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.076, optim0_lr0=1.720e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-03 23:12:32,306 (trainer:762) INFO: 34epoch:train:12501-13000batch: iter_time=1.710e-04, forward_time=0.214, loss_ctc=39.077, loss_att=26.918, acc=0.831, loss=30.566, backward_time=0.222, grad_norm=37.539, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.076, optim0_lr0=1.719e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-03 23:18:45,716 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 23:20:03,856 (trainer:762) INFO: 34epoch:train:13001-13500batch: iter_time=1.743e-04, forward_time=0.214, loss_ctc=39.649, loss_att=27.714, acc=0.825, loss=31.295, backward_time=0.222, grad_norm=37.864, clip=100.000, loss_scale=7.055e+33, optim_step_time=0.075, optim0_lr0=1.718e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-03 23:27:39,566 (trainer:762) INFO: 34epoch:train:13501-14000batch: iter_time=1.730e-04, forward_time=0.215, loss_ctc=41.085, loss_att=28.304, acc=0.827, loss=32.138, backward_time=0.223, grad_norm=39.122, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.079, optim0_lr0=1.717e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-03 23:35:15,384 (trainer:762) INFO: 34epoch:train:14001-14500batch: iter_time=1.566e-04, forward_time=0.213, loss_ctc=40.882, loss_att=28.049, acc=0.826, loss=31.899, backward_time=0.222, grad_norm=42.366, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.073, optim0_lr0=1.716e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-03 23:36:54,395 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-03 23:42:52,639 (trainer:762) INFO: 34epoch:train:14501-15000batch: iter_time=1.539e-04, forward_time=0.212, loss_ctc=40.325, loss_att=27.880, acc=0.826, loss=31.614, backward_time=0.222, grad_norm=39.326, clip=100.000, loss_scale=3.142e+33, optim_step_time=0.072, optim0_lr0=1.716e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-03 23:57:40,387 (trainer:361) INFO: 34epoch results: [train] iter_time=0.097, forward_time=0.211, loss_ctc=40.609, loss_att=27.886, acc=0.825, loss=31.703, backward_time=0.231, grad_norm=40.416, clip=100.000, loss_scale=2.951e+33, optim_step_time=0.075, optim0_lr0=1.728e-04, train_time=1.012, time=4 hours, 13 minutes and 25.91 seconds, total_count=510000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=36.440, cer_ctc=0.197, loss_att=28.467, acc=0.778, cer=0.192, wer=0.986, loss=30.859, time=14 minutes and 30.95 seconds, total_count=72522, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-03 23:57:50,008 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-03 23:57:50,041 (trainer:290) INFO: 35/45epoch started. Estimated time to finish: 2 days, 37 minutes and 52.59 seconds
+[ip-10-0-216-33:0/16] 2024-03-03 23:57:50,049 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-03 23:58:26,429 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-03 23:58:34,306 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ece60bbb0>)
+[ip-10-0-216-33:0/16] 2024-03-03 23:58:34,307 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-03 23:58:34,311 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 00:12:39,636 (trainer:762) INFO: 35epoch:train:1-500batch: iter_time=0.859, forward_time=0.213, loss_ctc=40.779, loss_att=27.818, acc=0.823, loss=31.706, backward_time=0.222, grad_norm=41.463, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.073, optim0_lr0=1.715e-04, train_time=1.779
+[ip-10-0-216-33:0/16] 2024-03-04 00:20:14,949 (trainer:762) INFO: 35epoch:train:501-1000batch: iter_time=1.607e-04, forward_time=0.213, loss_ctc=36.925, loss_att=25.542, acc=0.833, loss=28.957, backward_time=0.222, grad_norm=38.171, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.073, optim0_lr0=1.714e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-04 00:27:46,893 (trainer:762) INFO: 35epoch:train:1001-1500batch: iter_time=1.608e-04, forward_time=0.213, loss_ctc=39.956, loss_att=27.685, acc=0.832, loss=31.366, backward_time=0.223, grad_norm=38.732, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.072, optim0_lr0=1.713e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-04 00:35:16,626 (trainer:762) INFO: 35epoch:train:1501-2000batch: iter_time=1.585e-04, forward_time=0.212, loss_ctc=39.247, loss_att=26.890, acc=0.824, loss=30.597, backward_time=0.222, grad_norm=39.880, clip=100.000, loss_scale=4.642e+33, optim_step_time=0.072, optim0_lr0=1.712e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-04 00:42:49,482 (trainer:762) INFO: 35epoch:train:2001-2500batch: iter_time=1.680e-04, forward_time=0.213, loss_ctc=40.386, loss_att=27.812, acc=0.827, loss=31.584, backward_time=0.222, grad_norm=40.588, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.073, optim0_lr0=1.711e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-04 00:50:30,104 (trainer:762) INFO: 35epoch:train:2501-3000batch: iter_time=1.703e-04, forward_time=0.214, loss_ctc=43.786, loss_att=29.990, acc=0.826, loss=34.129, backward_time=0.223, grad_norm=41.730, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.072, optim0_lr0=1.711e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-04 00:54:08,429 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 00:58:11,734 (trainer:762) INFO: 35epoch:train:3001-3500batch: iter_time=1.689e-04, forward_time=0.212, loss_ctc=38.653, loss_att=26.460, acc=0.825, loss=30.117, backward_time=0.222, grad_norm=38.845, clip=100.000, loss_scale=3.819e+33, optim_step_time=0.072, optim0_lr0=1.710e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-04 01:05:48,228 (trainer:762) INFO: 35epoch:train:3501-4000batch: iter_time=1.706e-04, forward_time=0.213, loss_ctc=40.788, loss_att=27.713, acc=0.824, loss=31.635, backward_time=0.222, grad_norm=41.398, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.072, optim0_lr0=1.709e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-04 01:13:30,536 (trainer:762) INFO: 35epoch:train:4001-4500batch: iter_time=1.899e-04, forward_time=0.217, loss_ctc=39.885, loss_att=27.520, acc=0.833, loss=31.229, backward_time=0.223, grad_norm=38.965, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.083, optim0_lr0=1.708e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-04 01:21:10,662 (trainer:762) INFO: 35epoch:train:4501-5000batch: iter_time=1.901e-04, forward_time=0.215, loss_ctc=39.877, loss_att=27.630, acc=0.824, loss=31.304, backward_time=0.222, grad_norm=40.600, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.083, optim0_lr0=1.707e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-04 01:21:15,644 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 01:21:52,152 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 01:22:00,216 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e0dd728e0>)
+[ip-10-0-216-33:0/16] 2024-03-04 01:22:00,217 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-04 01:22:00,221 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 01:36:22,573 (trainer:762) INFO: 35epoch:train:5001-5500batch: iter_time=0.899, forward_time=0.214, loss_ctc=40.568, loss_att=27.810, acc=0.824, loss=31.637, backward_time=0.222, grad_norm=40.456, clip=100.000, loss_scale=3.967e+33, optim_step_time=0.078, optim0_lr0=1.706e-04, train_time=1.824
+[ip-10-0-216-33:0/16] 2024-03-04 01:38:57,845 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 01:43:54,672 (trainer:762) INFO: 35epoch:train:5501-6000batch: iter_time=1.588e-04, forward_time=0.213, loss_ctc=36.712, loss_att=25.436, acc=0.833, loss=28.819, backward_time=0.222, grad_norm=43.356, clip=100.000, loss_scale=3.491e+33, optim_step_time=0.072, optim0_lr0=1.706e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-04 01:51:23,797 (trainer:762) INFO: 35epoch:train:6001-6500batch: iter_time=1.643e-04, forward_time=0.212, loss_ctc=39.867, loss_att=27.564, acc=0.833, loss=31.255, backward_time=0.222, grad_norm=40.016, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.072, optim0_lr0=1.705e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-03-04 01:58:10,608 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 01:58:57,327 (trainer:762) INFO: 35epoch:train:6501-7000batch: iter_time=1.621e-04, forward_time=0.213, loss_ctc=39.101, loss_att=26.934, acc=0.823, loss=30.584, backward_time=0.222, grad_norm=39.794, clip=100.000, loss_scale=2.461e+33, optim_step_time=0.072, optim0_lr0=1.704e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-04 02:06:26,141 (trainer:762) INFO: 35epoch:train:7001-7500batch: iter_time=1.576e-04, forward_time=0.212, loss_ctc=40.088, loss_att=27.741, acc=0.827, loss=31.445, backward_time=0.223, grad_norm=40.562, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.072, optim0_lr0=1.703e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-03-04 02:14:04,916 (trainer:762) INFO: 35epoch:train:7501-8000batch: iter_time=1.631e-04, forward_time=0.213, loss_ctc=43.305, loss_att=29.777, acc=0.826, loss=33.836, backward_time=0.223, grad_norm=41.759, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.072, optim0_lr0=1.702e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-04 02:21:40,058 (trainer:762) INFO: 35epoch:train:8001-8500batch: iter_time=1.591e-04, forward_time=0.211, loss_ctc=38.570, loss_att=26.472, acc=0.825, loss=30.101, backward_time=0.222, grad_norm=40.548, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.073, optim0_lr0=1.701e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-04 02:29:15,854 (trainer:762) INFO: 35epoch:train:8501-9000batch: iter_time=1.637e-04, forward_time=0.213, loss_ctc=40.543, loss_att=27.525, acc=0.825, loss=31.431, backward_time=0.222, grad_norm=41.083, clip=100.000, loss_scale=1.433e+33, optim_step_time=0.072, optim0_lr0=1.701e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-04 02:36:47,925 (trainer:762) INFO: 35epoch:train:9001-9500batch: iter_time=1.638e-04, forward_time=0.213, loss_ctc=39.706, loss_att=27.439, acc=0.833, loss=31.119, backward_time=0.222, grad_norm=40.340, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.072, optim0_lr0=1.700e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-04 02:44:18,294 (trainer:762) INFO: 35epoch:train:9501-10000batch: iter_time=1.651e-04, forward_time=0.212, loss_ctc=39.981, loss_att=27.658, acc=0.824, loss=31.355, backward_time=0.222, grad_norm=40.987, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.072, optim0_lr0=1.699e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-04 02:44:26,198 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 02:45:03,156 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 02:45:11,310 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed06da5b0>)
+[ip-10-0-216-33:0/16] 2024-03-04 02:45:11,310 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-04 02:45:11,315 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 03:03:30,019 (trainer:762) INFO: 35epoch:train:10001-10500batch: iter_time=1.363, forward_time=0.213, loss_ctc=40.621, loss_att=27.833, acc=0.824, loss=31.669, backward_time=0.222, grad_norm=40.822, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.072, optim0_lr0=1.698e-04, train_time=2.303
+[ip-10-0-216-33:0/16] 2024-03-04 03:11:06,734 (trainer:762) INFO: 35epoch:train:10501-11000batch: iter_time=1.630e-04, forward_time=0.213, loss_ctc=36.671, loss_att=25.405, acc=0.833, loss=28.785, backward_time=0.223, grad_norm=37.641, clip=100.000, loss_scale=2.866e+33, optim_step_time=0.072, optim0_lr0=1.697e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-04 03:14:13,279 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 03:18:37,248 (trainer:762) INFO: 35epoch:train:11001-11500batch: iter_time=1.666e-04, forward_time=0.213, loss_ctc=39.673, loss_att=27.508, acc=0.833, loss=31.158, backward_time=0.223, grad_norm=39.700, clip=100.000, loss_scale=3.668e+33, optim_step_time=0.073, optim0_lr0=1.697e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-04 03:26:10,679 (trainer:762) INFO: 35epoch:train:11501-12000batch: iter_time=1.604e-04, forward_time=0.213, loss_ctc=38.812, loss_att=26.741, acc=0.825, loss=30.362, backward_time=0.222, grad_norm=39.442, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.073, optim0_lr0=1.696e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-04 03:33:41,991 (trainer:762) INFO: 35epoch:train:12001-12500batch: iter_time=1.645e-04, forward_time=0.213, loss_ctc=40.191, loss_att=27.679, acc=0.827, loss=31.432, backward_time=0.222, grad_norm=55.790, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.072, optim0_lr0=1.695e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-04 03:41:12,082 (trainer:762) INFO: 35epoch:train:12501-13000batch: iter_time=1.616e-04, forward_time=0.213, loss_ctc=43.220, loss_att=29.712, acc=0.827, loss=33.764, backward_time=0.223, grad_norm=40.560, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.073, optim0_lr0=1.694e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-04 03:48:43,706 (trainer:762) INFO: 35epoch:train:13001-13500batch: iter_time=1.580e-04, forward_time=0.213, loss_ctc=38.482, loss_att=26.391, acc=0.825, loss=30.019, backward_time=0.223, grad_norm=41.023, clip=100.000, loss_scale=4.117e+33, optim_step_time=0.072, optim0_lr0=1.693e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-04 03:52:15,224 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 03:53:22,563 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 03:56:10,680 (trainer:762) INFO: 35epoch:train:13501-14000batch: iter_time=1.636e-04, forward_time=0.212, loss_ctc=40.289, loss_att=27.378, acc=0.826, loss=31.252, backward_time=0.222, grad_norm=42.334, clip=100.000, loss_scale=3.336e+33, optim_step_time=0.072, optim0_lr0=1.693e-04, train_time=0.894
+[ip-10-0-216-33:0/16] 2024-03-04 04:03:46,342 (trainer:762) INFO: 35epoch:train:14001-14500batch: iter_time=1.605e-04, forward_time=0.214, loss_ctc=39.531, loss_att=27.375, acc=0.834, loss=31.022, backward_time=0.222, grad_norm=39.430, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.072, optim0_lr0=1.692e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-04 04:11:21,267 (trainer:762) INFO: 35epoch:train:14501-15000batch: iter_time=1.649e-04, forward_time=0.212, loss_ctc=39.551, loss_att=27.382, acc=0.824, loss=31.033, backward_time=0.222, grad_norm=41.090, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.072, optim0_lr0=1.691e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-04 04:25:48,149 (trainer:361) INFO: 35epoch results: [train] iter_time=0.104, forward_time=0.213, loss_ctc=39.859, loss_att=27.427, acc=0.827, loss=31.157, backward_time=0.222, grad_norm=40.903, clip=100.000, loss_scale=2.814e+33, optim_step_time=0.073, optim0_lr0=1.703e-04, train_time=1.014, time=4 hours, 13 minutes and 46.1 seconds, total_count=525000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=36.507, cer_ctc=0.198, loss_att=28.798, acc=0.764, cer=0.236, wer=0.991, loss=31.111, time=14 minutes and 11.65 seconds, total_count=74655, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-04 04:25:57,932 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-04 04:25:57,963 (trainer:290) INFO: 36/45epoch started. Estimated time to finish: 1 day, 20 hours and 13 minutes
+[ip-10-0-216-33:0/16] 2024-03-04 04:25:57,970 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 04:26:34,129 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 04:26:42,010 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ec3f15dc0>)
+[ip-10-0-216-33:0/16] 2024-03-04 04:26:42,011 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-04 04:26:42,015 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 04:40:35,501 (trainer:762) INFO: 36epoch:train:1-500batch: iter_time=0.842, forward_time=0.214, loss_ctc=42.433, loss_att=28.876, acc=0.821, loss=32.943, backward_time=0.222, grad_norm=44.437, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.690e-04, train_time=1.755
+[ip-10-0-216-33:0/16] 2024-03-04 04:48:11,942 (trainer:762) INFO: 36epoch:train:501-1000batch: iter_time=1.501e-04, forward_time=0.215, loss_ctc=38.738, loss_att=26.661, acc=0.826, loss=30.284, backward_time=0.223, grad_norm=39.887, clip=100.000, loss_scale=1.786e+33, optim_step_time=0.077, optim0_lr0=1.689e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-04 04:55:47,622 (trainer:762) INFO: 36epoch:train:1001-1500batch: iter_time=1.599e-04, forward_time=0.212, loss_ctc=39.304, loss_att=26.632, acc=0.831, loss=30.434, backward_time=0.223, grad_norm=39.217, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.073, optim0_lr0=1.689e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-04 04:57:45,836 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 05:03:24,831 (trainer:762) INFO: 36epoch:train:1501-2000batch: iter_time=1.700e-04, forward_time=0.214, loss_ctc=43.428, loss_att=30.244, acc=0.820, loss=34.199, backward_time=0.222, grad_norm=43.416, clip=100.000, loss_scale=1.631e+33, optim_step_time=0.073, optim0_lr0=1.688e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-04 05:10:57,309 (trainer:762) INFO: 36epoch:train:2001-2500batch: iter_time=1.599e-04, forward_time=0.213, loss_ctc=39.282, loss_att=27.054, acc=0.827, loss=30.723, backward_time=0.222, grad_norm=38.415, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.072, optim0_lr0=1.687e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-04 05:18:36,646 (trainer:762) INFO: 36epoch:train:2501-3000batch: iter_time=1.408e-04, forward_time=0.213, loss_ctc=39.528, loss_att=26.944, acc=0.828, loss=30.719, backward_time=0.222, grad_norm=37.815, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.076, optim0_lr0=1.686e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-04 05:26:13,295 (trainer:762) INFO: 36epoch:train:3001-3500batch: iter_time=1.363e-04, forward_time=0.213, loss_ctc=42.939, loss_att=30.051, acc=0.823, loss=33.917, backward_time=0.223, grad_norm=45.075, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.073, optim0_lr0=1.685e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-04 05:33:44,669 (trainer:762) INFO: 36epoch:train:3501-4000batch: iter_time=1.507e-04, forward_time=0.212, loss_ctc=40.505, loss_att=27.659, acc=0.821, loss=31.513, backward_time=0.222, grad_norm=40.392, clip=100.000, loss_scale=2.261e+33, optim_step_time=0.072, optim0_lr0=1.685e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-04 05:41:17,084 (trainer:762) INFO: 36epoch:train:4001-4500batch: iter_time=1.393e-04, forward_time=0.213, loss_ctc=39.745, loss_att=27.095, acc=0.825, loss=30.890, backward_time=0.223, grad_norm=40.147, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.074, optim0_lr0=1.684e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-04 05:48:48,981 (trainer:762) INFO: 36epoch:train:4501-5000batch: iter_time=1.432e-04, forward_time=0.213, loss_ctc=40.841, loss_att=27.738, acc=0.833, loss=31.669, backward_time=0.222, grad_norm=39.595, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.072, optim0_lr0=1.683e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-04 05:48:53,470 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 05:49:31,390 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 05:49:39,490 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfd57d90>)
+[ip-10-0-216-33:0/16] 2024-03-04 05:49:39,490 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-04 05:49:39,495 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 06:04:00,027 (trainer:762) INFO: 36epoch:train:5001-5500batch: iter_time=0.879, forward_time=0.213, loss_ctc=41.748, loss_att=28.557, acc=0.823, loss=32.514, backward_time=0.222, grad_norm=43.542, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.073, optim0_lr0=1.682e-04, train_time=1.822
+[ip-10-0-216-33:0/16] 2024-03-04 06:11:40,124 (trainer:762) INFO: 36epoch:train:5501-6000batch: iter_time=1.625e-04, forward_time=0.211, loss_ctc=38.610, loss_att=26.525, acc=0.826, loss=30.150, backward_time=0.222, grad_norm=40.421, clip=100.000, loss_scale=4.522e+33, optim_step_time=0.073, optim0_lr0=1.681e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-04 06:19:16,210 (trainer:762) INFO: 36epoch:train:6001-6500batch: iter_time=1.728e-04, forward_time=0.213, loss_ctc=38.842, loss_att=26.445, acc=0.832, loss=30.164, backward_time=0.223, grad_norm=38.550, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.073, optim0_lr0=1.681e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-04 06:26:34,664 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 06:26:52,103 (trainer:762) INFO: 36epoch:train:6501-7000batch: iter_time=1.629e-04, forward_time=0.211, loss_ctc=42.970, loss_att=29.970, acc=0.821, loss=33.870, backward_time=0.224, grad_norm=42.902, clip=100.000, loss_scale=5.093e+33, optim_step_time=0.073, optim0_lr0=1.680e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-04 06:34:21,791 (trainer:762) INFO: 36epoch:train:7001-7500batch: iter_time=1.718e-04, forward_time=0.214, loss_ctc=38.648, loss_att=26.737, acc=0.828, loss=30.311, backward_time=0.223, grad_norm=38.056, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.074, optim0_lr0=1.679e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-04 06:42:01,270 (trainer:762) INFO: 36epoch:train:7501-8000batch: iter_time=1.687e-04, forward_time=0.212, loss_ctc=39.450, loss_att=26.910, acc=0.828, loss=30.672, backward_time=0.222, grad_norm=36.970, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.073, optim0_lr0=1.678e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-04 06:49:40,589 (trainer:762) INFO: 36epoch:train:8001-8500batch: iter_time=1.620e-04, forward_time=0.212, loss_ctc=42.565, loss_att=29.810, acc=0.824, loss=33.637, backward_time=0.223, grad_norm=43.570, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.073, optim0_lr0=1.677e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-04 06:57:17,035 (trainer:762) INFO: 36epoch:train:8501-9000batch: iter_time=1.670e-04, forward_time=0.212, loss_ctc=40.005, loss_att=27.509, acc=0.823, loss=31.258, backward_time=0.222, grad_norm=39.325, clip=100.000, loss_scale=2.695e+33, optim_step_time=0.072, optim0_lr0=1.677e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-04 07:01:59,781 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 07:04:45,772 (trainer:762) INFO: 36epoch:train:9001-9500batch: iter_time=1.600e-04, forward_time=0.212, loss_ctc=39.597, loss_att=27.050, acc=0.826, loss=30.814, backward_time=0.223, grad_norm=40.304, clip=100.000, loss_scale=4.219e+33, optim_step_time=0.072, optim0_lr0=1.676e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-03-04 07:12:16,201 (trainer:762) INFO: 36epoch:train:9501-10000batch: iter_time=1.627e-04, forward_time=0.213, loss_ctc=40.447, loss_att=27.579, acc=0.834, loss=31.439, backward_time=0.223, grad_norm=39.651, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.072, optim0_lr0=1.675e-04, train_time=0.901
+[ip-10-0-216-33:0/16] 2024-03-04 07:12:21,076 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 07:12:58,136 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 07:13:06,253 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e843c48b0>)
+[ip-10-0-216-33:0/16] 2024-03-04 07:13:06,253 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-04 07:13:06,257 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 07:26:15,503 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 07:27:26,110 (trainer:762) INFO: 36epoch:train:10001-10500batch: iter_time=0.900, forward_time=0.214, loss_ctc=41.340, loss_att=28.360, acc=0.824, loss=32.254, backward_time=0.222, grad_norm=43.361, clip=100.000, loss_scale=2.391e+33, optim_step_time=0.072, optim0_lr0=1.674e-04, train_time=1.820
+[ip-10-0-216-33:0/16] 2024-03-04 07:35:05,361 (trainer:762) INFO: 36epoch:train:10501-11000batch: iter_time=1.666e-04, forward_time=0.213, loss_ctc=38.373, loss_att=26.453, acc=0.826, loss=30.029, backward_time=0.222, grad_norm=39.765, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.073, optim0_lr0=1.673e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-04 07:42:37,832 (trainer:762) INFO: 36epoch:train:11001-11500batch: iter_time=1.790e-04, forward_time=0.216, loss_ctc=38.689, loss_att=26.397, acc=0.832, loss=30.085, backward_time=0.223, grad_norm=39.076, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.077, optim0_lr0=1.673e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-04 07:50:13,905 (trainer:762) INFO: 36epoch:train:11501-12000batch: iter_time=1.708e-04, forward_time=0.213, loss_ctc=42.683, loss_att=29.862, acc=0.822, loss=33.708, backward_time=0.222, grad_norm=45.698, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.073, optim0_lr0=1.672e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-04 07:57:51,548 (trainer:762) INFO: 36epoch:train:12001-12500batch: iter_time=1.613e-04, forward_time=0.213, loss_ctc=38.628, loss_att=26.620, acc=0.828, loss=30.223, backward_time=0.221, grad_norm=38.351, clip=100.000, loss_scale=1.503e+33, optim_step_time=0.073, optim0_lr0=1.671e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-04 08:05:27,594 (trainer:762) INFO: 36epoch:train:12501-13000batch: iter_time=1.595e-04, forward_time=0.212, loss_ctc=39.347, loss_att=26.840, acc=0.829, loss=30.592, backward_time=0.222, grad_norm=37.230, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.072, optim0_lr0=1.670e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-04 08:13:07,847 (trainer:762) INFO: 36epoch:train:13001-13500batch: iter_time=1.685e-04, forward_time=0.214, loss_ctc=42.461, loss_att=29.816, acc=0.824, loss=33.609, backward_time=0.224, grad_norm=43.471, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.073, optim0_lr0=1.670e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-04 08:14:14,214 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 08:20:37,442 (trainer:762) INFO: 36epoch:train:13501-14000batch: iter_time=1.694e-04, forward_time=0.212, loss_ctc=39.916, loss_att=27.375, acc=0.823, loss=31.137, backward_time=0.223, grad_norm=40.815, clip=100.000, loss_scale=1.485e+33, optim_step_time=0.073, optim0_lr0=1.669e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-04 08:28:09,422 (trainer:762) INFO: 36epoch:train:14001-14500batch: iter_time=1.751e-04, forward_time=0.214, loss_ctc=39.337, loss_att=26.898, acc=0.826, loss=30.630, backward_time=0.222, grad_norm=40.434, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.073, optim0_lr0=1.668e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-04 08:35:40,866 (trainer:762) INFO: 36epoch:train:14501-15000batch: iter_time=1.688e-04, forward_time=0.213, loss_ctc=40.284, loss_att=27.450, acc=0.834, loss=31.300, backward_time=0.223, grad_norm=40.001, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.072, optim0_lr0=1.667e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-04 08:50:08,297 (trainer:361) INFO: 36epoch results: [train] iter_time=0.088, forward_time=0.213, loss_ctc=40.356, loss_att=27.737, acc=0.826, loss=31.523, backward_time=0.222, grad_norm=40.662, clip=100.000, loss_scale=2.347e+33, optim_step_time=0.073, optim0_lr0=1.679e-04, train_time=0.999, time=4 hours, 9 minutes and 57.8 seconds, total_count=540000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=35.410, cer_ctc=0.192, loss_att=28.720, acc=0.757, cer=0.267, wer=0.997, loss=30.727, time=14 minutes and 12.22 seconds, total_count=76788, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-04 08:50:18,313 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-04 08:50:18,358 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/31epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-04 08:50:18,359 (trainer:290) INFO: 37/45epoch started. Estimated time to finish: 1 day, 15 hours and 47 minutes
+[ip-10-0-216-33:0/16] 2024-03-04 08:50:18,367 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 08:50:54,764 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 08:51:02,759 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0291700>)
+[ip-10-0-216-33:0/16] 2024-03-04 08:51:02,759 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-04 08:51:02,764 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 09:04:54,220 (trainer:762) INFO: 37epoch:train:1-500batch: iter_time=0.837, forward_time=0.213, loss_ctc=39.833, loss_att=28.094, acc=0.826, loss=31.616, backward_time=0.228, grad_norm=43.256, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.666e-04, train_time=1.751
+[ip-10-0-216-33:0/16] 2024-03-04 09:12:27,045 (trainer:762) INFO: 37epoch:train:501-1000batch: iter_time=1.584e-04, forward_time=0.213, loss_ctc=39.236, loss_att=26.760, acc=0.825, loss=30.503, backward_time=0.227, grad_norm=42.666, clip=100.000, loss_scale=2.407e+33, optim_step_time=0.074, optim0_lr0=1.666e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-04 09:20:04,876 (trainer:762) INFO: 37epoch:train:1001-1500batch: iter_time=1.745e-04, forward_time=0.216, loss_ctc=40.789, loss_att=27.725, acc=0.828, loss=31.644, backward_time=0.226, grad_norm=40.046, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.665e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-04 09:27:37,968 (trainer:762) INFO: 37epoch:train:1501-2000batch: iter_time=1.551e-04, forward_time=0.209, loss_ctc=39.878, loss_att=27.329, acc=0.831, loss=31.094, backward_time=0.232, grad_norm=42.008, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.664e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-04 09:35:09,818 (trainer:762) INFO: 37epoch:train:2001-2500batch: iter_time=1.511e-04, forward_time=0.211, loss_ctc=38.591, loss_att=26.595, acc=0.830, loss=30.193, backward_time=0.236, grad_norm=39.372, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.663e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-04 09:36:26,753 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 09:39:52,924 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 09:42:48,472 (trainer:762) INFO: 37epoch:train:2501-3000batch: iter_time=1.523e-04, forward_time=0.210, loss_ctc=40.897, loss_att=27.933, acc=0.831, loss=31.822, backward_time=0.237, grad_norm=40.653, clip=100.000, loss_scale=2.135e+33, optim_step_time=0.075, optim0_lr0=1.663e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-04 09:46:53,172 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 09:50:23,106 (trainer:762) INFO: 37epoch:train:3001-3500batch: iter_time=1.451e-04, forward_time=0.209, loss_ctc=40.388, loss_att=27.682, acc=0.824, loss=31.494, backward_time=0.243, grad_norm=41.884, clip=100.000, loss_scale=9.989e+32, optim_step_time=0.075, optim0_lr0=1.662e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-04 09:57:57,202 (trainer:762) INFO: 37epoch:train:3501-4000batch: iter_time=1.507e-04, forward_time=0.210, loss_ctc=40.008, loss_att=27.337, acc=0.830, loss=31.138, backward_time=0.236, grad_norm=39.153, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.661e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-04 10:05:36,784 (trainer:762) INFO: 37epoch:train:4001-4500batch: iter_time=1.493e-04, forward_time=0.209, loss_ctc=42.677, loss_att=28.807, acc=0.826, loss=32.968, backward_time=0.241, grad_norm=42.185, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.660e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-04 10:13:11,411 (trainer:762) INFO: 37epoch:train:4501-5000batch: iter_time=1.463e-04, forward_time=0.210, loss_ctc=39.151, loss_att=26.990, acc=0.836, loss=30.638, backward_time=0.232, grad_norm=41.049, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.660e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-04 10:13:16,135 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 10:13:52,571 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 10:14:00,598 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ea105f130>)
+[ip-10-0-216-33:0/16] 2024-03-04 10:14:00,598 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-04 10:14:00,603 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 10:28:20,379 (trainer:762) INFO: 37epoch:train:5001-5500batch: iter_time=0.901, forward_time=0.209, loss_ctc=39.623, loss_att=27.912, acc=0.826, loss=31.425, backward_time=0.235, grad_norm=42.430, clip=100.000, loss_scale=9.476e+32, optim_step_time=0.075, optim0_lr0=1.659e-04, train_time=1.818
+[ip-10-0-216-33:0/16] 2024-03-04 10:35:53,868 (trainer:762) INFO: 37epoch:train:5501-6000batch: iter_time=1.530e-04, forward_time=0.210, loss_ctc=38.993, loss_att=26.669, acc=0.826, loss=30.366, backward_time=0.231, grad_norm=41.917, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.658e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-04 10:43:28,037 (trainer:762) INFO: 37epoch:train:6001-6500batch: iter_time=1.612e-04, forward_time=0.212, loss_ctc=40.470, loss_att=27.617, acc=0.828, loss=31.473, backward_time=0.223, grad_norm=47.501, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.074, optim0_lr0=1.657e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-04 10:51:09,684 (trainer:762) INFO: 37epoch:train:6501-7000batch: iter_time=1.602e-04, forward_time=0.213, loss_ctc=39.764, loss_att=27.255, acc=0.831, loss=31.008, backward_time=0.224, grad_norm=41.927, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.073, optim0_lr0=1.657e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-04 10:57:17,970 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 10:58:47,283 (trainer:762) INFO: 37epoch:train:7001-7500batch: iter_time=1.528e-04, forward_time=0.211, loss_ctc=38.828, loss_att=26.658, acc=0.830, loss=30.309, backward_time=0.222, grad_norm=39.031, clip=100.000, loss_scale=1.644e+33, optim_step_time=0.072, optim0_lr0=1.656e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-04 11:06:23,789 (trainer:762) INFO: 37epoch:train:7501-8000batch: iter_time=1.477e-04, forward_time=0.215, loss_ctc=40.548, loss_att=27.837, acc=0.833, loss=31.650, backward_time=0.222, grad_norm=40.102, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.074, optim0_lr0=1.655e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-04 11:13:55,304 (trainer:762) INFO: 37epoch:train:8001-8500batch: iter_time=1.729e-04, forward_time=0.212, loss_ctc=39.941, loss_att=27.413, acc=0.825, loss=31.172, backward_time=0.222, grad_norm=41.597, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.074, optim0_lr0=1.654e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-04 11:21:30,938 (trainer:762) INFO: 37epoch:train:8501-9000batch: iter_time=1.695e-04, forward_time=0.214, loss_ctc=39.829, loss_att=27.217, acc=0.830, loss=31.001, backward_time=0.222, grad_norm=40.428, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.654e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-04 11:29:10,950 (trainer:762) INFO: 37epoch:train:9001-9500batch: iter_time=1.673e-04, forward_time=0.212, loss_ctc=42.608, loss_att=28.800, acc=0.826, loss=32.942, backward_time=0.222, grad_norm=42.423, clip=100.000, loss_scale=1.547e+33, optim_step_time=0.072, optim0_lr0=1.653e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-04 11:33:56,055 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 11:36:49,859 (trainer:762) INFO: 37epoch:train:9501-10000batch: iter_time=1.633e-04, forward_time=0.212, loss_ctc=39.103, loss_att=26.951, acc=0.837, loss=30.597, backward_time=0.223, grad_norm=39.650, clip=100.000, loss_scale=2.104e+33, optim_step_time=0.072, optim0_lr0=1.652e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-04 11:36:56,575 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 11:37:33,870 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 11:37:42,858 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfdda310>)
+[ip-10-0-216-33:0/16] 2024-03-04 11:37:42,858 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-04 11:37:42,863 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 11:52:13,801 (trainer:762) INFO: 37epoch:train:10001-10500batch: iter_time=0.913, forward_time=0.211, loss_ctc=39.381, loss_att=27.755, acc=0.827, loss=31.243, backward_time=0.231, grad_norm=43.140, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.651e-04, train_time=1.848
+[ip-10-0-216-33:0/16] 2024-03-04 11:59:52,640 (trainer:762) INFO: 37epoch:train:10501-11000batch: iter_time=1.466e-04, forward_time=0.208, loss_ctc=38.656, loss_att=26.432, acc=0.827, loss=30.099, backward_time=0.232, grad_norm=42.003, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.651e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-04 12:07:34,669 (trainer:762) INFO: 37epoch:train:11001-11500batch: iter_time=1.488e-04, forward_time=0.210, loss_ctc=40.316, loss_att=27.519, acc=0.828, loss=31.358, backward_time=0.234, grad_norm=39.564, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.650e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-04 12:15:17,613 (trainer:762) INFO: 37epoch:train:11501-12000batch: iter_time=1.455e-04, forward_time=0.207, loss_ctc=39.501, loss_att=27.135, acc=0.832, loss=30.845, backward_time=0.233, grad_norm=40.581, clip=100.000, loss_scale=1.789e+33, optim_step_time=0.075, optim0_lr0=1.649e-04, train_time=0.926
+[ip-10-0-216-33:0/16] 2024-03-04 12:23:01,727 (trainer:762) INFO: 37epoch:train:12001-12500batch: iter_time=1.438e-04, forward_time=0.208, loss_ctc=38.353, loss_att=26.425, acc=0.831, loss=30.003, backward_time=0.232, grad_norm=40.067, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.648e-04, train_time=0.928
+[ip-10-0-216-33:0/16] 2024-03-04 12:30:36,713 (trainer:762) INFO: 37epoch:train:12501-13000batch: iter_time=1.433e-04, forward_time=0.209, loss_ctc=40.259, loss_att=27.692, acc=0.833, loss=31.463, backward_time=0.233, grad_norm=39.672, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.648e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-04 12:38:20,082 (trainer:762) INFO: 37epoch:train:13001-13500batch: iter_time=1.457e-04, forward_time=0.208, loss_ctc=40.053, loss_att=27.488, acc=0.825, loss=31.258, backward_time=0.236, grad_norm=40.705, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.647e-04, train_time=0.926
+[ip-10-0-216-33:0/16] 2024-03-04 12:42:21,069 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 12:46:00,013 (trainer:762) INFO: 37epoch:train:13501-14000batch: iter_time=1.487e-04, forward_time=0.207, loss_ctc=39.627, loss_att=27.071, acc=0.830, loss=30.838, backward_time=0.236, grad_norm=39.517, clip=100.000, loss_scale=1.974e+33, optim_step_time=0.075, optim0_lr0=1.646e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-04 12:53:43,543 (trainer:762) INFO: 37epoch:train:14001-14500batch: iter_time=1.448e-04, forward_time=0.209, loss_ctc=42.371, loss_att=28.626, acc=0.827, loss=32.750, backward_time=0.235, grad_norm=42.051, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.645e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-03-04 13:01:27,335 (trainer:762) INFO: 37epoch:train:14501-15000batch: iter_time=1.424e-04, forward_time=0.208, loss_ctc=38.783, loss_att=26.750, acc=0.837, loss=30.360, backward_time=0.233, grad_norm=40.196, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.645e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-03-04 13:15:39,098 (trainer:361) INFO: 37epoch results: [train] iter_time=0.089, forward_time=0.210, loss_ctc=39.949, loss_att=27.416, acc=0.829, loss=31.176, backward_time=0.231, grad_norm=41.226, clip=100.000, loss_scale=1.622e+33, optim_step_time=0.075, optim0_lr0=1.655e-04, train_time=1.004, time=4 hours, 11 minutes and 25.3 seconds, total_count=555000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=35.974, cer_ctc=0.197, loss_att=30.290, acc=0.752, cer=0.293, wer=0.995, loss=31.995, time=13 minutes and 55.1 seconds, total_count=78921, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-04 13:15:49,177 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-04 13:15:49,207 (trainer:290) INFO: 38/45epoch started. Estimated time to finish: 1 day, 11 hours and 22 minutes
+[ip-10-0-216-33:0/16] 2024-03-04 13:15:49,217 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 13:16:26,159 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 13:16:34,690 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ea43fe460>)
+[ip-10-0-216-33:0/16] 2024-03-04 13:16:34,690 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-04 13:16:34,695 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 13:30:28,796 (trainer:762) INFO: 38epoch:train:1-500batch: iter_time=0.840, forward_time=0.210, loss_ctc=38.871, loss_att=27.312, acc=0.829, loss=30.780, backward_time=0.234, grad_norm=41.337, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.644e-04, train_time=1.759
+[ip-10-0-216-33:0/16] 2024-03-04 13:38:05,175 (trainer:762) INFO: 38epoch:train:501-1000batch: iter_time=1.482e-04, forward_time=0.209, loss_ctc=40.672, loss_att=27.534, acc=0.827, loss=31.475, backward_time=0.233, grad_norm=42.438, clip=100.000, loss_scale=1.919e+33, optim_step_time=0.075, optim0_lr0=1.643e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-04 13:44:13,978 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 13:45:50,592 (trainer:762) INFO: 38epoch:train:1001-1500batch: iter_time=1.473e-04, forward_time=0.207, loss_ctc=42.723, loss_att=29.375, acc=0.821, loss=33.380, backward_time=0.231, grad_norm=44.545, clip=100.000, loss_scale=2.326e+33, optim_step_time=0.075, optim0_lr0=1.642e-04, train_time=0.931
+[ip-10-0-216-33:0/16] 2024-03-04 13:48:18,352 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 13:53:31,093 (trainer:762) INFO: 38epoch:train:1501-2000batch: iter_time=1.440e-04, forward_time=0.209, loss_ctc=38.999, loss_att=27.358, acc=0.822, loss=30.850, backward_time=0.235, grad_norm=42.788, clip=100.000, loss_scale=8.545e+32, optim_step_time=0.075, optim0_lr0=1.642e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-04 14:01:10,182 (trainer:762) INFO: 38epoch:train:2001-2500batch: iter_time=1.445e-04, forward_time=0.209, loss_ctc=39.047, loss_att=26.688, acc=0.833, loss=30.396, backward_time=0.233, grad_norm=41.273, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.641e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-04 14:08:52,635 (trainer:762) INFO: 38epoch:train:2501-3000batch: iter_time=1.442e-04, forward_time=0.210, loss_ctc=40.311, loss_att=28.063, acc=0.831, loss=31.737, backward_time=0.235, grad_norm=41.777, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.640e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-04 14:16:29,727 (trainer:762) INFO: 38epoch:train:3001-3500batch: iter_time=1.450e-04, forward_time=0.210, loss_ctc=40.087, loss_att=27.576, acc=0.829, loss=31.329, backward_time=0.236, grad_norm=40.299, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.639e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-04 14:24:05,794 (trainer:762) INFO: 38epoch:train:3501-4000batch: iter_time=1.446e-04, forward_time=0.209, loss_ctc=38.647, loss_att=26.372, acc=0.830, loss=30.055, backward_time=0.231, grad_norm=39.148, clip=100.000, loss_scale=1.092e+33, optim_step_time=0.075, optim0_lr0=1.639e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-04 14:31:50,816 (trainer:762) INFO: 38epoch:train:4001-4500batch: iter_time=1.451e-04, forward_time=0.210, loss_ctc=41.204, loss_att=28.346, acc=0.825, loss=32.204, backward_time=0.239, grad_norm=43.758, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.638e-04, train_time=0.930
+[ip-10-0-216-33:0/16] 2024-03-04 14:39:34,610 (trainer:762) INFO: 38epoch:train:4501-5000batch: iter_time=1.382e-04, forward_time=0.209, loss_ctc=38.825, loss_att=26.795, acc=0.830, loss=30.404, backward_time=0.229, grad_norm=38.939, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.637e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-03-04 14:39:40,565 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 14:40:17,154 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 14:40:25,079 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0616370>)
+[ip-10-0-216-33:0/16] 2024-03-04 14:40:25,080 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-04 14:40:25,084 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 14:54:49,542 (trainer:762) INFO: 38epoch:train:5001-5500batch: iter_time=0.894, forward_time=0.208, loss_ctc=38.578, loss_att=26.979, acc=0.831, loss=30.459, backward_time=0.234, grad_norm=40.801, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.636e-04, train_time=1.830
+[ip-10-0-216-33:0/16] 2024-03-04 15:02:28,335 (trainer:762) INFO: 38epoch:train:5501-6000batch: iter_time=1.471e-04, forward_time=0.208, loss_ctc=40.192, loss_att=27.339, acc=0.828, loss=31.195, backward_time=0.232, grad_norm=41.893, clip=100.000, loss_scale=2.183e+33, optim_step_time=0.075, optim0_lr0=1.636e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-04 15:08:58,028 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 15:10:02,118 (trainer:762) INFO: 38epoch:train:6001-6500batch: iter_time=1.462e-04, forward_time=0.210, loss_ctc=42.717, loss_att=29.326, acc=0.822, loss=33.343, backward_time=0.236, grad_norm=44.476, clip=100.000, loss_scale=2.411e+33, optim_step_time=0.075, optim0_lr0=1.635e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-04 15:17:37,264 (trainer:762) INFO: 38epoch:train:6501-7000batch: iter_time=1.457e-04, forward_time=0.208, loss_ctc=38.736, loss_att=27.152, acc=0.823, loss=30.627, backward_time=0.235, grad_norm=42.243, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.634e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-04 15:25:12,428 (trainer:762) INFO: 38epoch:train:7001-7500batch: iter_time=1.448e-04, forward_time=0.208, loss_ctc=38.862, loss_att=26.619, acc=0.834, loss=30.292, backward_time=0.234, grad_norm=41.739, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.634e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-04 15:32:48,088 (trainer:762) INFO: 38epoch:train:7501-8000batch: iter_time=1.473e-04, forward_time=0.209, loss_ctc=39.829, loss_att=27.907, acc=0.831, loss=31.483, backward_time=0.233, grad_norm=42.086, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.633e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-04 15:40:30,124 (trainer:762) INFO: 38epoch:train:8001-8500batch: iter_time=1.477e-04, forward_time=0.209, loss_ctc=39.747, loss_att=27.427, acc=0.830, loss=31.123, backward_time=0.234, grad_norm=39.590, clip=100.000, loss_scale=1.482e+33, optim_step_time=0.075, optim0_lr0=1.632e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-04 15:47:40,864 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 15:48:10,668 (trainer:762) INFO: 38epoch:train:8501-9000batch: iter_time=1.457e-04, forward_time=0.209, loss_ctc=38.397, loss_att=26.219, acc=0.831, loss=29.872, backward_time=0.231, grad_norm=39.544, clip=100.000, loss_scale=2.513e+33, optim_step_time=0.075, optim0_lr0=1.631e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-04 15:55:54,356 (trainer:762) INFO: 38epoch:train:9001-9500batch: iter_time=1.428e-04, forward_time=0.207, loss_ctc=40.931, loss_att=28.233, acc=0.825, loss=32.042, backward_time=0.237, grad_norm=43.748, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.631e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-03-04 16:03:27,308 (trainer:762) INFO: 38epoch:train:9501-10000batch: iter_time=1.429e-04, forward_time=0.209, loss_ctc=38.546, loss_att=26.783, acc=0.830, loss=30.312, backward_time=0.231, grad_norm=40.551, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.630e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-04 16:03:33,282 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 16:04:10,646 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 16:04:18,441 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0616460>)
+[ip-10-0-216-33:0/16] 2024-03-04 16:04:18,441 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-04 16:04:18,445 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 16:18:47,906 (trainer:762) INFO: 38epoch:train:10001-10500batch: iter_time=0.912, forward_time=0.210, loss_ctc=38.546, loss_att=26.878, acc=0.831, loss=30.379, backward_time=0.235, grad_norm=42.156, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.629e-04, train_time=1.841
+[ip-10-0-216-33:0/16] 2024-03-04 16:26:32,491 (trainer:762) INFO: 38epoch:train:10501-11000batch: iter_time=1.492e-04, forward_time=0.208, loss_ctc=40.097, loss_att=27.359, acc=0.828, loss=31.180, backward_time=0.234, grad_norm=48.920, clip=100.000, loss_scale=1.381e+33, optim_step_time=0.075, optim0_lr0=1.628e-04, train_time=0.929
+[ip-10-0-216-33:0/16] 2024-03-04 16:34:05,420 (trainer:762) INFO: 38epoch:train:11001-11500batch: iter_time=1.498e-04, forward_time=0.210, loss_ctc=42.268, loss_att=29.089, acc=0.823, loss=33.043, backward_time=0.234, grad_norm=43.688, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.628e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-04 16:41:41,735 (trainer:762) INFO: 38epoch:train:11501-12000batch: iter_time=1.506e-04, forward_time=0.209, loss_ctc=38.730, loss_att=27.157, acc=0.823, loss=30.629, backward_time=0.234, grad_norm=41.837, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.627e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-04 16:45:38,620 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 16:49:19,441 (trainer:762) INFO: 38epoch:train:12001-12500batch: iter_time=1.521e-04, forward_time=0.208, loss_ctc=38.661, loss_att=26.524, acc=0.834, loss=30.165, backward_time=0.233, grad_norm=40.833, clip=100.000, loss_scale=1.969e+33, optim_step_time=0.075, optim0_lr0=1.626e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-04 16:57:00,948 (trainer:762) INFO: 38epoch:train:12501-13000batch: iter_time=1.525e-04, forward_time=0.210, loss_ctc=39.987, loss_att=27.904, acc=0.831, loss=31.529, backward_time=0.234, grad_norm=42.071, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.626e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-04 16:59:57,585 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 17:04:39,355 (trainer:762) INFO: 38epoch:train:13001-13500batch: iter_time=1.513e-04, forward_time=0.209, loss_ctc=39.630, loss_att=27.505, acc=0.830, loss=31.142, backward_time=0.235, grad_norm=40.659, clip=100.000, loss_scale=8.975e+32, optim_step_time=0.075, optim0_lr0=1.625e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-04 17:12:14,655 (trainer:762) INFO: 38epoch:train:13501-14000batch: iter_time=1.480e-04, forward_time=0.211, loss_ctc=38.323, loss_att=26.203, acc=0.831, loss=29.839, backward_time=0.233, grad_norm=40.144, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.624e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-04 17:19:52,934 (trainer:762) INFO: 38epoch:train:14001-14500batch: iter_time=1.546e-04, forward_time=0.210, loss_ctc=41.058, loss_att=28.305, acc=0.824, loss=32.131, backward_time=0.234, grad_norm=44.927, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.623e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-04 17:26:12,397 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 17:27:23,821 (trainer:762) INFO: 38epoch:train:14501-15000batch: iter_time=1.505e-04, forward_time=0.211, loss_ctc=38.660, loss_att=26.744, acc=0.830, loss=30.319, backward_time=0.234, grad_norm=39.275, clip=100.000, loss_scale=5.970e+32, optim_step_time=0.075, optim0_lr0=1.623e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-04 17:41:34,477 (trainer:361) INFO: 38epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=39.729, loss_att=27.436, acc=0.828, loss=31.124, backward_time=0.234, grad_norm=41.916, clip=100.000, loss_scale=1.411e+33, optim_step_time=0.075, optim0_lr0=1.633e-04, train_time=1.006, time=4 hours, 11 minutes and 50.12 seconds, total_count=570000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=36.125, cer_ctc=0.195, loss_att=29.014, acc=0.761, cer=0.258, wer=0.995, loss=31.147, time=13 minutes and 54.82 seconds, total_count=81054, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-04 17:41:44,558 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-04 17:41:44,640 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/33epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-04 17:41:44,641 (trainer:290) INFO: 39/45epoch started. Estimated time to finish: 1 day, 6 hours and 57 minutes
+[ip-10-0-216-33:0/16] 2024-03-04 17:41:44,649 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 17:42:21,234 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 17:42:29,170 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf7bd340>)
+[ip-10-0-216-33:0/16] 2024-03-04 17:42:29,170 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-04 17:42:29,175 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 17:56:34,116 (trainer:762) INFO: 39epoch:train:1-500batch: iter_time=0.858, forward_time=0.210, loss_ctc=39.459, loss_att=27.234, acc=0.823, loss=30.901, backward_time=0.233, grad_norm=42.404, clip=100.000, loss_scale=3.245e+32, optim_step_time=0.075, optim0_lr0=1.622e-04, train_time=1.779
+[ip-10-0-216-33:0/16] 2024-03-04 18:04:11,025 (trainer:762) INFO: 39epoch:train:501-1000batch: iter_time=1.496e-04, forward_time=0.210, loss_ctc=41.569, loss_att=28.377, acc=0.827, loss=32.335, backward_time=0.232, grad_norm=41.558, clip=100.000, loss_scale=3.245e+32, optim_step_time=0.075, optim0_lr0=1.621e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-04 18:11:43,703 (trainer:762) INFO: 39epoch:train:1001-1500batch: iter_time=1.503e-04, forward_time=0.208, loss_ctc=41.115, loss_att=28.294, acc=0.824, loss=32.140, backward_time=0.235, grad_norm=42.230, clip=100.000, loss_scale=3.245e+32, optim_step_time=0.075, optim0_lr0=1.621e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-04 18:19:17,605 (trainer:762) INFO: 39epoch:train:1501-2000batch: iter_time=1.496e-04, forward_time=0.210, loss_ctc=40.173, loss_att=27.255, acc=0.832, loss=31.131, backward_time=0.234, grad_norm=39.595, clip=100.000, loss_scale=3.764e+32, optim_step_time=0.075, optim0_lr0=1.620e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-04 18:26:53,451 (trainer:762) INFO: 39epoch:train:2001-2500batch: iter_time=1.466e-04, forward_time=0.210, loss_ctc=39.764, loss_att=27.471, acc=0.831, loss=31.159, backward_time=0.234, grad_norm=40.816, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.619e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-04 18:34:30,326 (trainer:762) INFO: 39epoch:train:2501-3000batch: iter_time=1.463e-04, forward_time=0.209, loss_ctc=40.532, loss_att=27.573, acc=0.833, loss=31.461, backward_time=0.235, grad_norm=41.264, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.619e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-04 18:42:04,250 (trainer:762) INFO: 39epoch:train:3001-3500batch: iter_time=1.477e-04, forward_time=0.211, loss_ctc=40.270, loss_att=27.160, acc=0.836, loss=31.093, backward_time=0.229, grad_norm=41.483, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.618e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-04 18:49:41,841 (trainer:762) INFO: 39epoch:train:3501-4000batch: iter_time=1.433e-04, forward_time=0.209, loss_ctc=40.729, loss_att=27.488, acc=0.831, loss=31.460, backward_time=0.233, grad_norm=39.480, clip=100.000, loss_scale=7.529e+32, optim_step_time=0.075, optim0_lr0=1.617e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-04 18:57:15,158 (trainer:762) INFO: 39epoch:train:4001-4500batch: iter_time=1.449e-04, forward_time=0.209, loss_ctc=39.207, loss_att=27.210, acc=0.832, loss=30.809, backward_time=0.230, grad_norm=41.640, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.616e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-04 19:04:52,611 (trainer:762) INFO: 39epoch:train:4501-5000batch: iter_time=1.451e-04, forward_time=0.209, loss_ctc=38.002, loss_att=25.938, acc=0.827, loss=29.557, backward_time=0.242, grad_norm=40.422, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.616e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-04 19:04:58,940 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 19:05:35,157 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 19:05:43,270 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecffb8a60>)
+[ip-10-0-216-33:0/16] 2024-03-04 19:05:43,270 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-04 19:05:43,275 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 19:20:14,572 (trainer:762) INFO: 39epoch:train:5001-5500batch: iter_time=0.887, forward_time=0.208, loss_ctc=39.225, loss_att=27.219, acc=0.824, loss=30.821, backward_time=0.233, grad_norm=42.591, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.615e-04, train_time=1.844
+[ip-10-0-216-33:0/16] 2024-03-04 19:27:46,722 (trainer:762) INFO: 39epoch:train:5501-6000batch: iter_time=1.449e-04, forward_time=0.210, loss_ctc=41.163, loss_att=28.139, acc=0.827, loss=32.046, backward_time=0.231, grad_norm=41.204, clip=100.000, loss_scale=1.506e+33, optim_step_time=0.075, optim0_lr0=1.614e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-04 19:35:19,795 (trainer:762) INFO: 39epoch:train:6001-6500batch: iter_time=1.478e-04, forward_time=0.208, loss_ctc=41.068, loss_att=28.190, acc=0.823, loss=32.054, backward_time=0.232, grad_norm=42.628, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.614e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-04 19:42:52,508 (trainer:762) INFO: 39epoch:train:6501-7000batch: iter_time=1.436e-04, forward_time=0.210, loss_ctc=39.964, loss_att=27.142, acc=0.833, loss=30.989, backward_time=0.232, grad_norm=40.359, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.613e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-04 19:50:23,838 (trainer:762) INFO: 39epoch:train:7001-7500batch: iter_time=1.458e-04, forward_time=0.209, loss_ctc=39.603, loss_att=27.334, acc=0.831, loss=31.014, backward_time=0.233, grad_norm=40.170, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.612e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-04 19:57:56,652 (trainer:762) INFO: 39epoch:train:7501-8000batch: iter_time=1.459e-04, forward_time=0.210, loss_ctc=40.508, loss_att=27.519, acc=0.833, loss=31.416, backward_time=0.235, grad_norm=41.032, clip=100.000, loss_scale=3.012e+33, optim_step_time=0.075, optim0_lr0=1.611e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-04 20:05:33,207 (trainer:762) INFO: 39epoch:train:8001-8500batch: iter_time=1.423e-04, forward_time=0.208, loss_ctc=40.224, loss_att=27.110, acc=0.835, loss=31.044, backward_time=0.236, grad_norm=39.942, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.611e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-04 20:08:25,000 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 20:13:08,734 (trainer:762) INFO: 39epoch:train:8501-9000batch: iter_time=1.453e-04, forward_time=0.209, loss_ctc=40.613, loss_att=27.424, acc=0.832, loss=31.381, backward_time=0.235, grad_norm=38.063, clip=100.000, loss_scale=3.559e+33, optim_step_time=0.075, optim0_lr0=1.610e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-04 20:20:49,302 (trainer:762) INFO: 39epoch:train:9001-9500batch: iter_time=1.480e-04, forward_time=0.209, loss_ctc=38.744, loss_att=27.109, acc=0.832, loss=30.600, backward_time=0.232, grad_norm=40.034, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.609e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-04 20:28:25,459 (trainer:762) INFO: 39epoch:train:9501-10000batch: iter_time=1.480e-04, forward_time=0.210, loss_ctc=37.835, loss_att=25.929, acc=0.827, loss=29.501, backward_time=0.235, grad_norm=40.395, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.609e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-04 20:28:31,187 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 20:29:08,453 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 20:29:16,265 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e865b1c40>)
+[ip-10-0-216-33:0/16] 2024-03-04 20:29:16,266 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-04 20:29:16,270 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 20:43:42,446 (trainer:762) INFO: 39epoch:train:10001-10500batch: iter_time=0.881, forward_time=0.210, loss_ctc=39.075, loss_att=27.084, acc=0.824, loss=30.681, backward_time=0.234, grad_norm=42.437, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.608e-04, train_time=1.834
+[ip-10-0-216-33:0/16] 2024-03-04 20:46:59,891 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 20:51:20,786 (trainer:762) INFO: 39epoch:train:10501-11000batch: iter_time=1.498e-04, forward_time=0.208, loss_ctc=41.007, loss_att=28.060, acc=0.828, loss=31.944, backward_time=0.234, grad_norm=40.908, clip=100.000, loss_scale=2.742e+33, optim_step_time=0.075, optim0_lr0=1.607e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-04 20:58:59,039 (trainer:762) INFO: 39epoch:train:11001-11500batch: iter_time=1.559e-04, forward_time=0.209, loss_ctc=40.932, loss_att=28.179, acc=0.824, loss=32.005, backward_time=0.230, grad_norm=42.364, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.076, optim0_lr0=1.607e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-04 21:06:35,889 (trainer:762) INFO: 39epoch:train:11501-12000batch: iter_time=1.484e-04, forward_time=0.209, loss_ctc=39.957, loss_att=27.137, acc=0.833, loss=30.983, backward_time=0.232, grad_norm=39.483, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.606e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-04 21:14:16,331 (trainer:762) INFO: 39epoch:train:12001-12500batch: iter_time=1.518e-04, forward_time=0.209, loss_ctc=39.337, loss_att=27.162, acc=0.832, loss=30.815, backward_time=0.235, grad_norm=40.390, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.605e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-04 21:18:05,613 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 21:21:57,076 (trainer:762) INFO: 39epoch:train:12501-13000batch: iter_time=1.474e-04, forward_time=0.208, loss_ctc=40.062, loss_att=27.318, acc=0.834, loss=31.141, backward_time=0.233, grad_norm=41.114, clip=100.000, loss_scale=2.763e+33, optim_step_time=0.075, optim0_lr0=1.605e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-04 21:26:39,281 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 21:29:38,306 (trainer:762) INFO: 39epoch:train:13001-13500batch: iter_time=1.472e-04, forward_time=0.209, loss_ctc=40.104, loss_att=27.087, acc=0.836, loss=30.992, backward_time=0.233, grad_norm=40.529, clip=100.000, loss_scale=2.091e+33, optim_step_time=0.075, optim0_lr0=1.604e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-04 21:37:20,443 (trainer:762) INFO: 39epoch:train:13501-14000batch: iter_time=1.504e-04, forward_time=0.206, loss_ctc=40.234, loss_att=27.242, acc=0.832, loss=31.139, backward_time=0.232, grad_norm=38.084, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.603e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-04 21:45:02,101 (trainer:762) INFO: 39epoch:train:14001-14500batch: iter_time=1.445e-04, forward_time=0.208, loss_ctc=38.758, loss_att=27.093, acc=0.833, loss=30.592, backward_time=0.230, grad_norm=39.951, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.603e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-04 21:52:36,036 (trainer:762) INFO: 39epoch:train:14501-15000batch: iter_time=1.497e-04, forward_time=0.209, loss_ctc=37.658, loss_att=25.767, acc=0.828, loss=29.335, backward_time=0.237, grad_norm=39.813, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.602e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-04 22:06:52,299 (trainer:361) INFO: 39epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=39.896, loss_att=27.308, acc=0.830, loss=31.085, backward_time=0.233, grad_norm=40.746, clip=100.000, loss_scale=1.869e+33, optim_step_time=0.075, optim0_lr0=1.612e-04, train_time=1.003, time=4 hours, 11 minutes and 8.02 seconds, total_count=585000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=35.644, cer_ctc=0.193, loss_att=28.599, acc=0.775, cer=0.211, wer=0.983, loss=30.713, time=13 minutes and 59.33 seconds, total_count=83187, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-04 22:07:02,032 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-04 22:07:02,083 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/29epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-04 22:07:02,083 (trainer:290) INFO: 40/45epoch started. Estimated time to finish: 1 day, 2 hours and 32 minutes
+[ip-10-0-216-33:0/16] 2024-03-04 22:07:02,092 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 22:07:37,734 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 22:07:46,216 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf6d42b0>)
+[ip-10-0-216-33:0/16] 2024-03-04 22:07:46,216 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-04 22:07:46,221 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 22:21:44,820 (trainer:762) INFO: 40epoch:train:1-500batch: iter_time=0.839, forward_time=0.209, loss_ctc=43.383, loss_att=30.030, acc=0.822, loss=34.036, backward_time=0.234, grad_norm=44.503, clip=100.000, loss_scale=1.802e+33, optim_step_time=0.075, optim0_lr0=1.601e-04, train_time=1.765
+[ip-10-0-216-33:0/16] 2024-03-04 22:29:23,310 (trainer:762) INFO: 40epoch:train:501-1000batch: iter_time=1.503e-04, forward_time=0.209, loss_ctc=38.687, loss_att=26.347, acc=0.833, loss=30.049, backward_time=0.230, grad_norm=41.167, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.600e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-04 22:37:04,186 (trainer:762) INFO: 40epoch:train:1001-1500batch: iter_time=1.476e-04, forward_time=0.210, loss_ctc=40.140, loss_att=27.878, acc=0.834, loss=31.556, backward_time=0.234, grad_norm=41.503, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.600e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-04 22:44:44,273 (trainer:762) INFO: 40epoch:train:1501-2000batch: iter_time=1.463e-04, forward_time=0.209, loss_ctc=37.733, loss_att=26.367, acc=0.825, loss=29.777, backward_time=0.238, grad_norm=44.829, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.599e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-04 22:52:18,803 (trainer:762) INFO: 40epoch:train:2001-2500batch: iter_time=1.444e-04, forward_time=0.208, loss_ctc=37.950, loss_att=25.831, acc=0.833, loss=29.467, backward_time=0.235, grad_norm=38.988, clip=100.000, loss_scale=3.603e+33, optim_step_time=0.075, optim0_lr0=1.598e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-04 22:59:53,859 (trainer:762) INFO: 40epoch:train:2501-3000batch: iter_time=1.498e-04, forward_time=0.210, loss_ctc=38.357, loss_att=26.767, acc=0.833, loss=30.244, backward_time=0.235, grad_norm=42.169, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.598e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-04 23:07:32,614 (trainer:762) INFO: 40epoch:train:3001-3500batch: iter_time=1.475e-04, forward_time=0.210, loss_ctc=39.232, loss_att=27.175, acc=0.835, loss=30.792, backward_time=0.234, grad_norm=41.754, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.597e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-04 23:13:38,483 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-04 23:15:14,239 (trainer:762) INFO: 40epoch:train:3501-4000batch: iter_time=1.470e-04, forward_time=0.208, loss_ctc=41.650, loss_att=28.259, acc=0.828, loss=32.276, backward_time=0.232, grad_norm=42.106, clip=100.000, loss_scale=4.656e+33, optim_step_time=0.075, optim0_lr0=1.596e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-04 23:22:55,321 (trainer:762) INFO: 40epoch:train:4001-4500batch: iter_time=1.458e-04, forward_time=0.209, loss_ctc=39.991, loss_att=27.659, acc=0.828, loss=31.359, backward_time=0.237, grad_norm=44.580, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.596e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-04 23:30:34,393 (trainer:762) INFO: 40epoch:train:4501-5000batch: iter_time=1.433e-04, forward_time=0.208, loss_ctc=38.014, loss_att=25.799, acc=0.834, loss=29.463, backward_time=0.231, grad_norm=39.597, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.595e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-04 23:30:40,158 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-04 23:31:16,447 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-04 23:31:25,162 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf9a72e0>)
+[ip-10-0-216-33:0/16] 2024-03-04 23:31:25,163 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-04 23:31:25,169 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-04 23:45:55,087 (trainer:762) INFO: 40epoch:train:5001-5500batch: iter_time=0.882, forward_time=0.207, loss_ctc=42.761, loss_att=29.533, acc=0.824, loss=33.501, backward_time=0.241, grad_norm=49.138, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.594e-04, train_time=1.841
+[ip-10-0-216-33:0/16] 2024-03-04 23:53:27,874 (trainer:762) INFO: 40epoch:train:5501-6000batch: iter_time=1.452e-04, forward_time=0.210, loss_ctc=38.205, loss_att=26.052, acc=0.834, loss=29.698, backward_time=0.233, grad_norm=40.437, clip=100.000, loss_scale=3.131e+33, optim_step_time=0.075, optim0_lr0=1.594e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-05 00:01:00,399 (trainer:762) INFO: 40epoch:train:6001-6500batch: iter_time=1.456e-04, forward_time=0.210, loss_ctc=40.044, loss_att=27.824, acc=0.834, loss=31.490, backward_time=0.238, grad_norm=41.842, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.593e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-05 00:07:03,599 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 00:08:34,631 (trainer:762) INFO: 40epoch:train:6501-7000batch: iter_time=1.456e-04, forward_time=0.208, loss_ctc=37.371, loss_att=26.102, acc=0.826, loss=29.483, backward_time=0.237, grad_norm=43.885, clip=100.000, loss_scale=4.672e+33, optim_step_time=0.075, optim0_lr0=1.592e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-05 00:16:09,697 (trainer:762) INFO: 40epoch:train:7001-7500batch: iter_time=1.481e-04, forward_time=0.209, loss_ctc=38.051, loss_att=25.903, acc=0.833, loss=29.547, backward_time=0.233, grad_norm=39.799, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.592e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-05 00:23:43,453 (trainer:762) INFO: 40epoch:train:7501-8000batch: iter_time=1.505e-04, forward_time=0.208, loss_ctc=38.217, loss_att=26.841, acc=0.833, loss=30.254, backward_time=0.232, grad_norm=41.681, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.591e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-05 00:31:20,442 (trainer:762) INFO: 40epoch:train:8001-8500batch: iter_time=1.436e-04, forward_time=0.210, loss_ctc=38.965, loss_att=27.024, acc=0.836, loss=30.606, backward_time=0.232, grad_norm=41.773, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.590e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-05 00:38:53,581 (trainer:762) INFO: 40epoch:train:8501-9000batch: iter_time=1.488e-04, forward_time=0.208, loss_ctc=41.237, loss_att=28.035, acc=0.829, loss=31.996, backward_time=0.235, grad_norm=41.997, clip=100.000, loss_scale=3.115e+33, optim_step_time=0.075, optim0_lr0=1.590e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-05 00:39:21,979 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 00:46:25,897 (trainer:762) INFO: 40epoch:train:9001-9500batch: iter_time=1.453e-04, forward_time=0.210, loss_ctc=39.667, loss_att=27.459, acc=0.829, loss=31.121, backward_time=0.231, grad_norm=42.456, clip=100.000, loss_scale=2.752e+33, optim_step_time=0.075, optim0_lr0=1.589e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-05 00:53:58,244 (trainer:762) INFO: 40epoch:train:9501-10000batch: iter_time=1.445e-04, forward_time=0.208, loss_ctc=37.814, loss_att=25.597, acc=0.835, loss=29.262, backward_time=0.235, grad_norm=38.809, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.588e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-05 00:54:03,357 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 00:54:40,025 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 00:54:48,047 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5eb2221f10>)
+[ip-10-0-216-33:0/16] 2024-03-05 00:54:48,047 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-05 00:54:48,052 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 01:09:38,296 (trainer:762) INFO: 40epoch:train:10001-10500batch: iter_time=0.922, forward_time=0.210, loss_ctc=42.581, loss_att=29.397, acc=0.825, loss=33.352, backward_time=0.236, grad_norm=42.766, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.588e-04, train_time=1.880
+[ip-10-0-216-33:0/16] 2024-03-05 01:17:13,910 (trainer:762) INFO: 40epoch:train:10501-11000batch: iter_time=1.504e-04, forward_time=0.211, loss_ctc=38.100, loss_att=26.071, acc=0.834, loss=29.680, backward_time=0.231, grad_norm=40.923, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.587e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 01:19:33,199 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 01:19:45,871 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 01:24:51,814 (trainer:762) INFO: 40epoch:train:11001-11500batch: iter_time=1.523e-04, forward_time=0.208, loss_ctc=39.865, loss_att=27.672, acc=0.835, loss=31.330, backward_time=0.233, grad_norm=41.858, clip=100.000, loss_scale=2.359e+33, optim_step_time=0.075, optim0_lr0=1.586e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-05 01:32:31,807 (trainer:762) INFO: 40epoch:train:11501-12000batch: iter_time=1.485e-04, forward_time=0.210, loss_ctc=37.302, loss_att=26.101, acc=0.826, loss=29.461, backward_time=0.232, grad_norm=42.908, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.586e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-05 01:40:12,173 (trainer:762) INFO: 40epoch:train:12001-12500batch: iter_time=1.547e-04, forward_time=0.208, loss_ctc=37.717, loss_att=25.733, acc=0.834, loss=29.328, backward_time=0.239, grad_norm=41.070, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.585e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-05 01:47:46,141 (trainer:762) INFO: 40epoch:train:12501-13000batch: iter_time=1.513e-04, forward_time=0.210, loss_ctc=38.303, loss_att=26.869, acc=0.833, loss=30.299, backward_time=0.238, grad_norm=42.400, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.584e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-05 01:55:24,124 (trainer:762) INFO: 40epoch:train:13001-13500batch: iter_time=1.519e-04, forward_time=0.209, loss_ctc=38.880, loss_att=27.002, acc=0.836, loss=30.565, backward_time=0.234, grad_norm=41.320, clip=100.000, loss_scale=2.163e+33, optim_step_time=0.075, optim0_lr0=1.584e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-05 02:02:59,970 (trainer:762) INFO: 40epoch:train:13501-14000batch: iter_time=1.494e-04, forward_time=0.210, loss_ctc=41.363, loss_att=28.005, acc=0.829, loss=32.012, backward_time=0.238, grad_norm=41.584, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.583e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 02:10:31,382 (trainer:762) INFO: 40epoch:train:14001-14500batch: iter_time=1.468e-04, forward_time=0.210, loss_ctc=39.620, loss_att=27.361, acc=0.829, loss=31.039, backward_time=0.233, grad_norm=42.890, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.582e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-05 02:18:05,844 (trainer:762) INFO: 40epoch:train:14501-15000batch: iter_time=1.488e-04, forward_time=0.210, loss_ctc=37.562, loss_att=25.482, acc=0.835, loss=29.106, backward_time=0.235, grad_norm=39.013, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.076, optim0_lr0=1.582e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-05 02:32:19,189 (trainer:361) INFO: 40epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=39.292, loss_att=27.072, acc=0.831, loss=30.738, backward_time=0.235, grad_norm=41.991, clip=100.000, loss_scale=2.889e+33, optim_step_time=0.075, optim0_lr0=1.591e-04, train_time=1.004, time=4 hours, 11 minutes and 19.9 seconds, total_count=600000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=35.747, cer_ctc=0.193, loss_att=28.133, acc=0.771, cer=0.228, wer=0.989, loss=30.417, time=13 minutes and 56.91 seconds, total_count=85320, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-05 02:32:29,033 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-05 02:32:29,078 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/35epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-05 02:32:29,079 (trainer:290) INFO: 41/45epoch started. Estimated time to finish: 22 hours, 6 minutes and 42.16 seconds
+[ip-10-0-216-33:0/16] 2024-03-05 02:32:29,086 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 02:33:04,920 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 02:33:13,291 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfe6aeb0>)
+[ip-10-0-216-33:0/16] 2024-03-05 02:33:13,291 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-05 02:33:13,295 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 02:47:09,139 (trainer:762) INFO: 41epoch:train:1-500batch: iter_time=0.834, forward_time=0.209, loss_ctc=38.423, loss_att=26.354, acc=0.833, loss=29.974, backward_time=0.234, grad_norm=40.757, clip=100.000, loss_scale=4.325e+33, optim_step_time=0.075, optim0_lr0=1.581e-04, train_time=1.760
+[ip-10-0-216-33:0/16] 2024-03-05 02:47:12,720 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 02:53:56,846 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 02:54:50,789 (trainer:762) INFO: 41epoch:train:501-1000batch: iter_time=1.506e-04, forward_time=0.209, loss_ctc=38.293, loss_att=26.347, acc=0.837, loss=29.931, backward_time=0.234, grad_norm=41.973, clip=100.000, loss_scale=2.455e+33, optim_step_time=0.075, optim0_lr0=1.580e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-05 02:57:48,730 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 03:02:30,635 (trainer:762) INFO: 41epoch:train:1001-1500batch: iter_time=1.504e-04, forward_time=0.209, loss_ctc=37.140, loss_att=26.289, acc=0.831, loss=29.544, backward_time=0.238, grad_norm=41.560, clip=100.000, loss_scale=9.001e+32, optim_step_time=0.075, optim0_lr0=1.580e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-05 03:10:13,120 (trainer:762) INFO: 41epoch:train:1501-2000batch: iter_time=1.482e-04, forward_time=0.209, loss_ctc=40.464, loss_att=27.479, acc=0.833, loss=31.375, backward_time=0.238, grad_norm=42.544, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.579e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-05 03:17:54,059 (trainer:762) INFO: 41epoch:train:2001-2500batch: iter_time=1.493e-04, forward_time=0.206, loss_ctc=38.995, loss_att=26.783, acc=0.827, loss=30.447, backward_time=0.233, grad_norm=42.769, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.578e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-05 03:25:39,205 (trainer:762) INFO: 41epoch:train:2501-3000batch: iter_time=1.498e-04, forward_time=0.208, loss_ctc=39.537, loss_att=27.126, acc=0.831, loss=30.849, backward_time=0.235, grad_norm=41.892, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.578e-04, train_time=0.930
+[ip-10-0-216-33:0/16] 2024-03-05 03:33:18,623 (trainer:762) INFO: 41epoch:train:3001-3500batch: iter_time=1.461e-04, forward_time=0.208, loss_ctc=39.080, loss_att=26.914, acc=0.832, loss=30.563, backward_time=0.235, grad_norm=43.834, clip=100.000, loss_scale=1.046e+33, optim_step_time=0.075, optim0_lr0=1.577e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-05 03:34:32,573 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 03:40:53,870 (trainer:762) INFO: 41epoch:train:3501-4000batch: iter_time=1.470e-04, forward_time=0.209, loss_ctc=38.533, loss_att=26.562, acc=0.832, loss=30.153, backward_time=0.231, grad_norm=39.533, clip=100.000, loss_scale=7.518e+32, optim_step_time=0.075, optim0_lr0=1.576e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-05 03:48:31,599 (trainer:762) INFO: 41epoch:train:4001-4500batch: iter_time=1.468e-04, forward_time=0.209, loss_ctc=41.329, loss_att=28.429, acc=0.829, loss=32.299, backward_time=0.236, grad_norm=42.234, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.576e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-05 03:56:05,152 (trainer:762) INFO: 41epoch:train:4501-5000batch: iter_time=1.439e-04, forward_time=0.209, loss_ctc=38.907, loss_att=26.650, acc=0.829, loss=30.327, backward_time=0.234, grad_norm=41.096, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.575e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-05 03:56:10,960 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 03:56:47,620 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 03:56:55,452 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ea70c7a60>)
+[ip-10-0-216-33:0/16] 2024-03-05 03:56:55,452 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-05 03:56:55,457 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 04:11:27,539 (trainer:762) INFO: 41epoch:train:5001-5500batch: iter_time=0.879, forward_time=0.208, loss_ctc=38.020, loss_att=25.988, acc=0.834, loss=29.598, backward_time=0.234, grad_norm=40.255, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.574e-04, train_time=1.845
+[ip-10-0-216-33:0/16] 2024-03-05 04:19:11,826 (trainer:762) INFO: 41epoch:train:5501-6000batch: iter_time=1.448e-04, forward_time=0.208, loss_ctc=38.116, loss_att=26.283, acc=0.837, loss=29.833, backward_time=0.238, grad_norm=41.309, clip=100.000, loss_scale=1.194e+33, optim_step_time=0.075, optim0_lr0=1.574e-04, train_time=0.928
+[ip-10-0-216-33:0/16] 2024-03-05 04:26:49,213 (trainer:762) INFO: 41epoch:train:6001-6500batch: iter_time=1.485e-04, forward_time=0.208, loss_ctc=36.848, loss_att=26.066, acc=0.833, loss=29.301, backward_time=0.232, grad_norm=41.032, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.573e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-05 04:34:23,944 (trainer:762) INFO: 41epoch:train:6501-7000batch: iter_time=1.447e-04, forward_time=0.210, loss_ctc=40.276, loss_att=27.287, acc=0.834, loss=31.183, backward_time=0.233, grad_norm=41.773, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.573e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-05 04:41:58,871 (trainer:762) INFO: 41epoch:train:7001-7500batch: iter_time=1.519e-04, forward_time=0.209, loss_ctc=38.861, loss_att=26.659, acc=0.828, loss=30.320, backward_time=0.236, grad_norm=44.365, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.572e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-05 04:49:41,496 (trainer:762) INFO: 41epoch:train:7501-8000batch: iter_time=1.485e-04, forward_time=0.207, loss_ctc=39.366, loss_att=27.097, acc=0.831, loss=30.778, backward_time=0.234, grad_norm=42.838, clip=100.000, loss_scale=2.388e+33, optim_step_time=0.075, optim0_lr0=1.571e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-05 04:57:29,945 (trainer:762) INFO: 41epoch:train:8001-8500batch: iter_time=1.432e-04, forward_time=0.208, loss_ctc=38.798, loss_att=26.757, acc=0.832, loss=30.369, backward_time=0.235, grad_norm=41.589, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.571e-04, train_time=0.937
+[ip-10-0-216-33:0/16] 2024-03-05 05:05:12,299 (trainer:762) INFO: 41epoch:train:8501-9000batch: iter_time=1.493e-04, forward_time=0.206, loss_ctc=38.263, loss_att=26.464, acc=0.832, loss=30.003, backward_time=0.231, grad_norm=38.747, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.570e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-05 05:12:48,856 (trainer:762) INFO: 41epoch:train:9001-9500batch: iter_time=1.428e-04, forward_time=0.210, loss_ctc=41.209, loss_att=28.298, acc=0.830, loss=32.171, backward_time=0.232, grad_norm=41.304, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.569e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-05 05:20:20,071 (trainer:762) INFO: 41epoch:train:9501-10000batch: iter_time=1.450e-04, forward_time=0.209, loss_ctc=38.883, loss_att=26.571, acc=0.830, loss=30.265, backward_time=0.231, grad_norm=39.885, clip=100.000, loss_scale=4.777e+33, optim_step_time=0.075, optim0_lr0=1.569e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-05 05:20:26,027 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 05:21:02,565 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 05:21:10,417 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e8a856880>)
+[ip-10-0-216-33:0/16] 2024-03-05 05:21:10,417 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-05 05:21:10,422 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 05:35:58,818 (trainer:762) INFO: 41epoch:train:10001-10500batch: iter_time=0.884, forward_time=0.209, loss_ctc=37.944, loss_att=25.992, acc=0.835, loss=29.578, backward_time=0.230, grad_norm=40.022, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.568e-04, train_time=1.877
+[ip-10-0-216-33:0/16] 2024-03-05 05:43:35,811 (trainer:762) INFO: 41epoch:train:10501-11000batch: iter_time=1.473e-04, forward_time=0.208, loss_ctc=37.974, loss_att=26.109, acc=0.839, loss=29.668, backward_time=0.230, grad_norm=40.262, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.567e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-05 05:51:14,350 (trainer:762) INFO: 41epoch:train:11001-11500batch: iter_time=1.537e-04, forward_time=0.207, loss_ctc=36.847, loss_att=26.041, acc=0.832, loss=29.283, backward_time=0.240, grad_norm=41.606, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.567e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-05 05:55:29,732 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 05:58:52,393 (trainer:762) INFO: 41epoch:train:11501-12000batch: iter_time=1.480e-04, forward_time=0.209, loss_ctc=40.149, loss_att=27.192, acc=0.834, loss=31.079, backward_time=0.232, grad_norm=44.454, clip=100.000, loss_scale=7.232e+33, optim_step_time=0.075, optim0_lr0=1.566e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-05 06:06:27,135 (trainer:762) INFO: 41epoch:train:12001-12500batch: iter_time=1.511e-04, forward_time=0.210, loss_ctc=38.785, loss_att=26.574, acc=0.829, loss=30.237, backward_time=0.240, grad_norm=43.144, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.565e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-05 06:14:02,564 (trainer:762) INFO: 41epoch:train:12501-13000batch: iter_time=1.489e-04, forward_time=0.209, loss_ctc=39.217, loss_att=26.959, acc=0.832, loss=30.637, backward_time=0.235, grad_norm=42.044, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.565e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 06:21:36,009 (trainer:762) INFO: 41epoch:train:13001-13500batch: iter_time=1.483e-04, forward_time=0.208, loss_ctc=38.811, loss_att=26.738, acc=0.833, loss=30.360, backward_time=0.233, grad_norm=40.549, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.564e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-05 06:29:15,474 (trainer:762) INFO: 41epoch:train:13501-14000batch: iter_time=1.502e-04, forward_time=0.210, loss_ctc=38.492, loss_att=26.470, acc=0.832, loss=30.076, backward_time=0.238, grad_norm=38.974, clip=100.000, loss_scale=7.508e+33, optim_step_time=0.075, optim0_lr0=1.564e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-05 06:30:43,882 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 06:36:50,470 (trainer:762) INFO: 41epoch:train:14001-14500batch: iter_time=1.505e-04, forward_time=0.209, loss_ctc=40.770, loss_att=28.214, acc=0.831, loss=31.981, backward_time=0.233, grad_norm=41.266, clip=100.000, loss_scale=6.202e+33, optim_step_time=0.075, optim0_lr0=1.563e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-05 06:39:56,664 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 06:44:30,699 (trainer:762) INFO: 41epoch:train:14501-15000batch: iter_time=1.454e-04, forward_time=0.207, loss_ctc=38.709, loss_att=26.554, acc=0.830, loss=30.200, backward_time=0.236, grad_norm=40.028, clip=100.000, loss_scale=3.642e+33, optim_step_time=0.075, optim0_lr0=1.562e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-05 06:58:40,784 (trainer:361) INFO: 41epoch results: [train] iter_time=0.087, forward_time=0.208, loss_ctc=38.901, loss_att=26.775, acc=0.832, loss=30.413, backward_time=0.234, grad_norm=41.454, clip=100.000, loss_scale=2.972e+33, optim_step_time=0.075, optim0_lr0=1.572e-04, train_time=1.008, time=4 hours, 12 minutes and 17.14 seconds, total_count=615000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=34.715, cer_ctc=0.188, loss_att=27.154, acc=0.780, cer=0.212, wer=0.989, loss=29.422, time=13 minutes and 54.23 seconds, total_count=87453, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-05 06:58:50,775 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-05 06:58:50,831 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/32epoch.pth, exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/36epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-05 06:58:50,831 (trainer:290) INFO: 42/45epoch started. Estimated time to finish: 17 hours, 41 minutes and 27.86 seconds
+[ip-10-0-216-33:0/16] 2024-03-05 06:58:50,838 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 06:59:26,599 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 06:59:34,353 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecfcfbfd0>)
+[ip-10-0-216-33:0/16] 2024-03-05 06:59:34,354 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-05 06:59:34,358 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 07:13:34,497 (trainer:762) INFO: 42epoch:train:1-500batch: iter_time=0.847, forward_time=0.210, loss_ctc=40.042, loss_att=27.652, acc=0.835, loss=31.369, backward_time=0.231, grad_norm=43.584, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.562e-04, train_time=1.767
+[ip-10-0-216-33:0/16] 2024-03-05 07:19:26,979 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 07:21:10,867 (trainer:762) INFO: 42epoch:train:501-1000batch: iter_time=1.485e-04, forward_time=0.210, loss_ctc=41.029, loss_att=27.692, acc=0.824, loss=31.693, backward_time=0.233, grad_norm=43.151, clip=100.000, loss_scale=2.294e+33, optim_step_time=0.075, optim0_lr0=1.561e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-05 07:28:50,572 (trainer:762) INFO: 42epoch:train:1001-1500batch: iter_time=1.494e-04, forward_time=0.209, loss_ctc=39.558, loss_att=27.365, acc=0.829, loss=31.023, backward_time=0.232, grad_norm=44.833, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.560e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-05 07:36:30,472 (trainer:762) INFO: 42epoch:train:1501-2000batch: iter_time=1.478e-04, forward_time=0.209, loss_ctc=39.976, loss_att=27.094, acc=0.827, loss=30.959, backward_time=0.236, grad_norm=45.195, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.560e-04, train_time=0.920
+[ip-10-0-216-33:0/16] 2024-03-05 07:44:03,553 (trainer:762) INFO: 42epoch:train:2001-2500batch: iter_time=1.515e-04, forward_time=0.210, loss_ctc=37.847, loss_att=26.205, acc=0.830, loss=29.697, backward_time=0.234, grad_norm=41.558, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.559e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-05 07:50:57,748 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 07:51:44,375 (trainer:762) INFO: 42epoch:train:2501-3000batch: iter_time=1.512e-04, forward_time=0.209, loss_ctc=41.090, loss_att=28.012, acc=0.830, loss=31.935, backward_time=0.233, grad_norm=44.410, clip=100.000, loss_scale=1.465e+33, optim_step_time=0.075, optim0_lr0=1.558e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-05 07:57:53,577 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 07:59:22,932 (trainer:762) INFO: 42epoch:train:3001-3500batch: iter_time=1.476e-04, forward_time=0.209, loss_ctc=39.146, loss_att=27.130, acc=0.827, loss=30.735, backward_time=0.229, grad_norm=43.068, clip=100.000, loss_scale=1.173e+33, optim_step_time=0.075, optim0_lr0=1.558e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-05 08:07:06,119 (trainer:762) INFO: 42epoch:train:3501-4000batch: iter_time=1.518e-04, forward_time=0.208, loss_ctc=38.728, loss_att=26.485, acc=0.835, loss=30.158, backward_time=0.240, grad_norm=40.836, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.557e-04, train_time=0.926
+[ip-10-0-216-33:0/16] 2024-03-05 08:14:50,416 (trainer:762) INFO: 42epoch:train:4001-4500batch: iter_time=1.511e-04, forward_time=0.207, loss_ctc=40.766, loss_att=28.508, acc=0.822, loss=32.185, backward_time=0.235, grad_norm=45.488, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.557e-04, train_time=0.928
+[ip-10-0-216-33:0/16] 2024-03-05 08:22:30,245 (trainer:762) INFO: 42epoch:train:4501-5000batch: iter_time=1.432e-04, forward_time=0.208, loss_ctc=40.700, loss_att=27.513, acc=0.831, loss=31.469, backward_time=0.233, grad_norm=41.597, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.556e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-05 08:22:35,123 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 08:23:11,449 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 08:23:19,376 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecffcec70>)
+[ip-10-0-216-33:0/16] 2024-03-05 08:23:19,377 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-05 08:23:19,381 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 08:37:43,841 (trainer:762) INFO: 42epoch:train:5001-5500batch: iter_time=0.892, forward_time=0.210, loss_ctc=39.819, loss_att=27.488, acc=0.835, loss=31.187, backward_time=0.234, grad_norm=44.327, clip=100.000, loss_scale=7.737e+32, optim_step_time=0.075, optim0_lr0=1.555e-04, train_time=1.827
+[ip-10-0-216-33:0/16] 2024-03-05 08:45:22,779 (trainer:762) INFO: 42epoch:train:5501-6000batch: iter_time=1.502e-04, forward_time=0.207, loss_ctc=40.661, loss_att=27.562, acc=0.825, loss=31.492, backward_time=0.233, grad_norm=43.590, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.555e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-05 08:52:53,738 (trainer:762) INFO: 42epoch:train:6001-6500batch: iter_time=1.481e-04, forward_time=0.210, loss_ctc=39.343, loss_att=27.233, acc=0.830, loss=30.866, backward_time=0.232, grad_norm=45.172, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.554e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-05 09:00:31,281 (trainer:762) INFO: 42epoch:train:6501-7000batch: iter_time=1.480e-04, forward_time=0.209, loss_ctc=39.734, loss_att=27.019, acc=0.828, loss=30.833, backward_time=0.233, grad_norm=47.081, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.553e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-05 09:08:01,314 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 09:08:07,637 (trainer:762) INFO: 42epoch:train:7001-7500batch: iter_time=1.503e-04, forward_time=0.208, loss_ctc=37.437, loss_att=26.146, acc=0.831, loss=29.533, backward_time=0.236, grad_norm=41.006, clip=100.000, loss_scale=1.527e+33, optim_step_time=0.075, optim0_lr0=1.553e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-05 09:15:43,088 (trainer:762) INFO: 42epoch:train:7501-8000batch: iter_time=1.476e-04, forward_time=0.210, loss_ctc=40.482, loss_att=27.808, acc=0.830, loss=31.610, backward_time=0.233, grad_norm=44.436, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.552e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 09:23:20,726 (trainer:762) INFO: 42epoch:train:8001-8500batch: iter_time=1.547e-04, forward_time=0.208, loss_ctc=38.847, loss_att=26.959, acc=0.827, loss=30.525, backward_time=0.230, grad_norm=43.540, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.552e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-05 09:23:50,572 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 09:31:01,886 (trainer:762) INFO: 42epoch:train:8501-9000batch: iter_time=1.542e-04, forward_time=0.209, loss_ctc=38.493, loss_att=26.342, acc=0.836, loss=29.988, backward_time=0.238, grad_norm=41.801, clip=100.000, loss_scale=6.894e+32, optim_step_time=0.075, optim0_lr0=1.551e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-05 09:38:41,309 (trainer:762) INFO: 42epoch:train:9001-9500batch: iter_time=1.541e-04, forward_time=0.207, loss_ctc=40.431, loss_att=28.420, acc=0.822, loss=32.023, backward_time=0.230, grad_norm=45.819, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.550e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-05 09:46:19,460 (trainer:762) INFO: 42epoch:train:9501-10000batch: iter_time=1.496e-04, forward_time=0.209, loss_ctc=40.545, loss_att=27.431, acc=0.832, loss=31.365, backward_time=0.231, grad_norm=41.805, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.550e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-05 09:46:24,265 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 09:47:01,215 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 09:47:09,157 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e7f855520>)
+[ip-10-0-216-33:0/16] 2024-03-05 09:47:09,158 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-05 09:47:09,162 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 10:01:33,389 (trainer:762) INFO: 42epoch:train:10001-10500batch: iter_time=0.897, forward_time=0.211, loss_ctc=39.714, loss_att=27.429, acc=0.835, loss=31.115, backward_time=0.237, grad_norm=43.640, clip=100.000, loss_scale=6.490e+32, optim_step_time=0.075, optim0_lr0=1.549e-04, train_time=1.828
+[ip-10-0-216-33:0/16] 2024-03-05 10:09:10,266 (trainer:762) INFO: 42epoch:train:10501-11000batch: iter_time=1.472e-04, forward_time=0.208, loss_ctc=40.454, loss_att=27.481, acc=0.825, loss=31.372, backward_time=0.238, grad_norm=42.374, clip=100.000, loss_scale=1.257e+33, optim_step_time=0.075, optim0_lr0=1.548e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-05 10:16:41,963 (trainer:762) INFO: 42epoch:train:11001-11500batch: iter_time=1.583e-04, forward_time=0.210, loss_ctc=39.261, loss_att=27.221, acc=0.830, loss=30.833, backward_time=0.240, grad_norm=44.831, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.548e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-05 10:24:17,196 (trainer:762) INFO: 42epoch:train:11501-12000batch: iter_time=1.575e-04, forward_time=0.210, loss_ctc=39.664, loss_att=26.924, acc=0.829, loss=30.746, backward_time=0.237, grad_norm=43.116, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.547e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-05 10:31:51,702 (trainer:762) INFO: 42epoch:train:12001-12500batch: iter_time=1.535e-04, forward_time=0.210, loss_ctc=37.422, loss_att=25.995, acc=0.831, loss=29.423, backward_time=0.237, grad_norm=41.303, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.547e-04, train_time=0.909
+[ip-10-0-216-33:0/16] 2024-03-05 10:39:23,546 (trainer:762) INFO: 42epoch:train:12501-13000batch: iter_time=1.513e-04, forward_time=0.210, loss_ctc=40.202, loss_att=27.632, acc=0.831, loss=31.403, backward_time=0.233, grad_norm=43.550, clip=100.000, loss_scale=2.513e+33, optim_step_time=0.075, optim0_lr0=1.546e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-05 10:46:55,925 (trainer:762) INFO: 42epoch:train:13001-13500batch: iter_time=1.517e-04, forward_time=0.210, loss_ctc=38.933, loss_att=27.149, acc=0.827, loss=30.684, backward_time=0.232, grad_norm=42.532, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.545e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-05 10:54:28,555 (trainer:762) INFO: 42epoch:train:13501-14000batch: iter_time=1.543e-04, forward_time=0.209, loss_ctc=38.197, loss_att=26.220, acc=0.836, loss=29.813, backward_time=0.234, grad_norm=40.410, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.545e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-05 11:02:06,765 (trainer:762) INFO: 42epoch:train:14001-14500batch: iter_time=1.491e-04, forward_time=0.210, loss_ctc=40.495, loss_att=28.540, acc=0.822, loss=32.127, backward_time=0.233, grad_norm=45.036, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.544e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-05 11:02:52,252 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 11:09:47,295 (trainer:762) INFO: 42epoch:train:14501-15000batch: iter_time=1.485e-04, forward_time=0.208, loss_ctc=40.397, loss_att=27.442, acc=0.832, loss=31.329, backward_time=0.234, grad_norm=40.207, clip=100.000, loss_scale=2.685e+33, optim_step_time=0.075, optim0_lr0=1.544e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-05 11:23:58,258 (trainer:361) INFO: 42epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=39.647, loss_att=27.270, acc=0.830, loss=30.983, backward_time=0.234, grad_norm=43.310, clip=100.000, loss_scale=1.431e+33, optim_step_time=0.075, optim0_lr0=1.553e-04, train_time=1.004, time=4 hours, 11 minutes and 14.58 seconds, total_count=630000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=34.489, cer_ctc=0.187, loss_att=27.271, acc=0.759, cer=0.299, wer=1.000, loss=29.437, time=13 minutes and 52.47 seconds, total_count=89586, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-05 11:24:08,093 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-05 11:24:08,145 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/37epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-05 11:24:08,146 (trainer:290) INFO: 43/45epoch started. Estimated time to finish: 13 hours, 16 minutes and 5.55 seconds
+[ip-10-0-216-33:0/16] 2024-03-05 11:24:08,154 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 11:24:44,171 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 11:24:52,689 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0472310>)
+[ip-10-0-216-33:0/16] 2024-03-05 11:24:52,690 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-05 11:24:52,694 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 11:38:46,475 (trainer:762) INFO: 43epoch:train:1-500batch: iter_time=0.846, forward_time=0.211, loss_ctc=39.584, loss_att=27.068, acc=0.833, loss=30.823, backward_time=0.231, grad_norm=46.800, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.543e-04, train_time=1.756
+[ip-10-0-216-33:0/16] 2024-03-05 11:46:19,156 (trainer:762) INFO: 43epoch:train:501-1000batch: iter_time=1.525e-04, forward_time=0.210, loss_ctc=38.067, loss_att=26.907, acc=0.829, loss=30.255, backward_time=0.233, grad_norm=42.417, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.542e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-05 11:53:54,331 (trainer:762) INFO: 43epoch:train:1001-1500batch: iter_time=1.524e-04, forward_time=0.210, loss_ctc=41.357, loss_att=28.209, acc=0.829, loss=32.153, backward_time=0.232, grad_norm=42.680, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.542e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-05 12:01:29,580 (trainer:762) INFO: 43epoch:train:1501-2000batch: iter_time=1.546e-04, forward_time=0.210, loss_ctc=40.501, loss_att=27.717, acc=0.829, loss=31.552, backward_time=0.231, grad_norm=41.452, clip=100.000, loss_scale=4.933e+33, optim_step_time=0.075, optim0_lr0=1.541e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-05 12:06:15,404 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 12:08:59,138 (trainer:762) INFO: 43epoch:train:2001-2500batch: iter_time=1.495e-04, forward_time=0.210, loss_ctc=42.236, loss_att=28.294, acc=0.835, loss=32.477, backward_time=0.232, grad_norm=41.466, clip=100.000, loss_scale=4.240e+33, optim_step_time=0.075, optim0_lr0=1.540e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-05 12:16:35,567 (trainer:762) INFO: 43epoch:train:2501-3000batch: iter_time=1.508e-04, forward_time=0.211, loss_ctc=38.513, loss_att=26.707, acc=0.833, loss=30.249, backward_time=0.237, grad_norm=42.579, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.540e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-05 12:17:27,345 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 12:24:12,580 (trainer:762) INFO: 43epoch:train:3001-3500batch: iter_time=1.509e-04, forward_time=0.209, loss_ctc=39.644, loss_att=27.157, acc=0.828, loss=30.903, backward_time=0.239, grad_norm=46.906, clip=100.000, loss_scale=1.446e+33, optim_step_time=0.075, optim0_lr0=1.539e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-05 12:31:50,327 (trainer:762) INFO: 43epoch:train:3501-4000batch: iter_time=1.513e-04, forward_time=0.210, loss_ctc=39.265, loss_att=26.776, acc=0.837, loss=30.523, backward_time=0.235, grad_norm=41.954, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.539e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-05 12:39:25,956 (trainer:762) INFO: 43epoch:train:4001-4500batch: iter_time=1.520e-04, forward_time=0.211, loss_ctc=40.875, loss_att=27.971, acc=0.830, loss=31.842, backward_time=0.243, grad_norm=42.716, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.538e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 12:47:00,313 (trainer:762) INFO: 43epoch:train:4501-5000batch: iter_time=1.571e-04, forward_time=0.209, loss_ctc=36.927, loss_att=25.308, acc=0.832, loss=28.794, backward_time=0.237, grad_norm=40.362, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.537e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-05 12:47:05,678 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 12:47:42,015 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 12:47:49,925 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed00364c0>)
+[ip-10-0-216-33:0/16] 2024-03-05 12:47:49,925 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-05 12:47:49,930 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 13:02:10,663 (trainer:762) INFO: 43epoch:train:5001-5500batch: iter_time=0.899, forward_time=0.209, loss_ctc=39.475, loss_att=26.923, acc=0.834, loss=30.689, backward_time=0.237, grad_norm=42.135, clip=100.000, loss_scale=2.446e+33, optim_step_time=0.075, optim0_lr0=1.537e-04, train_time=1.820
+[ip-10-0-216-33:0/16] 2024-03-05 13:09:51,269 (trainer:762) INFO: 43epoch:train:5501-6000batch: iter_time=1.550e-04, forward_time=0.209, loss_ctc=37.761, loss_att=26.706, acc=0.830, loss=30.023, backward_time=0.238, grad_norm=43.096, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.536e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-05 13:17:22,289 (trainer:762) INFO: 43epoch:train:6001-6500batch: iter_time=1.542e-04, forward_time=0.209, loss_ctc=40.987, loss_att=28.072, acc=0.829, loss=31.946, backward_time=0.235, grad_norm=42.182, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.536e-04, train_time=0.902
+[ip-10-0-216-33:0/16] 2024-03-05 13:21:59,178 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 13:24:58,740 (trainer:762) INFO: 43epoch:train:6501-7000batch: iter_time=1.486e-04, forward_time=0.210, loss_ctc=40.135, loss_att=27.559, acc=0.830, loss=31.332, backward_time=0.235, grad_norm=42.030, clip=100.000, loss_scale=2.081e+33, optim_step_time=0.075, optim0_lr0=1.535e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-05 13:32:31,389 (trainer:762) INFO: 43epoch:train:7001-7500batch: iter_time=1.450e-04, forward_time=0.210, loss_ctc=41.786, loss_att=28.157, acc=0.836, loss=32.246, backward_time=0.233, grad_norm=41.018, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.534e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-05 13:40:09,792 (trainer:762) INFO: 43epoch:train:7501-8000batch: iter_time=1.472e-04, forward_time=0.210, loss_ctc=38.206, loss_att=26.544, acc=0.834, loss=30.043, backward_time=0.236, grad_norm=41.471, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.534e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-05 13:47:41,778 (trainer:762) INFO: 43epoch:train:8001-8500batch: iter_time=1.478e-04, forward_time=0.210, loss_ctc=39.388, loss_att=27.027, acc=0.829, loss=30.735, backward_time=0.232, grad_norm=42.274, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.533e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-05 13:55:10,013 (trainer:762) INFO: 43epoch:train:8501-9000batch: iter_time=1.506e-04, forward_time=0.209, loss_ctc=39.162, loss_att=26.779, acc=0.836, loss=30.494, backward_time=0.233, grad_norm=41.669, clip=100.000, loss_scale=1.812e+33, optim_step_time=0.075, optim0_lr0=1.533e-04, train_time=0.896
+[ip-10-0-216-33:0/16] 2024-03-05 14:02:42,990 (trainer:762) INFO: 43epoch:train:9001-9500batch: iter_time=1.481e-04, forward_time=0.211, loss_ctc=40.591, loss_att=27.836, acc=0.831, loss=31.663, backward_time=0.233, grad_norm=42.843, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.532e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-05 14:10:12,877 (trainer:762) INFO: 43epoch:train:9501-10000batch: iter_time=1.475e-04, forward_time=0.208, loss_ctc=36.821, loss_att=25.355, acc=0.832, loss=28.795, backward_time=0.235, grad_norm=40.189, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.531e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-05 14:10:18,396 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 14:10:55,410 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 14:11:03,386 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e8fe0d040>)
+[ip-10-0-216-33:0/16] 2024-03-05 14:11:03,386 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-05 14:11:03,391 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 14:27:51,833 (trainer:762) INFO: 43epoch:train:10001-10500batch: iter_time=1.203, forward_time=0.209, loss_ctc=39.288, loss_att=26.965, acc=0.835, loss=30.662, backward_time=0.232, grad_norm=41.793, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.531e-04, train_time=2.118
+[ip-10-0-216-33:0/16] 2024-03-05 14:28:33,500 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 14:35:29,761 (trainer:762) INFO: 43epoch:train:10501-11000batch: iter_time=1.492e-04, forward_time=0.209, loss_ctc=37.840, loss_att=26.610, acc=0.830, loss=29.979, backward_time=0.233, grad_norm=42.125, clip=100.000, loss_scale=1.410e+33, optim_step_time=0.075, optim0_lr0=1.530e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-05 14:43:01,495 (trainer:762) INFO: 43epoch:train:11001-11500batch: iter_time=1.504e-04, forward_time=0.209, loss_ctc=40.908, loss_att=27.985, acc=0.830, loss=31.862, backward_time=0.232, grad_norm=41.907, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.530e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-05 14:50:31,750 (trainer:762) INFO: 43epoch:train:11501-12000batch: iter_time=1.456e-04, forward_time=0.211, loss_ctc=40.181, loss_att=27.539, acc=0.830, loss=31.332, backward_time=0.227, grad_norm=41.639, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.529e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-05 14:58:04,565 (trainer:762) INFO: 43epoch:train:12001-12500batch: iter_time=1.476e-04, forward_time=0.210, loss_ctc=41.850, loss_att=28.150, acc=0.835, loss=32.260, backward_time=0.235, grad_norm=41.259, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.528e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-05 15:05:33,864 (trainer:762) INFO: 43epoch:train:12501-13000batch: iter_time=1.453e-04, forward_time=0.210, loss_ctc=38.142, loss_att=26.441, acc=0.834, loss=29.952, backward_time=0.228, grad_norm=42.711, clip=100.000, loss_scale=2.482e+33, optim_step_time=0.075, optim0_lr0=1.528e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-03-05 15:13:06,986 (trainer:762) INFO: 43epoch:train:13001-13500batch: iter_time=1.459e-04, forward_time=0.209, loss_ctc=39.434, loss_att=27.093, acc=0.829, loss=30.796, backward_time=0.227, grad_norm=42.236, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.527e-04, train_time=0.906
+[ip-10-0-216-33:0/16] 2024-03-05 15:20:39,825 (trainer:762) INFO: 43epoch:train:13501-14000batch: iter_time=1.466e-04, forward_time=0.208, loss_ctc=38.996, loss_att=26.709, acc=0.837, loss=30.395, backward_time=0.234, grad_norm=56.582, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.527e-04, train_time=0.905
+[ip-10-0-216-33:0/16] 2024-03-05 15:24:29,054 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 15:28:16,558 (trainer:762) INFO: 43epoch:train:14001-14500batch: iter_time=1.470e-04, forward_time=0.210, loss_ctc=40.432, loss_att=27.701, acc=0.831, loss=31.521, backward_time=0.233, grad_norm=43.043, clip=100.000, loss_scale=1.946e+33, optim_step_time=0.075, optim0_lr0=1.526e-04, train_time=0.913
+[ip-10-0-216-33:0/16] 2024-03-05 15:35:50,338 (trainer:762) INFO: 43epoch:train:14501-15000batch: iter_time=1.466e-04, forward_time=0.208, loss_ctc=36.849, loss_att=25.325, acc=0.832, loss=28.782, backward_time=0.229, grad_norm=40.684, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.525e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-05 15:50:17,575 (trainer:361) INFO: 43epoch results: [train] iter_time=0.098, forward_time=0.210, loss_ctc=39.507, loss_att=27.120, acc=0.832, loss=30.836, backward_time=0.234, grad_norm=42.740, clip=100.000, loss_scale=2.144e+33, optim_step_time=0.075, optim0_lr0=1.534e-04, train_time=1.007, time=4 hours, 11 minutes and 57.51 seconds, total_count=645000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=34.780, cer_ctc=0.188, loss_att=29.530, acc=0.772, cer=0.224, wer=0.990, loss=31.105, time=14 minutes and 11.61 seconds, total_count=91719, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-05 15:50:27,230 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-05 15:50:27,285 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/30epoch.pth, exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/38epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-05 15:50:27,286 (trainer:290) INFO: 44/45epoch started. Estimated time to finish: 8 hours, 50 minutes and 46.43 seconds
+[ip-10-0-216-33:0/16] 2024-03-05 15:50:27,293 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 15:51:03,322 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 15:51:11,206 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf381eb0>)
+[ip-10-0-216-33:0/16] 2024-03-05 15:51:11,207 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-05 15:51:11,211 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 16:05:19,209 (trainer:762) INFO: 44epoch:train:1-500batch: iter_time=0.847, forward_time=0.210, loss_ctc=41.452, loss_att=28.085, acc=0.829, loss=32.095, backward_time=0.231, grad_norm=43.882, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.525e-04, train_time=1.784
+[ip-10-0-216-33:0/16] 2024-03-05 16:13:01,742 (trainer:762) INFO: 44epoch:train:501-1000batch: iter_time=1.494e-04, forward_time=0.208, loss_ctc=36.366, loss_att=24.883, acc=0.835, loss=28.328, backward_time=0.233, grad_norm=39.112, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.524e-04, train_time=0.925
+[ip-10-0-216-33:0/16] 2024-03-05 16:20:43,711 (trainer:762) INFO: 44epoch:train:1001-1500batch: iter_time=1.522e-04, forward_time=0.208, loss_ctc=39.008, loss_att=26.809, acc=0.830, loss=30.469, backward_time=0.231, grad_norm=42.061, clip=100.000, loss_scale=1.947e+33, optim_step_time=0.075, optim0_lr0=1.524e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-05 16:28:19,447 (trainer:762) INFO: 44epoch:train:1501-2000batch: iter_time=1.474e-04, forward_time=0.210, loss_ctc=39.833, loss_att=27.209, acc=0.831, loss=30.996, backward_time=0.232, grad_norm=42.299, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.523e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 16:35:55,816 (trainer:762) INFO: 44epoch:train:2001-2500batch: iter_time=1.491e-04, forward_time=0.210, loss_ctc=40.341, loss_att=27.374, acc=0.832, loss=31.264, backward_time=0.235, grad_norm=44.211, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.523e-04, train_time=0.912
+[ip-10-0-216-33:0/16] 2024-03-05 16:43:34,531 (trainer:762) INFO: 44epoch:train:2501-3000batch: iter_time=1.503e-04, forward_time=0.208, loss_ctc=38.669, loss_att=26.310, acc=0.834, loss=30.018, backward_time=0.231, grad_norm=39.586, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.522e-04, train_time=0.917
+[ip-10-0-216-33:0/16] 2024-03-05 16:51:06,685 (trainer:762) INFO: 44epoch:train:3001-3500batch: iter_time=1.492e-04, forward_time=0.210, loss_ctc=38.861, loss_att=26.625, acc=0.830, loss=30.296, backward_time=0.237, grad_norm=41.660, clip=100.000, loss_scale=3.894e+33, optim_step_time=0.075, optim0_lr0=1.521e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-05 16:53:28,474 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 16:53:42,659 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 16:58:34,277 (trainer:762) INFO: 44epoch:train:3501-4000batch: iter_time=1.504e-04, forward_time=0.209, loss_ctc=39.708, loss_att=26.920, acc=0.835, loss=30.757, backward_time=0.235, grad_norm=40.963, clip=100.000, loss_scale=2.565e+33, optim_step_time=0.075, optim0_lr0=1.521e-04, train_time=0.895
+[ip-10-0-216-33:0/16] 2024-03-05 17:06:08,306 (trainer:762) INFO: 44epoch:train:4001-4500batch: iter_time=1.507e-04, forward_time=0.211, loss_ctc=39.374, loss_att=27.018, acc=0.831, loss=30.725, backward_time=0.235, grad_norm=43.309, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.520e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-05 17:13:40,114 (trainer:762) INFO: 44epoch:train:4501-5000batch: iter_time=1.475e-04, forward_time=0.210, loss_ctc=38.380, loss_att=26.098, acc=0.833, loss=29.783, backward_time=0.236, grad_norm=38.719, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.520e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-05 17:13:45,136 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 17:14:21,634 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 17:14:29,485 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ec5821220>)
+[ip-10-0-216-33:0/16] 2024-03-05 17:14:29,485 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-05 17:14:29,490 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 17:28:49,381 (trainer:762) INFO: 44epoch:train:5001-5500batch: iter_time=0.875, forward_time=0.211, loss_ctc=40.935, loss_att=27.804, acc=0.830, loss=31.743, backward_time=0.234, grad_norm=42.390, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.519e-04, train_time=1.818
+[ip-10-0-216-33:0/16] 2024-03-05 17:36:19,562 (trainer:762) INFO: 44epoch:train:5501-6000batch: iter_time=1.480e-04, forward_time=0.209, loss_ctc=36.236, loss_att=24.758, acc=0.836, loss=28.202, backward_time=0.234, grad_norm=40.024, clip=100.000, loss_scale=2.144e+33, optim_step_time=0.075, optim0_lr0=1.518e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-05 17:43:53,600 (trainer:762) INFO: 44epoch:train:6001-6500batch: iter_time=1.497e-04, forward_time=0.209, loss_ctc=38.627, loss_att=26.570, acc=0.831, loss=30.187, backward_time=0.235, grad_norm=42.030, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.518e-04, train_time=0.908
+[ip-10-0-216-33:0/16] 2024-03-05 17:51:32,685 (trainer:762) INFO: 44epoch:train:6501-7000batch: iter_time=1.466e-04, forward_time=0.209, loss_ctc=39.553, loss_att=26.981, acc=0.832, loss=30.752, backward_time=0.236, grad_norm=41.731, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.517e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-05 17:59:02,344 (trainer:762) INFO: 44epoch:train:7001-7500batch: iter_time=1.483e-04, forward_time=0.210, loss_ctc=40.091, loss_att=27.251, acc=0.833, loss=31.103, backward_time=0.231, grad_norm=42.831, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.517e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-05 18:02:34,077 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 18:05:04,424 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 18:06:41,551 (trainer:762) INFO: 44epoch:train:7501-8000batch: iter_time=1.452e-04, forward_time=0.209, loss_ctc=38.452, loss_att=26.209, acc=0.835, loss=29.882, backward_time=0.232, grad_norm=39.297, clip=100.000, loss_scale=2.607e+33, optim_step_time=0.075, optim0_lr0=1.516e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-05 18:14:25,421 (trainer:762) INFO: 44epoch:train:8001-8500batch: iter_time=1.442e-04, forward_time=0.208, loss_ctc=38.637, loss_att=26.549, acc=0.831, loss=30.175, backward_time=0.235, grad_norm=41.773, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.516e-04, train_time=0.927
+[ip-10-0-216-33:0/16] 2024-03-05 18:22:02,396 (trainer:762) INFO: 44epoch:train:8501-9000batch: iter_time=1.456e-04, forward_time=0.208, loss_ctc=39.508, loss_att=26.835, acc=0.836, loss=30.637, backward_time=0.231, grad_norm=42.155, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.515e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-05 18:29:43,808 (trainer:762) INFO: 44epoch:train:9001-9500batch: iter_time=1.444e-04, forward_time=0.207, loss_ctc=39.298, loss_att=27.033, acc=0.831, loss=30.712, backward_time=0.240, grad_norm=43.535, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.514e-04, train_time=0.923
+[ip-10-0-216-33:0/16] 2024-03-05 18:37:14,174 (trainer:762) INFO: 44epoch:train:9501-10000batch: iter_time=1.442e-04, forward_time=0.209, loss_ctc=38.315, loss_att=25.998, acc=0.833, loss=29.693, backward_time=0.238, grad_norm=39.109, clip=100.000, loss_scale=1.573e+33, optim_step_time=0.075, optim0_lr0=1.514e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-05 18:37:20,369 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 18:37:58,028 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 18:38:06,311 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e8b359190>)
+[ip-10-0-216-33:0/16] 2024-03-05 18:38:06,311 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-05 18:38:06,316 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 18:48:31,876 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 18:52:43,348 (trainer:762) INFO: 44epoch:train:10001-10500batch: iter_time=0.894, forward_time=0.210, loss_ctc=40.914, loss_att=27.796, acc=0.830, loss=31.731, backward_time=0.235, grad_norm=42.359, clip=100.000, loss_scale=1.868e+33, optim_step_time=0.075, optim0_lr0=1.513e-04, train_time=1.858
+[ip-10-0-216-33:0/16] 2024-03-05 19:00:18,648 (trainer:762) INFO: 44epoch:train:10501-11000batch: iter_time=1.460e-04, forward_time=0.208, loss_ctc=36.026, loss_att=24.595, acc=0.836, loss=28.024, backward_time=0.230, grad_norm=38.746, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.513e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-05 19:07:46,987 (trainer:762) INFO: 44epoch:train:11001-11500batch: iter_time=1.477e-04, forward_time=0.210, loss_ctc=38.606, loss_att=26.638, acc=0.831, loss=30.228, backward_time=0.230, grad_norm=41.787, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.512e-04, train_time=0.896
+[ip-10-0-216-33:0/16] 2024-03-05 19:15:18,653 (trainer:762) INFO: 44epoch:train:11501-12000batch: iter_time=1.505e-04, forward_time=0.209, loss_ctc=39.598, loss_att=27.013, acc=0.832, loss=30.789, backward_time=0.235, grad_norm=42.123, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.511e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-05 19:22:52,290 (trainer:762) INFO: 44epoch:train:12001-12500batch: iter_time=1.490e-04, forward_time=0.210, loss_ctc=39.935, loss_att=27.194, acc=0.833, loss=31.016, backward_time=0.234, grad_norm=42.378, clip=100.000, loss_scale=2.025e+33, optim_step_time=0.075, optim0_lr0=1.511e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-05 19:30:32,694 (trainer:762) INFO: 44epoch:train:12501-13000batch: iter_time=1.517e-04, forward_time=0.208, loss_ctc=38.253, loss_att=26.063, acc=0.835, loss=29.720, backward_time=0.237, grad_norm=39.557, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.510e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-05 19:38:14,721 (trainer:762) INFO: 44epoch:train:13001-13500batch: iter_time=1.475e-04, forward_time=0.208, loss_ctc=38.511, loss_att=26.421, acc=0.831, loss=30.048, backward_time=0.232, grad_norm=40.778, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.510e-04, train_time=0.924
+[ip-10-0-216-33:0/16] 2024-03-05 19:41:50,680 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 19:45:52,484 (trainer:762) INFO: 44epoch:train:13501-14000batch: iter_time=1.463e-04, forward_time=0.208, loss_ctc=39.468, loss_att=26.793, acc=0.836, loss=30.596, backward_time=0.230, grad_norm=40.670, clip=100.000, loss_scale=1.899e+33, optim_step_time=0.075, optim0_lr0=1.509e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-05 19:53:33,152 (trainer:762) INFO: 44epoch:train:14001-14500batch: iter_time=1.488e-04, forward_time=0.209, loss_ctc=39.127, loss_att=26.953, acc=0.832, loss=30.605, backward_time=0.235, grad_norm=43.558, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.509e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-05 20:01:08,924 (trainer:762) INFO: 44epoch:train:14501-15000batch: iter_time=1.480e-04, forward_time=0.208, loss_ctc=38.349, loss_att=26.012, acc=0.834, loss=29.713, backward_time=0.233, grad_norm=38.992, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.508e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 20:19:12,172 (trainer:361) INFO: 44epoch results: [train] iter_time=0.087, forward_time=0.209, loss_ctc=39.014, loss_att=26.627, acc=0.833, loss=30.343, backward_time=0.234, grad_norm=41.388, clip=100.000, loss_scale=1.939e+33, optim_step_time=0.075, optim0_lr0=1.516e-04, train_time=1.003, time=4 hours, 10 minutes and 56.69 seconds, total_count=660000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=35.323, cer_ctc=0.190, loss_att=26.708, acc=0.786, cer=0.208, wer=0.980, loss=29.292, time=17 minutes and 47.87 seconds, total_count=93852, gpu_max_cached_mem_GB=36.516
+[ip-10-0-216-33:0/16] 2024-03-05 20:19:22,245 (trainer:416) INFO: The best model has been updated: valid.acc, valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-05 20:19:22,297 (trainer:290) INFO: 45/45epoch started. Estimated time to finish: 4 hours, 25 minutes and 28.14 seconds
+[ip-10-0-216-33:0/16] 2024-03-05 20:19:22,306 (multiple_iter_factory:32) INFO: Building 0th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 20:19:59,079 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 20:20:07,192 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.0", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.0", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.0", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.0", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ecf122a00>)
+[ip-10-0-216-33:0/16] 2024-03-05 20:20:07,192 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.0, 
+[ip-10-0-216-33:0/16] 2024-03-05 20:20:07,197 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 20:34:03,600 (trainer:762) INFO: 45epoch:train:1-500batch: iter_time=0.850, forward_time=0.209, loss_ctc=41.322, loss_att=27.957, acc=0.830, loss=31.967, backward_time=0.235, grad_norm=75.942, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.507e-04, train_time=1.762
+[ip-10-0-216-33:0/16] 2024-03-05 20:41:39,025 (trainer:762) INFO: 45epoch:train:501-1000batch: iter_time=1.465e-04, forward_time=0.210, loss_ctc=41.245, loss_att=28.289, acc=0.828, loss=32.176, backward_time=0.234, grad_norm=53.569, clip=100.000, loss_scale=1.994e+33, optim_step_time=0.075, optim0_lr0=1.507e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 20:49:14,285 (trainer:762) INFO: 45epoch:train:1001-1500batch: iter_time=1.426e-04, forward_time=0.210, loss_ctc=38.194, loss_att=26.018, acc=0.839, loss=29.671, backward_time=0.234, grad_norm=42.852, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.506e-04, train_time=0.910
+[ip-10-0-216-33:0/16] 2024-03-05 20:56:52,131 (trainer:762) INFO: 45epoch:train:1501-2000batch: iter_time=1.434e-04, forward_time=0.207, loss_ctc=41.896, loss_att=28.453, acc=0.829, loss=32.486, backward_time=0.232, grad_norm=44.665, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.506e-04, train_time=0.915
+[ip-10-0-216-33:0/16] 2024-03-05 21:04:22,293 (trainer:762) INFO: 45epoch:train:2001-2500batch: iter_time=1.443e-04, forward_time=0.210, loss_ctc=39.119, loss_att=27.232, acc=0.826, loss=30.798, backward_time=0.228, grad_norm=44.750, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.505e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-05 21:12:02,165 (trainer:762) INFO: 45epoch:train:2501-3000batch: iter_time=1.416e-04, forward_time=0.208, loss_ctc=40.594, loss_att=27.275, acc=0.826, loss=31.271, backward_time=0.230, grad_norm=41.235, clip=100.000, loss_scale=3.988e+33, optim_step_time=0.075, optim0_lr0=1.505e-04, train_time=0.919
+[ip-10-0-216-33:0/16] 2024-03-05 21:14:15,078 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 21:16:33,931 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 21:19:37,594 (trainer:762) INFO: 45epoch:train:3001-3500batch: iter_time=1.471e-04, forward_time=0.207, loss_ctc=38.424, loss_att=26.225, acc=0.828, loss=29.885, backward_time=0.235, grad_norm=43.784, clip=100.000, loss_scale=2.839e+33, optim_step_time=0.075, optim0_lr0=1.504e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 21:27:15,835 (trainer:762) INFO: 45epoch:train:3501-4000batch: iter_time=1.450e-04, forward_time=0.209, loss_ctc=39.141, loss_att=27.363, acc=0.833, loss=30.897, backward_time=0.234, grad_norm=42.768, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.503e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-05 21:34:47,713 (trainer:762) INFO: 45epoch:train:4001-4500batch: iter_time=1.488e-04, forward_time=0.208, loss_ctc=41.121, loss_att=27.817, acc=0.827, loss=31.808, backward_time=0.232, grad_norm=43.922, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.075, optim0_lr0=1.503e-04, train_time=0.903
+[ip-10-0-216-33:0/16] 2024-03-05 21:42:21,269 (trainer:762) INFO: 45epoch:train:4501-5000batch: iter_time=1.464e-04, forward_time=0.209, loss_ctc=37.946, loss_att=26.352, acc=0.830, loss=29.830, backward_time=0.239, grad_norm=44.325, clip=100.000, loss_scale=1.298e+33, optim_step_time=0.076, optim0_lr0=1.502e-04, train_time=0.907
+[ip-10-0-216-33:0/16] 2024-03-05 21:42:26,171 (multiple_iter_factory:32) INFO: Building 1th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 21:43:02,984 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 21:43:10,939 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.1", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.1", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.1", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.1", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5ed0036e50>)
+[ip-10-0-216-33:0/16] 2024-03-05 21:43:10,939 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.1, 
+[ip-10-0-216-33:0/16] 2024-03-05 21:43:10,944 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 21:57:31,246 (trainer:762) INFO: 45epoch:train:5001-5500batch: iter_time=0.899, forward_time=0.209, loss_ctc=40.755, loss_att=27.503, acc=0.832, loss=31.478, backward_time=0.234, grad_norm=43.600, clip=100.000, loss_scale=1.815e+33, optim_step_time=0.075, optim0_lr0=1.502e-04, train_time=1.820
+[ip-10-0-216-33:0/16] 2024-03-05 22:05:06,810 (trainer:762) INFO: 45epoch:train:5501-6000batch: iter_time=1.496e-04, forward_time=0.208, loss_ctc=40.202, loss_att=27.893, acc=0.829, loss=31.586, backward_time=0.230, grad_norm=44.578, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.501e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 22:12:35,714 (trainer:762) INFO: 45epoch:train:6001-6500batch: iter_time=1.491e-04, forward_time=0.209, loss_ctc=38.005, loss_att=26.022, acc=0.840, loss=29.616, backward_time=0.230, grad_norm=42.936, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.501e-04, train_time=0.898
+[ip-10-0-216-33:0/16] 2024-03-05 22:20:14,940 (trainer:762) INFO: 45epoch:train:6501-7000batch: iter_time=1.481e-04, forward_time=0.209, loss_ctc=41.658, loss_att=28.298, acc=0.830, loss=32.306, backward_time=0.233, grad_norm=43.470, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.500e-04, train_time=0.918
+[ip-10-0-216-33:0/16] 2024-03-05 22:25:35,800 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 22:27:50,353 (trainer:762) INFO: 45epoch:train:7001-7500batch: iter_time=1.474e-04, forward_time=0.208, loss_ctc=38.722, loss_att=27.094, acc=0.826, loss=30.582, backward_time=0.236, grad_norm=45.978, clip=100.000, loss_scale=2.872e+33, optim_step_time=0.075, optim0_lr0=1.500e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 22:35:31,016 (trainer:762) INFO: 45epoch:train:7501-8000batch: iter_time=1.473e-04, forward_time=0.208, loss_ctc=40.264, loss_att=27.052, acc=0.827, loss=31.015, backward_time=0.233, grad_norm=41.721, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.499e-04, train_time=0.921
+[ip-10-0-216-33:0/16] 2024-03-05 22:43:06,416 (trainer:762) INFO: 45epoch:train:8001-8500batch: iter_time=1.434e-04, forward_time=0.206, loss_ctc=38.191, loss_att=26.153, acc=0.827, loss=29.764, backward_time=0.232, grad_norm=42.954, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.498e-04, train_time=0.911
+[ip-10-0-216-33:0/16] 2024-03-05 22:50:43,546 (trainer:762) INFO: 45epoch:train:8501-9000batch: iter_time=1.490e-04, forward_time=0.209, loss_ctc=38.758, loss_att=27.212, acc=0.834, loss=30.676, backward_time=0.231, grad_norm=44.139, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.498e-04, train_time=0.914
+[ip-10-0-216-33:0/16] 2024-03-05 22:58:24,744 (trainer:762) INFO: 45epoch:train:9001-9500batch: iter_time=1.512e-04, forward_time=0.208, loss_ctc=41.038, loss_att=27.776, acc=0.828, loss=31.755, backward_time=0.231, grad_norm=42.827, clip=100.000, loss_scale=3.349e+33, optim_step_time=0.075, optim0_lr0=1.497e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-05 23:05:54,351 (trainer:762) INFO: 45epoch:train:9501-10000batch: iter_time=1.500e-04, forward_time=0.208, loss_ctc=37.612, loss_att=26.195, acc=0.830, loss=29.620, backward_time=0.235, grad_norm=43.846, clip=100.000, loss_scale=5.192e+33, optim_step_time=0.075, optim0_lr0=1.497e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-05 23:05:59,294 (multiple_iter_factory:32) INFO: Building 2th iter-factory...
+[ip-10-0-216-33:0/16] 2024-03-05 23:06:37,604 (s2t:445) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
+[ip-10-0-216-33:0/16] 2024-03-05 23:06:45,721 (abs_task:1663) INFO: [train] dataset:
+ESPnetDataset(
+  speech: {"path": "exp/s2t_stats_raw_bpe50000/splits3/wav.scp/split.2", "type": "kaldi_ark"}
+  text_prev: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.prev/split.2", "type": "text"}
+  text_ctc: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text.ctc/split.2", "type": "text"}
+  text: {"path": "exp/s2t_stats_raw_bpe50000/splits3/text/split.2", "type": "text"}
+  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x7f5e8f600e20>)
+[ip-10-0-216-33:0/16] 2024-03-05 23:06:45,721 (abs_task:1664) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28118, batch_size=256, key_file=exp/s2t_stats_raw_bpe50000/splits3/speech_shape/split.2, 
+[ip-10-0-216-33:0/16] 2024-03-05 23:06:45,726 (abs_task:1665) INFO: [train] mini-batch sizes summary: N-batch=28118, mean=256.0, min=256, max=257
+[ip-10-0-216-33:0/16] 2024-03-05 23:17:45,881 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 23:21:13,627 (trainer:762) INFO: 45epoch:train:10001-10500batch: iter_time=0.902, forward_time=0.211, loss_ctc=40.862, loss_att=27.538, acc=0.832, loss=31.535, backward_time=0.241, grad_norm=43.395, clip=100.000, loss_scale=4.006e+33, optim_step_time=0.076, optim0_lr0=1.496e-04, train_time=1.838
+[ip-10-0-216-33:0/16] 2024-03-05 23:28:43,922 (trainer:762) INFO: 45epoch:train:10501-11000batch: iter_time=1.515e-04, forward_time=0.211, loss_ctc=40.233, loss_att=28.020, acc=0.830, loss=31.684, backward_time=0.235, grad_norm=44.782, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.496e-04, train_time=0.900
+[ip-10-0-216-33:0/16] 2024-03-05 23:36:13,482 (trainer:762) INFO: 45epoch:train:11001-11500batch: iter_time=1.586e-04, forward_time=0.210, loss_ctc=38.147, loss_att=25.999, acc=0.840, loss=29.643, backward_time=0.232, grad_norm=42.014, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.495e-04, train_time=0.899
+[ip-10-0-216-33:0/16] 2024-03-05 23:43:54,525 (trainer:762) INFO: 45epoch:train:11501-12000batch: iter_time=1.540e-04, forward_time=0.209, loss_ctc=41.423, loss_att=28.148, acc=0.830, loss=32.130, backward_time=0.230, grad_norm=49.793, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.495e-04, train_time=0.922
+[ip-10-0-216-33:0/16] 2024-03-05 23:51:32,864 (trainer:762) INFO: 45epoch:train:12001-12500batch: iter_time=1.507e-04, forward_time=0.209, loss_ctc=38.620, loss_att=26.947, acc=0.827, loss=30.449, backward_time=0.234, grad_norm=45.146, clip=100.000, loss_scale=3.780e+33, optim_step_time=0.075, optim0_lr0=1.494e-04, train_time=0.916
+[ip-10-0-216-33:0/16] 2024-03-05 23:57:06,050 (trainer:693) WARNING: The grad norm is nan. Skipping updating the model.
+[ip-10-0-216-33:0/16] 2024-03-05 23:59:01,448 (trainer:762) INFO: 45epoch:train:12501-13000batch: iter_time=1.567e-04, forward_time=0.210, loss_ctc=40.405, loss_att=27.146, acc=0.827, loss=31.124, backward_time=0.231, grad_norm=41.074, clip=100.000, loss_scale=4.521e+33, optim_step_time=0.075, optim0_lr0=1.493e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-03-06 00:06:28,409 (trainer:762) INFO: 45epoch:train:13001-13500batch: iter_time=1.567e-04, forward_time=0.210, loss_ctc=38.179, loss_att=26.144, acc=0.828, loss=29.755, backward_time=0.234, grad_norm=43.335, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.493e-04, train_time=0.894
+[ip-10-0-216-33:0/16] 2024-03-06 00:13:57,200 (trainer:762) INFO: 45epoch:train:13501-14000batch: iter_time=1.587e-04, forward_time=0.210, loss_ctc=38.398, loss_att=27.079, acc=0.834, loss=30.475, backward_time=0.232, grad_norm=41.724, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.492e-04, train_time=0.897
+[ip-10-0-216-33:0/16] 2024-03-06 00:21:29,316 (trainer:762) INFO: 45epoch:train:14001-14500batch: iter_time=1.515e-04, forward_time=0.210, loss_ctc=40.669, loss_att=27.592, acc=0.829, loss=31.515, backward_time=0.235, grad_norm=42.560, clip=100.000, loss_scale=2.596e+33, optim_step_time=0.075, optim0_lr0=1.492e-04, train_time=0.904
+[ip-10-0-216-33:0/16] 2024-03-06 00:28:57,349 (trainer:762) INFO: 45epoch:train:14501-15000batch: iter_time=1.641e-04, forward_time=0.210, loss_ctc=37.689, loss_att=26.302, acc=0.830, loss=29.718, backward_time=0.233, grad_norm=43.027, clip=100.000, loss_scale=3.266e+33, optim_step_time=0.075, optim0_lr0=1.491e-04, train_time=0.896
+[ip-10-0-216-33:0/16] 2024-03-06 00:43:10,684 (trainer:361) INFO: 45epoch results: [train] iter_time=0.088, forward_time=0.209, loss_ctc=39.628, loss_att=27.170, acc=0.830, loss=30.907, backward_time=0.233, grad_norm=45.024, clip=100.000, loss_scale=2.725e+33, optim_step_time=0.075, optim0_lr0=1.499e-04, train_time=0.998, time=4 hours, 9 minutes and 58.03 seconds, total_count=675000, gpu_max_cached_mem_GB=36.516, [valid] loss_ctc=33.741, cer_ctc=0.183, loss_att=25.673, acc=0.767, cer=0.299, wer=0.999, loss=28.093, time=13 minutes and 50.04 seconds, total_count=95985, gpu_max_cached_mem_GB=36.516
+ip-10-0-216-33:311885:312456 [5] NCCL INFO [Service thread] Connection closed by localRank 5
+ip-10-0-216-33:311882:312458 [2] NCCL INFO [Service thread] Connection closed by localRank 2
+ip-10-0-216-33:311886:312461 [6] NCCL INFO [Service thread] Connection closed by localRank 6
+ip-10-0-216-33:311885:311885 [5] NCCL INFO comm 0x45e56440 rank 5 nranks 16 cudaDev 5 busId 901d0 - Abort COMPLETE
+ip-10-0-216-33:311887:312460 [7] NCCL INFO [Service thread] Connection closed by localRank 7
+ip-10-0-216-33:311883:312462 [3] NCCL INFO [Service thread] Connection closed by localRank 3
+ip-10-0-216-33:311884:312459 [4] NCCL INFO [Service thread] Connection closed by localRank 4
+ip-10-0-216-33:311881:312457 [1] NCCL INFO [Service thread] Connection closed by localRank 1
+ip-10-0-216-33:311886:311886 [6] NCCL INFO comm 0x42c0fbc0 rank 6 nranks 16 cudaDev 6 busId a01c0 - Abort COMPLETE
+ip-10-0-216-33:311884:311884 [4] NCCL INFO comm 0x9d5b200 rank 4 nranks 16 cudaDev 4 busId 901c0 - Abort COMPLETE
+ip-10-0-216-33:311881:311881 [1] NCCL INFO comm 0x9a92260 rank 1 nranks 16 cudaDev 1 busId 101d0 - Abort COMPLETE
+ip-10-0-216-33:311883:311883 [3] NCCL INFO comm 0x6f267bc0 rank 3 nranks 16 cudaDev 3 busId 201d0 - Abort COMPLETE
+ip-10-0-216-33:311882:311882 [2] NCCL INFO comm 0x511a1b90 rank 2 nranks 16 cudaDev 2 busId 201c0 - Abort COMPLETE
+ip-10-0-216-33:311887:311887 [7] NCCL INFO comm 0x6ee1f170 rank 7 nranks 16 cudaDev 7 busId a01d0 - Abort COMPLETE
+ip-10-0-224-187:3788822:3789277 [4] NCCL INFO [Service thread] Connection closed by localRank 4
+ip-10-0-224-187:3788824:3789274 [6] NCCL INFO [Service thread] Connection closed by localRank 6
+ip-10-0-224-187:3788820:3789275 [2] NCCL INFO [Service thread] Connection closed by localRank 2
+ip-10-0-224-187:3788823:3789280 [5] NCCL INFO [Service thread] Connection closed by localRank 5
+ip-10-0-224-187:3788826:3789279 [7] NCCL INFO [Service thread] Connection closed by localRank 7
+ip-10-0-224-187:3788821:3789276 [3] NCCL INFO [Service thread] Connection closed by localRank 3
+ip-10-0-224-187:3788819:3789281 [1] NCCL INFO [Service thread] Connection closed by localRank 1
+ip-10-0-224-187:3788818:3789278 [0] NCCL INFO [Service thread] Connection closed by localRank 0
+ip-10-0-224-187:3788822:3788822 [4] NCCL INFO comm 0x51831900 rank 12 nranks 16 cudaDev 4 busId 901c0 - Abort COMPLETE
+ip-10-0-224-187:3788824:3788824 [6] NCCL INFO comm 0x704da1d0 rank 14 nranks 16 cudaDev 6 busId a01c0 - Abort COMPLETE
+ip-10-0-224-187:3788820:3788820 [2] NCCL INFO comm 0x965e2c0 rank 10 nranks 16 cudaDev 2 busId 201c0 - Abort COMPLETE
+ip-10-0-224-187:3788823:3788823 [5] NCCL INFO comm 0xa4dcd80 rank 13 nranks 16 cudaDev 5 busId 901d0 - Abort COMPLETE
+ip-10-0-224-187:3788826:3788826 [7] NCCL INFO comm 0x42154f70 rank 15 nranks 16 cudaDev 7 busId a01d0 - Abort COMPLETE
+ip-10-0-224-187:3788821:3788821 [3] NCCL INFO comm 0x729fe240 rank 11 nranks 16 cudaDev 3 busId 201d0 - Abort COMPLETE
+ip-10-0-224-187:3788819:3788819 [1] NCCL INFO comm 0x1236d920 rank 9 nranks 16 cudaDev 1 busId 101d0 - Abort COMPLETE
+ip-10-0-224-187:3788818:3788818 [0] NCCL INFO comm 0x804dad0 rank 8 nranks 16 cudaDev 0 busId 101c0 - Abort COMPLETE
+[ip-10-0-216-33:0/16] 2024-03-06 00:43:20,608 (trainer:416) INFO: The best model has been updated: valid.total_count
+[ip-10-0-216-33:0/16] 2024-03-06 00:43:20,662 (trainer:470) INFO: The model files were removed: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/40epoch.pth
+[ip-10-0-216-33:0/16] 2024-03-06 00:43:20,662 (trainer:488) INFO: The training was finished at 45 epochs 
+[ip-10-0-216-33:0/16] 2024-03-06 00:43:20,688 (average_nbest_models:69) INFO: Averaging 5best models: criterion="valid.acc": exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/valid.acc.ave_5best.pth
+[ip-10-0-216-33:0/16] 2024-03-06 00:44:08,799 (average_nbest_models:69) INFO: Averaging 5best models: criterion="valid.total_count": exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/valid.total_count.ave_5best.pth
+wandb: - 0.000 MB of 0.000 MB uploaded
wandb: WARNING Ensure read and write access to run files dir: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/wandb/run-20240226_220402-zw9g4dhv/files, control this via the WANDB_DIR env var. See https://docs.wandb.ai/guides/track/environment-variables
+wandb: \ 0.000 MB of 0.000 MB uploaded
wandb: | 0.000 MB of 1.212 MB uploaded
wandb: / 0.580 MB of 1.212 MB uploaded
wandb: - 0.583 MB of 1.212 MB uploaded
wandb: \ 0.583 MB of 1.212 MB uploaded
wandb: | 0.645 MB of 1.212 MB uploaded
wandb: / 0.645 MB of 1.212 MB uploaded
wandb: - 0.645 MB of 1.212 MB uploaded
wandb: \ 0.645 MB of 1.212 MB uploaded
wandb: | 0.645 MB of 1.212 MB uploaded
wandb: / 0.645 MB of 1.212 MB uploaded
wandb: - 0.645 MB of 1.212 MB uploaded
wandb: \ 0.645 MB of 1.212 MB uploaded
wandb: | 0.645 MB of 1.212 MB uploaded
wandb: / 1.212 MB of 1.212 MB uploaded
wandb:                                                                                
+wandb: 
+wandb: Run history:
+wandb:                                   epoch ▁▁▁▁▂▂▂▂▂▂▃▃▃▃▃▄▄▄▄▄▅▅▅▅▅▅▆▆▆▆▆▇▇▇▇▇▇███
+wandb:                               iteration ▁▁▁▂▂▂▂▂▂▃▃▃▃▃▄▄▄▄▄▄▅▅▅▅▅▅▆▆▆▆▆▇▇▇▇▇▇███
+wandb:                             metrics/acc ▁▂▆▆▇▇▇▇▇███████████████████████████████
+wandb:                   metrics/backward_time ▄▆▆████▇▇▇▇▇█▇█▄▇▇██▇▇█▇▇▇█▇▇▇▁▁▅▇▇██▆▆▇
+wandb:                            metrics/clip ▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:                    metrics/forward_time ▁▃▃▃▃▃▂▃▃▃▃▄▃▃▄▂▄▃▃▄▄▃▄▃▄▃▂▃▄▄██▆▃▄▄▃▅▃▃
+wandb:                       metrics/grad_norm █▆▃▂▁▁▁▁▁▁▁▁▁▁▁▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▃▃▃▃▃▃▃▃
+wandb:                       metrics/iter_time ▂▁▁▂▂▁▁▁█▄▂▁▁▁▁█▄▂▁▁▂▂▁▁▁▂▁▂▂▂▁▁▂▂▁▁▁█▄▂
+wandb:                            metrics/loss █▆▃▂▂▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:                        metrics/loss_att █▆▃▂▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:                        metrics/loss_ctc █▄▃▂▂▂▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:                      metrics/loss_scale ▁▁▁▁▁▁▁▁▁▁▁▇▅▅▆▅▄▅▅█▅▅▅▇▇▅▅▄▃▄▄▄▃▃▃▄▄▄▃▄
+wandb:                      metrics/optim0_lr0 ▁▂▆█▇▇▆▆▅▅▅▅▅▄▄▄▄▄▄▄▄▄▄▄▃▃▃▃▃▃▃▃▃▃▃▃▃▃▃▃
+wandb:                 metrics/optim_step_time ▂███▇▆▆▇▇▇▇▆▆▆▇▆▆▆▆▆▆▆▆▆▆▆▆▆▆▆▃▁▄▆▆▆▆▆▆▆
+wandb:                   train/train_acc_epoch ▁▃▆▆▇▇▇▇▇▇██████████████████████████████
+wandb:         train/train_backward_time_epoch ▄▆▇████▇█▇▇▇█▇██▇█▇▇▇▇▇▇▇▇▇▇▇▇▁▁▆▇▇██▇▇▇
+wandb:                  train/train_clip_epoch ▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:          train/train_forward_time_epoch ▁▃▃▃▃▃▂▂▂▃▃▃▂▃▃▃▂▄▃▃▃▃▃▃▃▂▂▂▃▃██▅▃▃▃▂▃▄▃
+wandb: train/train_gpu_max_cached_mem_GB_epoch ▁▁▁▁▁▁▁▁▁▁██████████████████████████████
+wandb:             train/train_grad_norm_epoch █▆▃▂▁▁▁▁▁▁▁▁▁▁▂▂▂▂▂▂▂▂▂▂▃▂▃▃▃▃▃▃▃▃▃▃▃▃▃▄
+wandb:             train/train_iter_time_epoch ▂▁▁▅▂▂▂▁▂▁▁▂▅▇▂▂▂▂▂▂▆▂▂▂▂▂▂▂▂▂█▂▂▂▂▂▂▂▆▂
+wandb:              train/train_loss_att_epoch █▅▂▂▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:              train/train_loss_ctc_epoch █▄▃▃▂▂▂▂▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:                  train/train_loss_epoch █▅▃▂▂▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:            train/train_loss_scale_epoch ▁▁▁▁▁▁▁▁▁▁▃▆▆▆▇▆▅▆▆▇▇▆▅▇█▅▅▅▄▅▅▄▃▃▃▅▅▃▄▅
+wandb:            train/train_optim0_lr0_epoch ▁▃▇█▇▇▆▆▆▅▅▅▅▄▄▄▄▄▄▄▄▄▄▄▃▃▃▃▃▃▃▃▃▃▃▃▃▃▃▃
+wandb:       train/train_optim_step_time_epoch ▂███▇▇▇▇▇▇▇▇▇▇▇▇▆▆▆▇▇▇▆▇▇▇▇▇▇▇▁▁▅▇▇▇▇▇▇▇
+wandb:                        train/train_time ▁▁▁▂▂▁▁▁█▄▂▁▁▂▁█▄▂▁▁▂▂▂▁▁▂▁▂▂▂▁▁▂▂▁▁▁█▄▂
+wandb:            train/train_train_time_epoch ▁▁▂▅▄▃▄▄▄▄▃▃▆▇▃▃▃▄▄▃█▄▄▃▃▄▄▄▄▄▇▃▄▅▄▄▅▄▅▃
+wandb:                   valid/valid_acc_epoch ▁▃▅▆▆▆▆▇▇▇▇▇▇▇▇██▇▇█▇█▇█████████████████
+wandb:               valid/valid_cer_ctc_epoch █▇▅▄▃▃▃▂▂▂▂▂▂▂▂▂▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:                   valid/valid_cer_epoch █▆▅▄▄▂▃▃▃▂▃▃▃▂▂▁▁▂▂▁▂▁▂▂▁▁▁▁▁▁▁▂▂▂▁▁▁▂▁▂
+wandb: valid/valid_gpu_max_cached_mem_GB_epoch ▁▁▁▁▁▁▁▁▁▁██████████████████████████████
+wandb:              valid/valid_loss_att_epoch █▅▃▂▂▂▂▂▂▂▂▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:              valid/valid_loss_ctc_epoch █▆▅▄▃▃▃▂▂▂▂▂▂▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:                  valid/valid_loss_epoch █▅▄▃▂▂▂▂▂▂▂▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
+wandb:                   valid/valid_wer_epoch █████████████▇▇▅▅█▇▅█▆█▇▄▄▄▄▄▃▄▇▆▆▁▄▄█▄█
+wandb: 
+wandb: Run summary:
+wandb:                                   epoch 45
+wandb:                               iteration 675000
+wandb:                             metrics/acc 0.83021
+wandb:                   metrics/backward_time 0.23303
+wandb:                            metrics/clip 100.0
+wandb:                    metrics/forward_time 0.209
+wandb:                       metrics/grad_norm 45.02424
+wandb:                       metrics/iter_time 0.08849
+wandb:                            metrics/loss 30.90718
+wandb:                        metrics/loss_att 27.16983
+wandb:                        metrics/loss_ctc 39.62768
+wandb:                      metrics/loss_scale 2.724960328658875e+33
+wandb:                      metrics/optim0_lr0 0.00015
+wandb:                 metrics/optim_step_time 0.07528
+wandb:                   train/train_acc_epoch 0.83021
+wandb:         train/train_backward_time_epoch 0.23303
+wandb:                  train/train_clip_epoch 100.0
+wandb:          train/train_forward_time_epoch 0.209
+wandb: train/train_gpu_max_cached_mem_GB_epoch 36.51562
+wandb:             train/train_grad_norm_epoch 45.02424
+wandb:             train/train_iter_time_epoch 0.08849
+wandb:              train/train_loss_att_epoch 27.16983
+wandb:              train/train_loss_ctc_epoch 39.62768
+wandb:                  train/train_loss_epoch 30.90718
+wandb:            train/train_loss_scale_epoch 2.724960328658875e+33
+wandb:            train/train_optim0_lr0_epoch 0.00015
+wandb:       train/train_optim_step_time_epoch 0.07528
+wandb:                        train/train_time 0.99808
+wandb:            train/train_train_time_epoch 0.99808
+wandb:                   valid/valid_acc_epoch 0.76696
+wandb:               valid/valid_cer_ctc_epoch 0.18327
+wandb:                   valid/valid_cer_epoch 0.29893
+wandb: valid/valid_gpu_max_cached_mem_GB_epoch 36.51562
+wandb:              valid/valid_loss_att_epoch 25.67251
+wandb:              valid/valid_loss_ctc_epoch 33.74075
+wandb:                  valid/valid_loss_epoch 28.09298
+wandb:                   valid/valid_wer_epoch 0.99948
+wandb: 
+wandb: 🚀 View run _weka_home-pengyf_espnet-owsm-train_egs2_owsm_v3.1_licensefree_nosa_s2t1 at: https://stability.wandb.io/pengyf/ESPnet_S2TTask/runs/zw9g4dhv
+wandb: Synced 4 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)
+wandb: Find logs at: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/wandb/run-20240226_220402-zw9g4dhv/logs
+ip-10-0-216-33:311880:312455 [0] NCCL INFO [Service thread] Connection closed by localRank 0
+ip-10-0-216-33:311880:311880 [0] NCCL INFO comm 0x90d2b80 rank 0 nranks 16 cudaDev 0 busId 101c0 - Abort COMPLETE
+# Accounting: begin_time=1708985012
+# Accounting: end_time=1709685882
+# Accounting: time=700870 threads=1
+# Finished at Wed Mar 6 00:44:42 UTC 2024 with status 0
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/valid.acc.ave_5best.pth b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/valid.acc.ave_5best.pth
new file mode 100644
index 0000000000000000000000000000000000000000..69b6c20bc70130ca3e7ec7c3a9aa55479575a6fe
--- /dev/null
+++ b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/valid.acc.ave_5best.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03a1dbd2d9498777c9de4f97ab2faae563f6d13f89f4cc85481c916038eb1646
+size 1466924749
diff --git a/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/valid.total_count.ave_5best.pth b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/valid.total_count.ave_5best.pth
new file mode 100644
index 0000000000000000000000000000000000000000..111b33536da6abe8c52773f79061e45e8189d9f9
--- /dev/null
+++ b/exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/valid.total_count.ave_5best.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2412ee08e7f7c2dfd2a19f26289d93a92ae1f3beb7bc319e98cd74a6b05347e1
+size 1466929645
diff --git a/meta.yaml b/meta.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..bb588f6fae3290e9589dc83cbc5ff28e2b74f0d7
--- /dev/null
+++ b/meta.yaml
@@ -0,0 +1,8 @@
+espnet: '202310'
+files:
+  s2t_model_file: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/valid.acc.ave_5best.pth
+python: 3.10.13 | packaged by conda-forge | (main, Dec 23 2023, 15:36:39) [GCC 12.3.0]
+timestamp: 1725247130.37343
+torch: 1.13.1
+yaml_files:
+  s2t_train_config: exp/s2t_train_s2t_ebf_conv2d_size768_e9_d9_piecewise_lr5e-4_warmup60k_flashattn_raw_bpe50000/config.yaml