Upload fine-tuned SpeechT5 model for Darija
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- added_tokens.json +4 -0
- checkpoint-100/added_tokens.json +4 -0
- checkpoint-100/config.json +91 -0
- checkpoint-100/generation_config.json +9 -0
- checkpoint-100/model.safetensors +3 -0
- checkpoint-100/optimizer.pt +3 -0
- checkpoint-100/preprocessor_config.json +19 -0
- checkpoint-100/rng_state.pth +3 -0
- checkpoint-100/scaler.pt +3 -0
- checkpoint-100/scheduler.pt +3 -0
- checkpoint-100/special_tokens_map.json +13 -0
- checkpoint-100/spm_char.model +3 -0
- checkpoint-100/tokenizer_config.json +64 -0
- checkpoint-100/trainer_state.json +70 -0
- checkpoint-100/training_args.bin +3 -0
- checkpoint-1000/added_tokens.json +4 -0
- checkpoint-1000/config.json +91 -0
- checkpoint-1000/generation_config.json +9 -0
- checkpoint-1000/model.safetensors +3 -0
- checkpoint-1000/optimizer.pt +3 -0
- checkpoint-1000/preprocessor_config.json +19 -0
- checkpoint-1000/rng_state.pth +3 -0
- checkpoint-1000/scaler.pt +3 -0
- checkpoint-1000/scheduler.pt +3 -0
- checkpoint-1000/special_tokens_map.json +13 -0
- checkpoint-1000/spm_char.model +3 -0
- checkpoint-1000/tokenizer_config.json +64 -0
- checkpoint-1000/trainer_state.json +394 -0
- checkpoint-1000/training_args.bin +3 -0
- checkpoint-1100/added_tokens.json +4 -0
- checkpoint-1100/config.json +91 -0
- checkpoint-1100/generation_config.json +9 -0
- checkpoint-1100/model.safetensors +3 -0
- checkpoint-1100/optimizer.pt +3 -0
- checkpoint-1100/preprocessor_config.json +19 -0
- checkpoint-1100/rng_state.pth +3 -0
- checkpoint-1100/scaler.pt +3 -0
- checkpoint-1100/scheduler.pt +3 -0
- checkpoint-1100/special_tokens_map.json +13 -0
- checkpoint-1100/spm_char.model +3 -0
- checkpoint-1100/tokenizer_config.json +64 -0
- checkpoint-1100/trainer_state.json +430 -0
- checkpoint-1100/training_args.bin +3 -0
- checkpoint-1200/added_tokens.json +4 -0
- checkpoint-1200/config.json +91 -0
- checkpoint-1200/generation_config.json +9 -0
- checkpoint-1200/model.safetensors +3 -0
- checkpoint-1200/optimizer.pt +3 -0
- checkpoint-1200/preprocessor_config.json +19 -0
- checkpoint-1200/rng_state.pth +3 -0
added_tokens.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<ctc_blank>": 80,
|
| 3 |
+
"<mask>": 79
|
| 4 |
+
}
|
checkpoint-100/added_tokens.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<ctc_blank>": 80,
|
| 3 |
+
"<mask>": 79
|
| 4 |
+
}
|
checkpoint-100/config.json
ADDED
|
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.1,
|
| 3 |
+
"apply_spec_augment": true,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"SpeechT5ForTextToSpeech"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"conv_bias": false,
|
| 10 |
+
"conv_dim": [
|
| 11 |
+
512,
|
| 12 |
+
512,
|
| 13 |
+
512,
|
| 14 |
+
512,
|
| 15 |
+
512,
|
| 16 |
+
512,
|
| 17 |
+
512
|
| 18 |
+
],
|
| 19 |
+
"conv_kernel": [
|
| 20 |
+
10,
|
| 21 |
+
3,
|
| 22 |
+
3,
|
| 23 |
+
3,
|
| 24 |
+
3,
|
| 25 |
+
2,
|
| 26 |
+
2
|
| 27 |
+
],
|
| 28 |
+
"conv_stride": [
|
| 29 |
+
5,
|
| 30 |
+
2,
|
| 31 |
+
2,
|
| 32 |
+
2,
|
| 33 |
+
2,
|
| 34 |
+
2,
|
| 35 |
+
2
|
| 36 |
+
],
|
| 37 |
+
"decoder_attention_heads": 12,
|
| 38 |
+
"decoder_ffn_dim": 3072,
|
| 39 |
+
"decoder_layerdrop": 0.1,
|
| 40 |
+
"decoder_layers": 6,
|
| 41 |
+
"decoder_start_token_id": 2,
|
| 42 |
+
"encoder_attention_heads": 12,
|
| 43 |
+
"encoder_ffn_dim": 3072,
|
| 44 |
+
"encoder_layerdrop": 0.1,
|
| 45 |
+
"encoder_layers": 12,
|
| 46 |
+
"encoder_max_relative_position": 160,
|
| 47 |
+
"eos_token_id": 2,
|
| 48 |
+
"feat_extract_activation": "gelu",
|
| 49 |
+
"feat_extract_norm": "group",
|
| 50 |
+
"feat_proj_dropout": 0.0,
|
| 51 |
+
"guided_attention_loss_num_heads": 2,
|
| 52 |
+
"guided_attention_loss_scale": 10.0,
|
| 53 |
+
"guided_attention_loss_sigma": 0.4,
|
| 54 |
+
"hidden_act": "gelu",
|
| 55 |
+
"hidden_dropout": 0.1,
|
| 56 |
+
"hidden_size": 768,
|
| 57 |
+
"initializer_range": 0.02,
|
| 58 |
+
"is_encoder_decoder": true,
|
| 59 |
+
"layer_norm_eps": 1e-05,
|
| 60 |
+
"mask_feature_length": 10,
|
| 61 |
+
"mask_feature_min_masks": 0,
|
| 62 |
+
"mask_feature_prob": 0.0,
|
| 63 |
+
"mask_time_length": 10,
|
| 64 |
+
"mask_time_min_masks": 2,
|
| 65 |
+
"mask_time_prob": 0.05,
|
| 66 |
+
"max_length": null,
|
| 67 |
+
"max_speech_positions": 1876,
|
| 68 |
+
"max_text_positions": 600,
|
| 69 |
+
"model_type": "speecht5",
|
| 70 |
+
"num_conv_pos_embedding_groups": 16,
|
| 71 |
+
"num_conv_pos_embeddings": 128,
|
| 72 |
+
"num_feat_extract_layers": 7,
|
| 73 |
+
"num_mel_bins": 80,
|
| 74 |
+
"pad_token_id": 1,
|
| 75 |
+
"positional_dropout": 0.1,
|
| 76 |
+
"reduction_factor": 2,
|
| 77 |
+
"scale_embedding": false,
|
| 78 |
+
"speaker_embedding_dim": 512,
|
| 79 |
+
"speech_decoder_postnet_dropout": 0.5,
|
| 80 |
+
"speech_decoder_postnet_kernel": 5,
|
| 81 |
+
"speech_decoder_postnet_layers": 5,
|
| 82 |
+
"speech_decoder_postnet_units": 256,
|
| 83 |
+
"speech_decoder_prenet_dropout": 0.5,
|
| 84 |
+
"speech_decoder_prenet_layers": 2,
|
| 85 |
+
"speech_decoder_prenet_units": 256,
|
| 86 |
+
"torch_dtype": "float32",
|
| 87 |
+
"transformers_version": "4.51.3",
|
| 88 |
+
"use_cache": false,
|
| 89 |
+
"use_guided_attention_loss": true,
|
| 90 |
+
"vocab_size": 81
|
| 91 |
+
}
|
checkpoint-100/generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"decoder_start_token_id": 2,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"max_length": 1876,
|
| 7 |
+
"pad_token_id": 1,
|
| 8 |
+
"transformers_version": "4.51.3"
|
| 9 |
+
}
|
checkpoint-100/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fb6553e92a72586d91d093f55b6a9199b220142530d4119a59eb780f2f2279d4
|
| 3 |
+
size 577789320
|
checkpoint-100/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2776de7e595c01b01144144e11b239e14780005182b29974ab0fc5bd28f0be2
|
| 3 |
+
size 1155772233
|
checkpoint-100/preprocessor_config.json
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": false,
|
| 3 |
+
"feature_extractor_type": "SpeechT5FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"fmax": 7600,
|
| 6 |
+
"fmin": 80,
|
| 7 |
+
"frame_signal_scale": 1.0,
|
| 8 |
+
"hop_length": 16,
|
| 9 |
+
"mel_floor": 1e-10,
|
| 10 |
+
"num_mel_bins": 80,
|
| 11 |
+
"padding_side": "right",
|
| 12 |
+
"padding_value": 0.0,
|
| 13 |
+
"processor_class": "SpeechT5Processor",
|
| 14 |
+
"reduction_factor": 2,
|
| 15 |
+
"return_attention_mask": true,
|
| 16 |
+
"sampling_rate": 16000,
|
| 17 |
+
"win_function": "hann_window",
|
| 18 |
+
"win_length": 64
|
| 19 |
+
}
|
checkpoint-100/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12807153eba72090d89aa5b95ac47f066dfe60e3bf0a7d11b7526af98aaf8082
|
| 3 |
+
size 14244
|
checkpoint-100/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47a1e6580d10177520a05ee0bb0d089956e70982a368fccd47e44cd17443e5d4
|
| 3 |
+
size 988
|
checkpoint-100/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78269da4a5a8ba9b5fd7917471207cb70ebbc4970f118ce561fab7c27acd633f
|
| 3 |
+
size 1064
|
checkpoint-100/special_tokens_map.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"eos_token": "</s>",
|
| 4 |
+
"mask_token": {
|
| 5 |
+
"content": "<mask>",
|
| 6 |
+
"lstrip": true,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false
|
| 10 |
+
},
|
| 11 |
+
"pad_token": "<pad>",
|
| 12 |
+
"unk_token": "<unk>"
|
| 13 |
+
}
|
checkpoint-100/spm_char.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7fcc48f3e225f627b1641db410ceb0c8649bd2b0c982e150b03f8be3728ab560
|
| 3 |
+
size 238473
|
checkpoint-100/tokenizer_config.json
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"79": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": true,
|
| 38 |
+
"normalized": true,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
},
|
| 43 |
+
"80": {
|
| 44 |
+
"content": "<ctc_blank>",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": true,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": false
|
| 50 |
+
}
|
| 51 |
+
},
|
| 52 |
+
"bos_token": "<s>",
|
| 53 |
+
"clean_up_tokenization_spaces": false,
|
| 54 |
+
"eos_token": "</s>",
|
| 55 |
+
"extra_special_tokens": {},
|
| 56 |
+
"mask_token": "<mask>",
|
| 57 |
+
"model_max_length": 600,
|
| 58 |
+
"normalize": false,
|
| 59 |
+
"pad_token": "<pad>",
|
| 60 |
+
"processor_class": "SpeechT5Processor",
|
| 61 |
+
"sp_model_kwargs": {},
|
| 62 |
+
"tokenizer_class": "SpeechT5Tokenizer",
|
| 63 |
+
"unk_token": "<unk>"
|
| 64 |
+
}
|
checkpoint-100/trainer_state.json
ADDED
|
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 100,
|
| 3 |
+
"best_metric": 0.5272690057754517,
|
| 4 |
+
"best_model_checkpoint": "./speecht5_finetuned_Darija/checkpoint-100",
|
| 5 |
+
"epoch": 0.2790373212417161,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 100,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.06975933031042902,
|
| 14 |
+
"grad_norm": 4.014983654022217,
|
| 15 |
+
"learning_rate": 2.4e-05,
|
| 16 |
+
"loss": 1.2037,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.13951866062085805,
|
| 21 |
+
"grad_norm": 4.588967800140381,
|
| 22 |
+
"learning_rate": 4.8e-05,
|
| 23 |
+
"loss": 0.8946,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.20927799093128707,
|
| 28 |
+
"grad_norm": 4.026744842529297,
|
| 29 |
+
"learning_rate": 7.3e-05,
|
| 30 |
+
"loss": 0.6685,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.2790373212417161,
|
| 35 |
+
"grad_norm": 3.659444808959961,
|
| 36 |
+
"learning_rate": 9.8e-05,
|
| 37 |
+
"loss": 0.6168,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.2790373212417161,
|
| 42 |
+
"eval_loss": 0.5272690057754517,
|
| 43 |
+
"eval_runtime": 41.6965,
|
| 44 |
+
"eval_samples_per_second": 30.578,
|
| 45 |
+
"eval_steps_per_second": 15.301,
|
| 46 |
+
"step": 100
|
| 47 |
+
}
|
| 48 |
+
],
|
| 49 |
+
"logging_steps": 25,
|
| 50 |
+
"max_steps": 2000,
|
| 51 |
+
"num_input_tokens_seen": 0,
|
| 52 |
+
"num_train_epochs": 6,
|
| 53 |
+
"save_steps": 100,
|
| 54 |
+
"stateful_callbacks": {
|
| 55 |
+
"TrainerControl": {
|
| 56 |
+
"args": {
|
| 57 |
+
"should_epoch_stop": false,
|
| 58 |
+
"should_evaluate": false,
|
| 59 |
+
"should_log": false,
|
| 60 |
+
"should_save": true,
|
| 61 |
+
"should_training_stop": false
|
| 62 |
+
},
|
| 63 |
+
"attributes": {}
|
| 64 |
+
}
|
| 65 |
+
},
|
| 66 |
+
"total_flos": 121755491146080.0,
|
| 67 |
+
"train_batch_size": 4,
|
| 68 |
+
"trial_name": null,
|
| 69 |
+
"trial_params": null
|
| 70 |
+
}
|
checkpoint-100/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:354a8264c86a8844409c6dae12ec0813199720c50856795c887a58d69c89f6a3
|
| 3 |
+
size 5496
|
checkpoint-1000/added_tokens.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<ctc_blank>": 80,
|
| 3 |
+
"<mask>": 79
|
| 4 |
+
}
|
checkpoint-1000/config.json
ADDED
|
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.1,
|
| 3 |
+
"apply_spec_augment": true,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"SpeechT5ForTextToSpeech"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"conv_bias": false,
|
| 10 |
+
"conv_dim": [
|
| 11 |
+
512,
|
| 12 |
+
512,
|
| 13 |
+
512,
|
| 14 |
+
512,
|
| 15 |
+
512,
|
| 16 |
+
512,
|
| 17 |
+
512
|
| 18 |
+
],
|
| 19 |
+
"conv_kernel": [
|
| 20 |
+
10,
|
| 21 |
+
3,
|
| 22 |
+
3,
|
| 23 |
+
3,
|
| 24 |
+
3,
|
| 25 |
+
2,
|
| 26 |
+
2
|
| 27 |
+
],
|
| 28 |
+
"conv_stride": [
|
| 29 |
+
5,
|
| 30 |
+
2,
|
| 31 |
+
2,
|
| 32 |
+
2,
|
| 33 |
+
2,
|
| 34 |
+
2,
|
| 35 |
+
2
|
| 36 |
+
],
|
| 37 |
+
"decoder_attention_heads": 12,
|
| 38 |
+
"decoder_ffn_dim": 3072,
|
| 39 |
+
"decoder_layerdrop": 0.1,
|
| 40 |
+
"decoder_layers": 6,
|
| 41 |
+
"decoder_start_token_id": 2,
|
| 42 |
+
"encoder_attention_heads": 12,
|
| 43 |
+
"encoder_ffn_dim": 3072,
|
| 44 |
+
"encoder_layerdrop": 0.1,
|
| 45 |
+
"encoder_layers": 12,
|
| 46 |
+
"encoder_max_relative_position": 160,
|
| 47 |
+
"eos_token_id": 2,
|
| 48 |
+
"feat_extract_activation": "gelu",
|
| 49 |
+
"feat_extract_norm": "group",
|
| 50 |
+
"feat_proj_dropout": 0.0,
|
| 51 |
+
"guided_attention_loss_num_heads": 2,
|
| 52 |
+
"guided_attention_loss_scale": 10.0,
|
| 53 |
+
"guided_attention_loss_sigma": 0.4,
|
| 54 |
+
"hidden_act": "gelu",
|
| 55 |
+
"hidden_dropout": 0.1,
|
| 56 |
+
"hidden_size": 768,
|
| 57 |
+
"initializer_range": 0.02,
|
| 58 |
+
"is_encoder_decoder": true,
|
| 59 |
+
"layer_norm_eps": 1e-05,
|
| 60 |
+
"mask_feature_length": 10,
|
| 61 |
+
"mask_feature_min_masks": 0,
|
| 62 |
+
"mask_feature_prob": 0.0,
|
| 63 |
+
"mask_time_length": 10,
|
| 64 |
+
"mask_time_min_masks": 2,
|
| 65 |
+
"mask_time_prob": 0.05,
|
| 66 |
+
"max_length": null,
|
| 67 |
+
"max_speech_positions": 1876,
|
| 68 |
+
"max_text_positions": 600,
|
| 69 |
+
"model_type": "speecht5",
|
| 70 |
+
"num_conv_pos_embedding_groups": 16,
|
| 71 |
+
"num_conv_pos_embeddings": 128,
|
| 72 |
+
"num_feat_extract_layers": 7,
|
| 73 |
+
"num_mel_bins": 80,
|
| 74 |
+
"pad_token_id": 1,
|
| 75 |
+
"positional_dropout": 0.1,
|
| 76 |
+
"reduction_factor": 2,
|
| 77 |
+
"scale_embedding": false,
|
| 78 |
+
"speaker_embedding_dim": 512,
|
| 79 |
+
"speech_decoder_postnet_dropout": 0.5,
|
| 80 |
+
"speech_decoder_postnet_kernel": 5,
|
| 81 |
+
"speech_decoder_postnet_layers": 5,
|
| 82 |
+
"speech_decoder_postnet_units": 256,
|
| 83 |
+
"speech_decoder_prenet_dropout": 0.5,
|
| 84 |
+
"speech_decoder_prenet_layers": 2,
|
| 85 |
+
"speech_decoder_prenet_units": 256,
|
| 86 |
+
"torch_dtype": "float32",
|
| 87 |
+
"transformers_version": "4.51.3",
|
| 88 |
+
"use_cache": false,
|
| 89 |
+
"use_guided_attention_loss": true,
|
| 90 |
+
"vocab_size": 81
|
| 91 |
+
}
|
checkpoint-1000/generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"decoder_start_token_id": 2,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"max_length": 1876,
|
| 7 |
+
"pad_token_id": 1,
|
| 8 |
+
"transformers_version": "4.51.3"
|
| 9 |
+
}
|
checkpoint-1000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6bd1d418bda64e0e8ed30deaa7cb664e7b1390f49c8b1e53b08dd2ccc7054fed
|
| 3 |
+
size 577789320
|
checkpoint-1000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2446e1c51d5336d5f4cc55369c8d19a28c8534ad195d79bdd509d471a174892e
|
| 3 |
+
size 1155772233
|
checkpoint-1000/preprocessor_config.json
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": false,
|
| 3 |
+
"feature_extractor_type": "SpeechT5FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"fmax": 7600,
|
| 6 |
+
"fmin": 80,
|
| 7 |
+
"frame_signal_scale": 1.0,
|
| 8 |
+
"hop_length": 16,
|
| 9 |
+
"mel_floor": 1e-10,
|
| 10 |
+
"num_mel_bins": 80,
|
| 11 |
+
"padding_side": "right",
|
| 12 |
+
"padding_value": 0.0,
|
| 13 |
+
"processor_class": "SpeechT5Processor",
|
| 14 |
+
"reduction_factor": 2,
|
| 15 |
+
"return_attention_mask": true,
|
| 16 |
+
"sampling_rate": 16000,
|
| 17 |
+
"win_function": "hann_window",
|
| 18 |
+
"win_length": 64
|
| 19 |
+
}
|
checkpoint-1000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5bfb31f1c0c64295dcc8e2a2d5384b7671c7ee22b92201a77e9916531139053c
|
| 3 |
+
size 14244
|
checkpoint-1000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:af112fad71ad8006e06c966d5edaeac57be34c7e7b6a2d76dfcb59852a9edfd7
|
| 3 |
+
size 988
|
checkpoint-1000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c98fe8b515ebd10ec044c5e3f95f19e6ca7262605bf74f191121573bcdaf65fe
|
| 3 |
+
size 1064
|
checkpoint-1000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"eos_token": "</s>",
|
| 4 |
+
"mask_token": {
|
| 5 |
+
"content": "<mask>",
|
| 6 |
+
"lstrip": true,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false
|
| 10 |
+
},
|
| 11 |
+
"pad_token": "<pad>",
|
| 12 |
+
"unk_token": "<unk>"
|
| 13 |
+
}
|
checkpoint-1000/spm_char.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7fcc48f3e225f627b1641db410ceb0c8649bd2b0c982e150b03f8be3728ab560
|
| 3 |
+
size 238473
|
checkpoint-1000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"79": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": true,
|
| 38 |
+
"normalized": true,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
},
|
| 43 |
+
"80": {
|
| 44 |
+
"content": "<ctc_blank>",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": true,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": false
|
| 50 |
+
}
|
| 51 |
+
},
|
| 52 |
+
"bos_token": "<s>",
|
| 53 |
+
"clean_up_tokenization_spaces": false,
|
| 54 |
+
"eos_token": "</s>",
|
| 55 |
+
"extra_special_tokens": {},
|
| 56 |
+
"mask_token": "<mask>",
|
| 57 |
+
"model_max_length": 600,
|
| 58 |
+
"normalize": false,
|
| 59 |
+
"pad_token": "<pad>",
|
| 60 |
+
"processor_class": "SpeechT5Processor",
|
| 61 |
+
"sp_model_kwargs": {},
|
| 62 |
+
"tokenizer_class": "SpeechT5Tokenizer",
|
| 63 |
+
"unk_token": "<unk>"
|
| 64 |
+
}
|
checkpoint-1000/trainer_state.json
ADDED
|
@@ -0,0 +1,394 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1000,
|
| 3 |
+
"best_metric": 0.44482412934303284,
|
| 4 |
+
"best_model_checkpoint": "./speecht5_finetuned_Darija/checkpoint-1000",
|
| 5 |
+
"epoch": 2.7868852459016393,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 1000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.06975933031042902,
|
| 14 |
+
"grad_norm": 4.014983654022217,
|
| 15 |
+
"learning_rate": 2.4e-05,
|
| 16 |
+
"loss": 1.2037,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.13951866062085805,
|
| 21 |
+
"grad_norm": 4.588967800140381,
|
| 22 |
+
"learning_rate": 4.8e-05,
|
| 23 |
+
"loss": 0.8946,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.20927799093128707,
|
| 28 |
+
"grad_norm": 4.026744842529297,
|
| 29 |
+
"learning_rate": 7.3e-05,
|
| 30 |
+
"loss": 0.6685,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.2790373212417161,
|
| 35 |
+
"grad_norm": 3.659444808959961,
|
| 36 |
+
"learning_rate": 9.8e-05,
|
| 37 |
+
"loss": 0.6168,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.2790373212417161,
|
| 42 |
+
"eval_loss": 0.5272690057754517,
|
| 43 |
+
"eval_runtime": 41.6965,
|
| 44 |
+
"eval_samples_per_second": 30.578,
|
| 45 |
+
"eval_steps_per_second": 15.301,
|
| 46 |
+
"step": 100
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.3487966515521451,
|
| 50 |
+
"grad_norm": 6.453115940093994,
|
| 51 |
+
"learning_rate": 9.878947368421053e-05,
|
| 52 |
+
"loss": 0.5847,
|
| 53 |
+
"step": 125
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.41855598186257414,
|
| 57 |
+
"grad_norm": 3.9114222526550293,
|
| 58 |
+
"learning_rate": 9.747368421052632e-05,
|
| 59 |
+
"loss": 0.5667,
|
| 60 |
+
"step": 150
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 0.4883153121730031,
|
| 64 |
+
"grad_norm": 5.056523323059082,
|
| 65 |
+
"learning_rate": 9.615789473684212e-05,
|
| 66 |
+
"loss": 0.5662,
|
| 67 |
+
"step": 175
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 0.5580746424834322,
|
| 71 |
+
"grad_norm": 5.494399070739746,
|
| 72 |
+
"learning_rate": 9.48421052631579e-05,
|
| 73 |
+
"loss": 0.5699,
|
| 74 |
+
"step": 200
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.5580746424834322,
|
| 78 |
+
"eval_loss": 0.5231854915618896,
|
| 79 |
+
"eval_runtime": 37.0744,
|
| 80 |
+
"eval_samples_per_second": 34.39,
|
| 81 |
+
"eval_steps_per_second": 17.209,
|
| 82 |
+
"step": 200
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 0.6278339727938612,
|
| 86 |
+
"grad_norm": 2.748530864715576,
|
| 87 |
+
"learning_rate": 9.352631578947368e-05,
|
| 88 |
+
"loss": 0.5448,
|
| 89 |
+
"step": 225
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.6975933031042902,
|
| 93 |
+
"grad_norm": 3.9691007137298584,
|
| 94 |
+
"learning_rate": 9.221052631578948e-05,
|
| 95 |
+
"loss": 0.5358,
|
| 96 |
+
"step": 250
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 0.7673526334147193,
|
| 100 |
+
"grad_norm": 3.5358879566192627,
|
| 101 |
+
"learning_rate": 9.089473684210526e-05,
|
| 102 |
+
"loss": 0.5245,
|
| 103 |
+
"step": 275
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 0.8371119637251483,
|
| 107 |
+
"grad_norm": 2.21895432472229,
|
| 108 |
+
"learning_rate": 8.957894736842106e-05,
|
| 109 |
+
"loss": 0.5231,
|
| 110 |
+
"step": 300
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.8371119637251483,
|
| 114 |
+
"eval_loss": 0.47762706875801086,
|
| 115 |
+
"eval_runtime": 36.3582,
|
| 116 |
+
"eval_samples_per_second": 35.068,
|
| 117 |
+
"eval_steps_per_second": 17.548,
|
| 118 |
+
"step": 300
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 0.9068712940355773,
|
| 122 |
+
"grad_norm": 3.315195322036743,
|
| 123 |
+
"learning_rate": 8.826315789473684e-05,
|
| 124 |
+
"loss": 0.5349,
|
| 125 |
+
"step": 325
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.9766306243460062,
|
| 129 |
+
"grad_norm": 2.936798572540283,
|
| 130 |
+
"learning_rate": 8.694736842105264e-05,
|
| 131 |
+
"loss": 0.5248,
|
| 132 |
+
"step": 350
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 1.0446459713986747,
|
| 136 |
+
"grad_norm": 2.950599431991577,
|
| 137 |
+
"learning_rate": 8.563157894736843e-05,
|
| 138 |
+
"loss": 0.5101,
|
| 139 |
+
"step": 375
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"epoch": 1.1144053017091036,
|
| 143 |
+
"grad_norm": 4.972070693969727,
|
| 144 |
+
"learning_rate": 8.431578947368422e-05,
|
| 145 |
+
"loss": 0.515,
|
| 146 |
+
"step": 400
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 1.1144053017091036,
|
| 150 |
+
"eval_loss": 0.49294987320899963,
|
| 151 |
+
"eval_runtime": 35.8818,
|
| 152 |
+
"eval_samples_per_second": 35.533,
|
| 153 |
+
"eval_steps_per_second": 17.781,
|
| 154 |
+
"step": 400
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 1.1841646320195327,
|
| 158 |
+
"grad_norm": 2.6673648357391357,
|
| 159 |
+
"learning_rate": 8.3e-05,
|
| 160 |
+
"loss": 0.5115,
|
| 161 |
+
"step": 425
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.2539239623299616,
|
| 165 |
+
"grad_norm": 3.388873338699341,
|
| 166 |
+
"learning_rate": 8.16842105263158e-05,
|
| 167 |
+
"loss": 0.5064,
|
| 168 |
+
"step": 450
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.3236832926403905,
|
| 172 |
+
"grad_norm": 2.4961979389190674,
|
| 173 |
+
"learning_rate": 8.036842105263158e-05,
|
| 174 |
+
"loss": 0.5028,
|
| 175 |
+
"step": 475
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 1.3934426229508197,
|
| 179 |
+
"grad_norm": 2.7970707416534424,
|
| 180 |
+
"learning_rate": 7.905263157894737e-05,
|
| 181 |
+
"loss": 0.4975,
|
| 182 |
+
"step": 500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.3934426229508197,
|
| 186 |
+
"eval_loss": 0.4633351266384125,
|
| 187 |
+
"eval_runtime": 36.5245,
|
| 188 |
+
"eval_samples_per_second": 34.908,
|
| 189 |
+
"eval_steps_per_second": 17.468,
|
| 190 |
+
"step": 500
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 1.4632019532612488,
|
| 194 |
+
"grad_norm": 2.774756908416748,
|
| 195 |
+
"learning_rate": 7.773684210526317e-05,
|
| 196 |
+
"loss": 0.498,
|
| 197 |
+
"step": 525
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.5329612835716777,
|
| 201 |
+
"grad_norm": 5.114898204803467,
|
| 202 |
+
"learning_rate": 7.642105263157895e-05,
|
| 203 |
+
"loss": 0.4974,
|
| 204 |
+
"step": 550
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 1.6027206138821066,
|
| 208 |
+
"grad_norm": 3.1757712364196777,
|
| 209 |
+
"learning_rate": 7.510526315789475e-05,
|
| 210 |
+
"loss": 0.5068,
|
| 211 |
+
"step": 575
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 1.6724799441925358,
|
| 215 |
+
"grad_norm": 2.1318249702453613,
|
| 216 |
+
"learning_rate": 7.378947368421053e-05,
|
| 217 |
+
"loss": 0.5003,
|
| 218 |
+
"step": 600
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.6724799441925358,
|
| 222 |
+
"eval_loss": 0.45842912793159485,
|
| 223 |
+
"eval_runtime": 36.8677,
|
| 224 |
+
"eval_samples_per_second": 34.583,
|
| 225 |
+
"eval_steps_per_second": 17.305,
|
| 226 |
+
"step": 600
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 1.742239274502965,
|
| 230 |
+
"grad_norm": 3.224851131439209,
|
| 231 |
+
"learning_rate": 7.247368421052631e-05,
|
| 232 |
+
"loss": 0.4954,
|
| 233 |
+
"step": 625
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.8119986048133938,
|
| 237 |
+
"grad_norm": 2.2048707008361816,
|
| 238 |
+
"learning_rate": 7.115789473684211e-05,
|
| 239 |
+
"loss": 0.4954,
|
| 240 |
+
"step": 650
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.8817579351238227,
|
| 244 |
+
"grad_norm": 3.310940742492676,
|
| 245 |
+
"learning_rate": 6.98421052631579e-05,
|
| 246 |
+
"loss": 0.4898,
|
| 247 |
+
"step": 675
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 1.9515172654342519,
|
| 251 |
+
"grad_norm": 3.6913022994995117,
|
| 252 |
+
"learning_rate": 6.852631578947369e-05,
|
| 253 |
+
"loss": 0.4881,
|
| 254 |
+
"step": 700
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.9515172654342519,
|
| 258 |
+
"eval_loss": 0.4606294631958008,
|
| 259 |
+
"eval_runtime": 36.611,
|
| 260 |
+
"eval_samples_per_second": 34.826,
|
| 261 |
+
"eval_steps_per_second": 17.426,
|
| 262 |
+
"step": 700
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 2.01953261248692,
|
| 266 |
+
"grad_norm": 3.2030959129333496,
|
| 267 |
+
"learning_rate": 6.721052631578948e-05,
|
| 268 |
+
"loss": 0.4727,
|
| 269 |
+
"step": 725
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 2.0892919427973493,
|
| 273 |
+
"grad_norm": 3.458202362060547,
|
| 274 |
+
"learning_rate": 6.589473684210526e-05,
|
| 275 |
+
"loss": 0.4809,
|
| 276 |
+
"step": 750
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 2.159051273107778,
|
| 280 |
+
"grad_norm": 1.9468854665756226,
|
| 281 |
+
"learning_rate": 6.457894736842106e-05,
|
| 282 |
+
"loss": 0.4749,
|
| 283 |
+
"step": 775
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 2.228810603418207,
|
| 287 |
+
"grad_norm": 3.895932912826538,
|
| 288 |
+
"learning_rate": 6.331578947368422e-05,
|
| 289 |
+
"loss": 0.4803,
|
| 290 |
+
"step": 800
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 2.228810603418207,
|
| 294 |
+
"eval_loss": 0.4526459872722626,
|
| 295 |
+
"eval_runtime": 37.4188,
|
| 296 |
+
"eval_samples_per_second": 34.074,
|
| 297 |
+
"eval_steps_per_second": 17.05,
|
| 298 |
+
"step": 800
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 2.2985699337286363,
|
| 302 |
+
"grad_norm": 3.0414257049560547,
|
| 303 |
+
"learning_rate": 6.2e-05,
|
| 304 |
+
"loss": 0.4774,
|
| 305 |
+
"step": 825
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 2.3683292640390654,
|
| 309 |
+
"grad_norm": 2.6127874851226807,
|
| 310 |
+
"learning_rate": 6.0684210526315785e-05,
|
| 311 |
+
"loss": 0.4745,
|
| 312 |
+
"step": 850
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 2.438088594349494,
|
| 316 |
+
"grad_norm": 3.0023088455200195,
|
| 317 |
+
"learning_rate": 5.936842105263158e-05,
|
| 318 |
+
"loss": 0.4791,
|
| 319 |
+
"step": 875
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 2.5078479246599232,
|
| 323 |
+
"grad_norm": 2.730001926422119,
|
| 324 |
+
"learning_rate": 5.805263157894737e-05,
|
| 325 |
+
"loss": 0.4785,
|
| 326 |
+
"step": 900
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 2.5078479246599232,
|
| 330 |
+
"eval_loss": 0.448452353477478,
|
| 331 |
+
"eval_runtime": 37.2061,
|
| 332 |
+
"eval_samples_per_second": 34.269,
|
| 333 |
+
"eval_steps_per_second": 17.148,
|
| 334 |
+
"step": 900
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 2.5776072549703524,
|
| 338 |
+
"grad_norm": 2.2317161560058594,
|
| 339 |
+
"learning_rate": 5.6736842105263166e-05,
|
| 340 |
+
"loss": 0.4732,
|
| 341 |
+
"step": 925
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 2.647366585280781,
|
| 345 |
+
"grad_norm": 3.1052346229553223,
|
| 346 |
+
"learning_rate": 5.542105263157895e-05,
|
| 347 |
+
"loss": 0.4801,
|
| 348 |
+
"step": 950
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 2.71712591559121,
|
| 352 |
+
"grad_norm": 2.309736967086792,
|
| 353 |
+
"learning_rate": 5.410526315789474e-05,
|
| 354 |
+
"loss": 0.4789,
|
| 355 |
+
"step": 975
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 2.7868852459016393,
|
| 359 |
+
"grad_norm": 2.5499184131622314,
|
| 360 |
+
"learning_rate": 5.2789473684210534e-05,
|
| 361 |
+
"loss": 0.4733,
|
| 362 |
+
"step": 1000
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 2.7868852459016393,
|
| 366 |
+
"eval_loss": 0.44482412934303284,
|
| 367 |
+
"eval_runtime": 36.9591,
|
| 368 |
+
"eval_samples_per_second": 34.498,
|
| 369 |
+
"eval_steps_per_second": 17.262,
|
| 370 |
+
"step": 1000
|
| 371 |
+
}
|
| 372 |
+
],
|
| 373 |
+
"logging_steps": 25,
|
| 374 |
+
"max_steps": 2000,
|
| 375 |
+
"num_input_tokens_seen": 0,
|
| 376 |
+
"num_train_epochs": 6,
|
| 377 |
+
"save_steps": 100,
|
| 378 |
+
"stateful_callbacks": {
|
| 379 |
+
"TrainerControl": {
|
| 380 |
+
"args": {
|
| 381 |
+
"should_epoch_stop": false,
|
| 382 |
+
"should_evaluate": false,
|
| 383 |
+
"should_log": false,
|
| 384 |
+
"should_save": true,
|
| 385 |
+
"should_training_stop": false
|
| 386 |
+
},
|
| 387 |
+
"attributes": {}
|
| 388 |
+
}
|
| 389 |
+
},
|
| 390 |
+
"total_flos": 1231346591934624.0,
|
| 391 |
+
"train_batch_size": 4,
|
| 392 |
+
"trial_name": null,
|
| 393 |
+
"trial_params": null
|
| 394 |
+
}
|
checkpoint-1000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:354a8264c86a8844409c6dae12ec0813199720c50856795c887a58d69c89f6a3
|
| 3 |
+
size 5496
|
checkpoint-1100/added_tokens.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<ctc_blank>": 80,
|
| 3 |
+
"<mask>": 79
|
| 4 |
+
}
|
checkpoint-1100/config.json
ADDED
|
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.1,
|
| 3 |
+
"apply_spec_augment": true,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"SpeechT5ForTextToSpeech"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"conv_bias": false,
|
| 10 |
+
"conv_dim": [
|
| 11 |
+
512,
|
| 12 |
+
512,
|
| 13 |
+
512,
|
| 14 |
+
512,
|
| 15 |
+
512,
|
| 16 |
+
512,
|
| 17 |
+
512
|
| 18 |
+
],
|
| 19 |
+
"conv_kernel": [
|
| 20 |
+
10,
|
| 21 |
+
3,
|
| 22 |
+
3,
|
| 23 |
+
3,
|
| 24 |
+
3,
|
| 25 |
+
2,
|
| 26 |
+
2
|
| 27 |
+
],
|
| 28 |
+
"conv_stride": [
|
| 29 |
+
5,
|
| 30 |
+
2,
|
| 31 |
+
2,
|
| 32 |
+
2,
|
| 33 |
+
2,
|
| 34 |
+
2,
|
| 35 |
+
2
|
| 36 |
+
],
|
| 37 |
+
"decoder_attention_heads": 12,
|
| 38 |
+
"decoder_ffn_dim": 3072,
|
| 39 |
+
"decoder_layerdrop": 0.1,
|
| 40 |
+
"decoder_layers": 6,
|
| 41 |
+
"decoder_start_token_id": 2,
|
| 42 |
+
"encoder_attention_heads": 12,
|
| 43 |
+
"encoder_ffn_dim": 3072,
|
| 44 |
+
"encoder_layerdrop": 0.1,
|
| 45 |
+
"encoder_layers": 12,
|
| 46 |
+
"encoder_max_relative_position": 160,
|
| 47 |
+
"eos_token_id": 2,
|
| 48 |
+
"feat_extract_activation": "gelu",
|
| 49 |
+
"feat_extract_norm": "group",
|
| 50 |
+
"feat_proj_dropout": 0.0,
|
| 51 |
+
"guided_attention_loss_num_heads": 2,
|
| 52 |
+
"guided_attention_loss_scale": 10.0,
|
| 53 |
+
"guided_attention_loss_sigma": 0.4,
|
| 54 |
+
"hidden_act": "gelu",
|
| 55 |
+
"hidden_dropout": 0.1,
|
| 56 |
+
"hidden_size": 768,
|
| 57 |
+
"initializer_range": 0.02,
|
| 58 |
+
"is_encoder_decoder": true,
|
| 59 |
+
"layer_norm_eps": 1e-05,
|
| 60 |
+
"mask_feature_length": 10,
|
| 61 |
+
"mask_feature_min_masks": 0,
|
| 62 |
+
"mask_feature_prob": 0.0,
|
| 63 |
+
"mask_time_length": 10,
|
| 64 |
+
"mask_time_min_masks": 2,
|
| 65 |
+
"mask_time_prob": 0.05,
|
| 66 |
+
"max_length": null,
|
| 67 |
+
"max_speech_positions": 1876,
|
| 68 |
+
"max_text_positions": 600,
|
| 69 |
+
"model_type": "speecht5",
|
| 70 |
+
"num_conv_pos_embedding_groups": 16,
|
| 71 |
+
"num_conv_pos_embeddings": 128,
|
| 72 |
+
"num_feat_extract_layers": 7,
|
| 73 |
+
"num_mel_bins": 80,
|
| 74 |
+
"pad_token_id": 1,
|
| 75 |
+
"positional_dropout": 0.1,
|
| 76 |
+
"reduction_factor": 2,
|
| 77 |
+
"scale_embedding": false,
|
| 78 |
+
"speaker_embedding_dim": 512,
|
| 79 |
+
"speech_decoder_postnet_dropout": 0.5,
|
| 80 |
+
"speech_decoder_postnet_kernel": 5,
|
| 81 |
+
"speech_decoder_postnet_layers": 5,
|
| 82 |
+
"speech_decoder_postnet_units": 256,
|
| 83 |
+
"speech_decoder_prenet_dropout": 0.5,
|
| 84 |
+
"speech_decoder_prenet_layers": 2,
|
| 85 |
+
"speech_decoder_prenet_units": 256,
|
| 86 |
+
"torch_dtype": "float32",
|
| 87 |
+
"transformers_version": "4.51.3",
|
| 88 |
+
"use_cache": false,
|
| 89 |
+
"use_guided_attention_loss": true,
|
| 90 |
+
"vocab_size": 81
|
| 91 |
+
}
|
checkpoint-1100/generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"decoder_start_token_id": 2,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"max_length": 1876,
|
| 7 |
+
"pad_token_id": 1,
|
| 8 |
+
"transformers_version": "4.51.3"
|
| 9 |
+
}
|
checkpoint-1100/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1515289757ca3e4edf8ab503847c453cb879c54361bf683b91cb154d85ecff21
|
| 3 |
+
size 577789320
|
checkpoint-1100/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:68cd3ba8d42097f1fa875b22aa99c4107cf1630ed6b2d7a19e302625357b31aa
|
| 3 |
+
size 1155772233
|
checkpoint-1100/preprocessor_config.json
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": false,
|
| 3 |
+
"feature_extractor_type": "SpeechT5FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"fmax": 7600,
|
| 6 |
+
"fmin": 80,
|
| 7 |
+
"frame_signal_scale": 1.0,
|
| 8 |
+
"hop_length": 16,
|
| 9 |
+
"mel_floor": 1e-10,
|
| 10 |
+
"num_mel_bins": 80,
|
| 11 |
+
"padding_side": "right",
|
| 12 |
+
"padding_value": 0.0,
|
| 13 |
+
"processor_class": "SpeechT5Processor",
|
| 14 |
+
"reduction_factor": 2,
|
| 15 |
+
"return_attention_mask": true,
|
| 16 |
+
"sampling_rate": 16000,
|
| 17 |
+
"win_function": "hann_window",
|
| 18 |
+
"win_length": 64
|
| 19 |
+
}
|
checkpoint-1100/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f71aa4ef2aa2b2c50e5991c7f0f8138e6145332189d025ab84237ffc15c1c50
|
| 3 |
+
size 14244
|
checkpoint-1100/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eb2d44f23aea7387eb8f6a37ccd0f9b573e828163294e7c2e7e472da44c76139
|
| 3 |
+
size 988
|
checkpoint-1100/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f4c135e02d69cec69ecce91a0b5c8766bcc46f992297c01dbad30db73c80b97
|
| 3 |
+
size 1064
|
checkpoint-1100/special_tokens_map.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"eos_token": "</s>",
|
| 4 |
+
"mask_token": {
|
| 5 |
+
"content": "<mask>",
|
| 6 |
+
"lstrip": true,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false
|
| 10 |
+
},
|
| 11 |
+
"pad_token": "<pad>",
|
| 12 |
+
"unk_token": "<unk>"
|
| 13 |
+
}
|
checkpoint-1100/spm_char.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7fcc48f3e225f627b1641db410ceb0c8649bd2b0c982e150b03f8be3728ab560
|
| 3 |
+
size 238473
|
checkpoint-1100/tokenizer_config.json
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"79": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": true,
|
| 38 |
+
"normalized": true,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
},
|
| 43 |
+
"80": {
|
| 44 |
+
"content": "<ctc_blank>",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": true,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": false
|
| 50 |
+
}
|
| 51 |
+
},
|
| 52 |
+
"bos_token": "<s>",
|
| 53 |
+
"clean_up_tokenization_spaces": false,
|
| 54 |
+
"eos_token": "</s>",
|
| 55 |
+
"extra_special_tokens": {},
|
| 56 |
+
"mask_token": "<mask>",
|
| 57 |
+
"model_max_length": 600,
|
| 58 |
+
"normalize": false,
|
| 59 |
+
"pad_token": "<pad>",
|
| 60 |
+
"processor_class": "SpeechT5Processor",
|
| 61 |
+
"sp_model_kwargs": {},
|
| 62 |
+
"tokenizer_class": "SpeechT5Tokenizer",
|
| 63 |
+
"unk_token": "<unk>"
|
| 64 |
+
}
|
checkpoint-1100/trainer_state.json
ADDED
|
@@ -0,0 +1,430 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1100,
|
| 3 |
+
"best_metric": 0.4392930865287781,
|
| 4 |
+
"best_model_checkpoint": "./speecht5_finetuned_Darija/checkpoint-1100",
|
| 5 |
+
"epoch": 3.0641785838855946,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 1100,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.06975933031042902,
|
| 14 |
+
"grad_norm": 4.014983654022217,
|
| 15 |
+
"learning_rate": 2.4e-05,
|
| 16 |
+
"loss": 1.2037,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.13951866062085805,
|
| 21 |
+
"grad_norm": 4.588967800140381,
|
| 22 |
+
"learning_rate": 4.8e-05,
|
| 23 |
+
"loss": 0.8946,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.20927799093128707,
|
| 28 |
+
"grad_norm": 4.026744842529297,
|
| 29 |
+
"learning_rate": 7.3e-05,
|
| 30 |
+
"loss": 0.6685,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.2790373212417161,
|
| 35 |
+
"grad_norm": 3.659444808959961,
|
| 36 |
+
"learning_rate": 9.8e-05,
|
| 37 |
+
"loss": 0.6168,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.2790373212417161,
|
| 42 |
+
"eval_loss": 0.5272690057754517,
|
| 43 |
+
"eval_runtime": 41.6965,
|
| 44 |
+
"eval_samples_per_second": 30.578,
|
| 45 |
+
"eval_steps_per_second": 15.301,
|
| 46 |
+
"step": 100
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.3487966515521451,
|
| 50 |
+
"grad_norm": 6.453115940093994,
|
| 51 |
+
"learning_rate": 9.878947368421053e-05,
|
| 52 |
+
"loss": 0.5847,
|
| 53 |
+
"step": 125
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.41855598186257414,
|
| 57 |
+
"grad_norm": 3.9114222526550293,
|
| 58 |
+
"learning_rate": 9.747368421052632e-05,
|
| 59 |
+
"loss": 0.5667,
|
| 60 |
+
"step": 150
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 0.4883153121730031,
|
| 64 |
+
"grad_norm": 5.056523323059082,
|
| 65 |
+
"learning_rate": 9.615789473684212e-05,
|
| 66 |
+
"loss": 0.5662,
|
| 67 |
+
"step": 175
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 0.5580746424834322,
|
| 71 |
+
"grad_norm": 5.494399070739746,
|
| 72 |
+
"learning_rate": 9.48421052631579e-05,
|
| 73 |
+
"loss": 0.5699,
|
| 74 |
+
"step": 200
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.5580746424834322,
|
| 78 |
+
"eval_loss": 0.5231854915618896,
|
| 79 |
+
"eval_runtime": 37.0744,
|
| 80 |
+
"eval_samples_per_second": 34.39,
|
| 81 |
+
"eval_steps_per_second": 17.209,
|
| 82 |
+
"step": 200
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 0.6278339727938612,
|
| 86 |
+
"grad_norm": 2.748530864715576,
|
| 87 |
+
"learning_rate": 9.352631578947368e-05,
|
| 88 |
+
"loss": 0.5448,
|
| 89 |
+
"step": 225
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.6975933031042902,
|
| 93 |
+
"grad_norm": 3.9691007137298584,
|
| 94 |
+
"learning_rate": 9.221052631578948e-05,
|
| 95 |
+
"loss": 0.5358,
|
| 96 |
+
"step": 250
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 0.7673526334147193,
|
| 100 |
+
"grad_norm": 3.5358879566192627,
|
| 101 |
+
"learning_rate": 9.089473684210526e-05,
|
| 102 |
+
"loss": 0.5245,
|
| 103 |
+
"step": 275
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 0.8371119637251483,
|
| 107 |
+
"grad_norm": 2.21895432472229,
|
| 108 |
+
"learning_rate": 8.957894736842106e-05,
|
| 109 |
+
"loss": 0.5231,
|
| 110 |
+
"step": 300
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.8371119637251483,
|
| 114 |
+
"eval_loss": 0.47762706875801086,
|
| 115 |
+
"eval_runtime": 36.3582,
|
| 116 |
+
"eval_samples_per_second": 35.068,
|
| 117 |
+
"eval_steps_per_second": 17.548,
|
| 118 |
+
"step": 300
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 0.9068712940355773,
|
| 122 |
+
"grad_norm": 3.315195322036743,
|
| 123 |
+
"learning_rate": 8.826315789473684e-05,
|
| 124 |
+
"loss": 0.5349,
|
| 125 |
+
"step": 325
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.9766306243460062,
|
| 129 |
+
"grad_norm": 2.936798572540283,
|
| 130 |
+
"learning_rate": 8.694736842105264e-05,
|
| 131 |
+
"loss": 0.5248,
|
| 132 |
+
"step": 350
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 1.0446459713986747,
|
| 136 |
+
"grad_norm": 2.950599431991577,
|
| 137 |
+
"learning_rate": 8.563157894736843e-05,
|
| 138 |
+
"loss": 0.5101,
|
| 139 |
+
"step": 375
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"epoch": 1.1144053017091036,
|
| 143 |
+
"grad_norm": 4.972070693969727,
|
| 144 |
+
"learning_rate": 8.431578947368422e-05,
|
| 145 |
+
"loss": 0.515,
|
| 146 |
+
"step": 400
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 1.1144053017091036,
|
| 150 |
+
"eval_loss": 0.49294987320899963,
|
| 151 |
+
"eval_runtime": 35.8818,
|
| 152 |
+
"eval_samples_per_second": 35.533,
|
| 153 |
+
"eval_steps_per_second": 17.781,
|
| 154 |
+
"step": 400
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 1.1841646320195327,
|
| 158 |
+
"grad_norm": 2.6673648357391357,
|
| 159 |
+
"learning_rate": 8.3e-05,
|
| 160 |
+
"loss": 0.5115,
|
| 161 |
+
"step": 425
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.2539239623299616,
|
| 165 |
+
"grad_norm": 3.388873338699341,
|
| 166 |
+
"learning_rate": 8.16842105263158e-05,
|
| 167 |
+
"loss": 0.5064,
|
| 168 |
+
"step": 450
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.3236832926403905,
|
| 172 |
+
"grad_norm": 2.4961979389190674,
|
| 173 |
+
"learning_rate": 8.036842105263158e-05,
|
| 174 |
+
"loss": 0.5028,
|
| 175 |
+
"step": 475
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 1.3934426229508197,
|
| 179 |
+
"grad_norm": 2.7970707416534424,
|
| 180 |
+
"learning_rate": 7.905263157894737e-05,
|
| 181 |
+
"loss": 0.4975,
|
| 182 |
+
"step": 500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.3934426229508197,
|
| 186 |
+
"eval_loss": 0.4633351266384125,
|
| 187 |
+
"eval_runtime": 36.5245,
|
| 188 |
+
"eval_samples_per_second": 34.908,
|
| 189 |
+
"eval_steps_per_second": 17.468,
|
| 190 |
+
"step": 500
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 1.4632019532612488,
|
| 194 |
+
"grad_norm": 2.774756908416748,
|
| 195 |
+
"learning_rate": 7.773684210526317e-05,
|
| 196 |
+
"loss": 0.498,
|
| 197 |
+
"step": 525
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.5329612835716777,
|
| 201 |
+
"grad_norm": 5.114898204803467,
|
| 202 |
+
"learning_rate": 7.642105263157895e-05,
|
| 203 |
+
"loss": 0.4974,
|
| 204 |
+
"step": 550
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 1.6027206138821066,
|
| 208 |
+
"grad_norm": 3.1757712364196777,
|
| 209 |
+
"learning_rate": 7.510526315789475e-05,
|
| 210 |
+
"loss": 0.5068,
|
| 211 |
+
"step": 575
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 1.6724799441925358,
|
| 215 |
+
"grad_norm": 2.1318249702453613,
|
| 216 |
+
"learning_rate": 7.378947368421053e-05,
|
| 217 |
+
"loss": 0.5003,
|
| 218 |
+
"step": 600
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.6724799441925358,
|
| 222 |
+
"eval_loss": 0.45842912793159485,
|
| 223 |
+
"eval_runtime": 36.8677,
|
| 224 |
+
"eval_samples_per_second": 34.583,
|
| 225 |
+
"eval_steps_per_second": 17.305,
|
| 226 |
+
"step": 600
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 1.742239274502965,
|
| 230 |
+
"grad_norm": 3.224851131439209,
|
| 231 |
+
"learning_rate": 7.247368421052631e-05,
|
| 232 |
+
"loss": 0.4954,
|
| 233 |
+
"step": 625
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.8119986048133938,
|
| 237 |
+
"grad_norm": 2.2048707008361816,
|
| 238 |
+
"learning_rate": 7.115789473684211e-05,
|
| 239 |
+
"loss": 0.4954,
|
| 240 |
+
"step": 650
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.8817579351238227,
|
| 244 |
+
"grad_norm": 3.310940742492676,
|
| 245 |
+
"learning_rate": 6.98421052631579e-05,
|
| 246 |
+
"loss": 0.4898,
|
| 247 |
+
"step": 675
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 1.9515172654342519,
|
| 251 |
+
"grad_norm": 3.6913022994995117,
|
| 252 |
+
"learning_rate": 6.852631578947369e-05,
|
| 253 |
+
"loss": 0.4881,
|
| 254 |
+
"step": 700
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.9515172654342519,
|
| 258 |
+
"eval_loss": 0.4606294631958008,
|
| 259 |
+
"eval_runtime": 36.611,
|
| 260 |
+
"eval_samples_per_second": 34.826,
|
| 261 |
+
"eval_steps_per_second": 17.426,
|
| 262 |
+
"step": 700
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 2.01953261248692,
|
| 266 |
+
"grad_norm": 3.2030959129333496,
|
| 267 |
+
"learning_rate": 6.721052631578948e-05,
|
| 268 |
+
"loss": 0.4727,
|
| 269 |
+
"step": 725
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 2.0892919427973493,
|
| 273 |
+
"grad_norm": 3.458202362060547,
|
| 274 |
+
"learning_rate": 6.589473684210526e-05,
|
| 275 |
+
"loss": 0.4809,
|
| 276 |
+
"step": 750
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 2.159051273107778,
|
| 280 |
+
"grad_norm": 1.9468854665756226,
|
| 281 |
+
"learning_rate": 6.457894736842106e-05,
|
| 282 |
+
"loss": 0.4749,
|
| 283 |
+
"step": 775
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 2.228810603418207,
|
| 287 |
+
"grad_norm": 3.895932912826538,
|
| 288 |
+
"learning_rate": 6.331578947368422e-05,
|
| 289 |
+
"loss": 0.4803,
|
| 290 |
+
"step": 800
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 2.228810603418207,
|
| 294 |
+
"eval_loss": 0.4526459872722626,
|
| 295 |
+
"eval_runtime": 37.4188,
|
| 296 |
+
"eval_samples_per_second": 34.074,
|
| 297 |
+
"eval_steps_per_second": 17.05,
|
| 298 |
+
"step": 800
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 2.2985699337286363,
|
| 302 |
+
"grad_norm": 3.0414257049560547,
|
| 303 |
+
"learning_rate": 6.2e-05,
|
| 304 |
+
"loss": 0.4774,
|
| 305 |
+
"step": 825
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 2.3683292640390654,
|
| 309 |
+
"grad_norm": 2.6127874851226807,
|
| 310 |
+
"learning_rate": 6.0684210526315785e-05,
|
| 311 |
+
"loss": 0.4745,
|
| 312 |
+
"step": 850
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 2.438088594349494,
|
| 316 |
+
"grad_norm": 3.0023088455200195,
|
| 317 |
+
"learning_rate": 5.936842105263158e-05,
|
| 318 |
+
"loss": 0.4791,
|
| 319 |
+
"step": 875
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 2.5078479246599232,
|
| 323 |
+
"grad_norm": 2.730001926422119,
|
| 324 |
+
"learning_rate": 5.805263157894737e-05,
|
| 325 |
+
"loss": 0.4785,
|
| 326 |
+
"step": 900
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 2.5078479246599232,
|
| 330 |
+
"eval_loss": 0.448452353477478,
|
| 331 |
+
"eval_runtime": 37.2061,
|
| 332 |
+
"eval_samples_per_second": 34.269,
|
| 333 |
+
"eval_steps_per_second": 17.148,
|
| 334 |
+
"step": 900
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 2.5776072549703524,
|
| 338 |
+
"grad_norm": 2.2317161560058594,
|
| 339 |
+
"learning_rate": 5.6736842105263166e-05,
|
| 340 |
+
"loss": 0.4732,
|
| 341 |
+
"step": 925
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 2.647366585280781,
|
| 345 |
+
"grad_norm": 3.1052346229553223,
|
| 346 |
+
"learning_rate": 5.542105263157895e-05,
|
| 347 |
+
"loss": 0.4801,
|
| 348 |
+
"step": 950
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 2.71712591559121,
|
| 352 |
+
"grad_norm": 2.309736967086792,
|
| 353 |
+
"learning_rate": 5.410526315789474e-05,
|
| 354 |
+
"loss": 0.4789,
|
| 355 |
+
"step": 975
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 2.7868852459016393,
|
| 359 |
+
"grad_norm": 2.5499184131622314,
|
| 360 |
+
"learning_rate": 5.2789473684210534e-05,
|
| 361 |
+
"loss": 0.4733,
|
| 362 |
+
"step": 1000
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 2.7868852459016393,
|
| 366 |
+
"eval_loss": 0.44482412934303284,
|
| 367 |
+
"eval_runtime": 36.9591,
|
| 368 |
+
"eval_samples_per_second": 34.498,
|
| 369 |
+
"eval_steps_per_second": 17.262,
|
| 370 |
+
"step": 1000
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 2.8566445762120685,
|
| 374 |
+
"grad_norm": 3.4246134757995605,
|
| 375 |
+
"learning_rate": 5.1473684210526317e-05,
|
| 376 |
+
"loss": 0.4778,
|
| 377 |
+
"step": 1025
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 2.9264039065224976,
|
| 381 |
+
"grad_norm": 2.3219287395477295,
|
| 382 |
+
"learning_rate": 5.0157894736842106e-05,
|
| 383 |
+
"loss": 0.4748,
|
| 384 |
+
"step": 1050
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 2.9961632368329263,
|
| 388 |
+
"grad_norm": 4.338850021362305,
|
| 389 |
+
"learning_rate": 4.8842105263157895e-05,
|
| 390 |
+
"loss": 0.4738,
|
| 391 |
+
"step": 1075
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 3.0641785838855946,
|
| 395 |
+
"grad_norm": 2.8604490756988525,
|
| 396 |
+
"learning_rate": 4.7526315789473684e-05,
|
| 397 |
+
"loss": 0.4542,
|
| 398 |
+
"step": 1100
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 3.0641785838855946,
|
| 402 |
+
"eval_loss": 0.4392930865287781,
|
| 403 |
+
"eval_runtime": 37.1499,
|
| 404 |
+
"eval_samples_per_second": 34.32,
|
| 405 |
+
"eval_steps_per_second": 17.174,
|
| 406 |
+
"step": 1100
|
| 407 |
+
}
|
| 408 |
+
],
|
| 409 |
+
"logging_steps": 25,
|
| 410 |
+
"max_steps": 2000,
|
| 411 |
+
"num_input_tokens_seen": 0,
|
| 412 |
+
"num_train_epochs": 6,
|
| 413 |
+
"save_steps": 100,
|
| 414 |
+
"stateful_callbacks": {
|
| 415 |
+
"TrainerControl": {
|
| 416 |
+
"args": {
|
| 417 |
+
"should_epoch_stop": false,
|
| 418 |
+
"should_evaluate": false,
|
| 419 |
+
"should_log": false,
|
| 420 |
+
"should_save": true,
|
| 421 |
+
"should_training_stop": false
|
| 422 |
+
},
|
| 423 |
+
"attributes": {}
|
| 424 |
+
}
|
| 425 |
+
},
|
| 426 |
+
"total_flos": 1351207069461216.0,
|
| 427 |
+
"train_batch_size": 4,
|
| 428 |
+
"trial_name": null,
|
| 429 |
+
"trial_params": null
|
| 430 |
+
}
|
checkpoint-1100/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:354a8264c86a8844409c6dae12ec0813199720c50856795c887a58d69c89f6a3
|
| 3 |
+
size 5496
|
checkpoint-1200/added_tokens.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<ctc_blank>": 80,
|
| 3 |
+
"<mask>": 79
|
| 4 |
+
}
|
checkpoint-1200/config.json
ADDED
|
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.1,
|
| 3 |
+
"apply_spec_augment": true,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"SpeechT5ForTextToSpeech"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"conv_bias": false,
|
| 10 |
+
"conv_dim": [
|
| 11 |
+
512,
|
| 12 |
+
512,
|
| 13 |
+
512,
|
| 14 |
+
512,
|
| 15 |
+
512,
|
| 16 |
+
512,
|
| 17 |
+
512
|
| 18 |
+
],
|
| 19 |
+
"conv_kernel": [
|
| 20 |
+
10,
|
| 21 |
+
3,
|
| 22 |
+
3,
|
| 23 |
+
3,
|
| 24 |
+
3,
|
| 25 |
+
2,
|
| 26 |
+
2
|
| 27 |
+
],
|
| 28 |
+
"conv_stride": [
|
| 29 |
+
5,
|
| 30 |
+
2,
|
| 31 |
+
2,
|
| 32 |
+
2,
|
| 33 |
+
2,
|
| 34 |
+
2,
|
| 35 |
+
2
|
| 36 |
+
],
|
| 37 |
+
"decoder_attention_heads": 12,
|
| 38 |
+
"decoder_ffn_dim": 3072,
|
| 39 |
+
"decoder_layerdrop": 0.1,
|
| 40 |
+
"decoder_layers": 6,
|
| 41 |
+
"decoder_start_token_id": 2,
|
| 42 |
+
"encoder_attention_heads": 12,
|
| 43 |
+
"encoder_ffn_dim": 3072,
|
| 44 |
+
"encoder_layerdrop": 0.1,
|
| 45 |
+
"encoder_layers": 12,
|
| 46 |
+
"encoder_max_relative_position": 160,
|
| 47 |
+
"eos_token_id": 2,
|
| 48 |
+
"feat_extract_activation": "gelu",
|
| 49 |
+
"feat_extract_norm": "group",
|
| 50 |
+
"feat_proj_dropout": 0.0,
|
| 51 |
+
"guided_attention_loss_num_heads": 2,
|
| 52 |
+
"guided_attention_loss_scale": 10.0,
|
| 53 |
+
"guided_attention_loss_sigma": 0.4,
|
| 54 |
+
"hidden_act": "gelu",
|
| 55 |
+
"hidden_dropout": 0.1,
|
| 56 |
+
"hidden_size": 768,
|
| 57 |
+
"initializer_range": 0.02,
|
| 58 |
+
"is_encoder_decoder": true,
|
| 59 |
+
"layer_norm_eps": 1e-05,
|
| 60 |
+
"mask_feature_length": 10,
|
| 61 |
+
"mask_feature_min_masks": 0,
|
| 62 |
+
"mask_feature_prob": 0.0,
|
| 63 |
+
"mask_time_length": 10,
|
| 64 |
+
"mask_time_min_masks": 2,
|
| 65 |
+
"mask_time_prob": 0.05,
|
| 66 |
+
"max_length": null,
|
| 67 |
+
"max_speech_positions": 1876,
|
| 68 |
+
"max_text_positions": 600,
|
| 69 |
+
"model_type": "speecht5",
|
| 70 |
+
"num_conv_pos_embedding_groups": 16,
|
| 71 |
+
"num_conv_pos_embeddings": 128,
|
| 72 |
+
"num_feat_extract_layers": 7,
|
| 73 |
+
"num_mel_bins": 80,
|
| 74 |
+
"pad_token_id": 1,
|
| 75 |
+
"positional_dropout": 0.1,
|
| 76 |
+
"reduction_factor": 2,
|
| 77 |
+
"scale_embedding": false,
|
| 78 |
+
"speaker_embedding_dim": 512,
|
| 79 |
+
"speech_decoder_postnet_dropout": 0.5,
|
| 80 |
+
"speech_decoder_postnet_kernel": 5,
|
| 81 |
+
"speech_decoder_postnet_layers": 5,
|
| 82 |
+
"speech_decoder_postnet_units": 256,
|
| 83 |
+
"speech_decoder_prenet_dropout": 0.5,
|
| 84 |
+
"speech_decoder_prenet_layers": 2,
|
| 85 |
+
"speech_decoder_prenet_units": 256,
|
| 86 |
+
"torch_dtype": "float32",
|
| 87 |
+
"transformers_version": "4.51.3",
|
| 88 |
+
"use_cache": false,
|
| 89 |
+
"use_guided_attention_loss": true,
|
| 90 |
+
"vocab_size": 81
|
| 91 |
+
}
|
checkpoint-1200/generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"decoder_start_token_id": 2,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"max_length": 1876,
|
| 7 |
+
"pad_token_id": 1,
|
| 8 |
+
"transformers_version": "4.51.3"
|
| 9 |
+
}
|
checkpoint-1200/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7fbca962afe1a543093eefad83a2dd786ee2d242affa44a3b914cad4cfebcc50
|
| 3 |
+
size 577789320
|
checkpoint-1200/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:204df246de4beea7e03d2c3c094f12c4dd92bfee2b05e47685ae808666f19384
|
| 3 |
+
size 1155772233
|
checkpoint-1200/preprocessor_config.json
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": false,
|
| 3 |
+
"feature_extractor_type": "SpeechT5FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"fmax": 7600,
|
| 6 |
+
"fmin": 80,
|
| 7 |
+
"frame_signal_scale": 1.0,
|
| 8 |
+
"hop_length": 16,
|
| 9 |
+
"mel_floor": 1e-10,
|
| 10 |
+
"num_mel_bins": 80,
|
| 11 |
+
"padding_side": "right",
|
| 12 |
+
"padding_value": 0.0,
|
| 13 |
+
"processor_class": "SpeechT5Processor",
|
| 14 |
+
"reduction_factor": 2,
|
| 15 |
+
"return_attention_mask": true,
|
| 16 |
+
"sampling_rate": 16000,
|
| 17 |
+
"win_function": "hann_window",
|
| 18 |
+
"win_length": 64
|
| 19 |
+
}
|
checkpoint-1200/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5bb1f5229a40e00865bb1dc2808a875e380f81ae55762e1c43d7e879fdefce78
|
| 3 |
+
size 14244
|