Commit
•
4f1c7a8
1
Parent(s):
95e760c
Training in progress, step 32000
Browse files- README.md +356 -0
- added_tokens.json +7 -0
- all_results.json +14 -0
- config.json +28 -0
- eval_results.json +9 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +9 -0
- tokenizer.json +0 -0
- tokenizer_config.json +62 -0
- train_results.json +8 -0
- trainer_state.json +3328 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
ADDED
|
@@ -0,0 +1,356 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
base_model: cardiffnlp/twitter-roberta-base-2019-90m
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
model-index:
|
| 7 |
+
- name: 2020-Q1-50p-filtered
|
| 8 |
+
results: []
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 12 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 13 |
+
|
| 14 |
+
# 2020-Q1-50p-filtered
|
| 15 |
+
|
| 16 |
+
This model is a fine-tuned version of [cardiffnlp/twitter-roberta-base-2019-90m](https://huggingface.co/cardiffnlp/twitter-roberta-base-2019-90m) on an unknown dataset.
|
| 17 |
+
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 2.4514
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 4.1e-07
|
| 38 |
+
- train_batch_size: 16
|
| 39 |
+
- eval_batch_size: 16
|
| 40 |
+
- seed: 42
|
| 41 |
+
- optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-08
|
| 42 |
+
- lr_scheduler_type: linear
|
| 43 |
+
- training_steps: 2400000
|
| 44 |
+
|
| 45 |
+
### Training results
|
| 46 |
+
|
| 47 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 48 |
+
|:-------------:|:-----:|:-------:|:---------------:|
|
| 49 |
+
| No log | 0.03 | 8000 | 2.8937 |
|
| 50 |
+
| 3.073 | 0.07 | 16000 | 2.7660 |
|
| 51 |
+
| 3.073 | 0.1 | 24000 | 2.7233 |
|
| 52 |
+
| 2.8244 | 0.13 | 32000 | 2.6878 |
|
| 53 |
+
| 2.8244 | 0.16 | 40000 | 2.6520 |
|
| 54 |
+
| 2.7542 | 0.2 | 48000 | 2.6300 |
|
| 55 |
+
| 2.7542 | 0.23 | 56000 | 2.6135 |
|
| 56 |
+
| 2.7083 | 0.26 | 64000 | 2.6068 |
|
| 57 |
+
| 2.7083 | 0.3 | 72000 | 2.5854 |
|
| 58 |
+
| 2.6752 | 0.33 | 80000 | 2.5755 |
|
| 59 |
+
| 2.6752 | 0.36 | 88000 | 2.5721 |
|
| 60 |
+
| 2.6657 | 0.39 | 96000 | 2.5709 |
|
| 61 |
+
| 2.6657 | 0.43 | 104000 | 2.5656 |
|
| 62 |
+
| 2.6534 | 0.46 | 112000 | 2.5558 |
|
| 63 |
+
| 2.6534 | 0.49 | 120000 | 2.5496 |
|
| 64 |
+
| 2.646 | 0.52 | 128000 | 2.5471 |
|
| 65 |
+
| 2.646 | 0.56 | 136000 | 2.5408 |
|
| 66 |
+
| 2.625 | 0.59 | 144000 | 2.5315 |
|
| 67 |
+
| 2.625 | 0.62 | 152000 | 2.5365 |
|
| 68 |
+
| 2.6222 | 0.66 | 160000 | 2.5372 |
|
| 69 |
+
| 2.6222 | 0.69 | 168000 | 2.5342 |
|
| 70 |
+
| 2.6256 | 0.72 | 176000 | 2.5308 |
|
| 71 |
+
| 2.6256 | 0.75 | 184000 | 2.5312 |
|
| 72 |
+
| 2.6074 | 0.79 | 192000 | 2.5228 |
|
| 73 |
+
| 2.6074 | 0.82 | 200000 | 2.5292 |
|
| 74 |
+
| 2.6071 | 0.85 | 208000 | 2.5295 |
|
| 75 |
+
| 2.6071 | 0.89 | 216000 | 2.5235 |
|
| 76 |
+
| 2.5955 | 0.92 | 224000 | 2.5219 |
|
| 77 |
+
| 2.5955 | 0.95 | 232000 | 2.5191 |
|
| 78 |
+
| 2.6036 | 0.98 | 240000 | 2.5171 |
|
| 79 |
+
| 2.6036 | 1.02 | 248000 | 2.5102 |
|
| 80 |
+
| 2.6046 | 1.05 | 256000 | 2.5070 |
|
| 81 |
+
| 2.6046 | 1.08 | 264000 | 2.5109 |
|
| 82 |
+
| 2.5892 | 1.11 | 272000 | 2.5105 |
|
| 83 |
+
| 2.5892 | 1.15 | 280000 | 2.5087 |
|
| 84 |
+
| 2.5929 | 1.18 | 288000 | 2.5094 |
|
| 85 |
+
| 2.5929 | 1.21 | 296000 | 2.5086 |
|
| 86 |
+
| 2.5857 | 1.25 | 304000 | 2.4991 |
|
| 87 |
+
| 2.5857 | 1.28 | 312000 | 2.5089 |
|
| 88 |
+
| 2.5828 | 1.31 | 320000 | 2.5017 |
|
| 89 |
+
| 2.5828 | 1.34 | 328000 | 2.5039 |
|
| 90 |
+
| 2.5812 | 1.38 | 336000 | 2.5065 |
|
| 91 |
+
| 2.5812 | 1.41 | 344000 | 2.5083 |
|
| 92 |
+
| 2.5775 | 1.44 | 352000 | 2.5099 |
|
| 93 |
+
| 2.5775 | 1.48 | 360000 | 2.5079 |
|
| 94 |
+
| 2.5711 | 1.51 | 368000 | 2.4922 |
|
| 95 |
+
| 2.5711 | 1.54 | 376000 | 2.5012 |
|
| 96 |
+
| 2.5797 | 1.57 | 384000 | 2.4999 |
|
| 97 |
+
| 2.5797 | 1.61 | 392000 | 2.4881 |
|
| 98 |
+
| 2.5718 | 1.64 | 400000 | 2.4960 |
|
| 99 |
+
| 2.5718 | 1.67 | 408000 | 2.4908 |
|
| 100 |
+
| 2.5627 | 1.7 | 416000 | 2.4971 |
|
| 101 |
+
| 2.5627 | 1.74 | 424000 | 2.4916 |
|
| 102 |
+
| 2.5641 | 1.77 | 432000 | 2.4971 |
|
| 103 |
+
| 2.5641 | 1.8 | 440000 | 2.4954 |
|
| 104 |
+
| 2.5633 | 1.84 | 448000 | 2.4860 |
|
| 105 |
+
| 2.5633 | 1.87 | 456000 | 2.4894 |
|
| 106 |
+
| 2.5676 | 1.9 | 464000 | 2.4893 |
|
| 107 |
+
| 2.5676 | 1.93 | 472000 | 2.4884 |
|
| 108 |
+
| 2.5687 | 1.97 | 480000 | 2.4921 |
|
| 109 |
+
| 2.5687 | 2.0 | 488000 | 2.4873 |
|
| 110 |
+
| 2.5633 | 2.03 | 496000 | 2.4919 |
|
| 111 |
+
| 2.5633 | 2.07 | 504000 | 2.4821 |
|
| 112 |
+
| 2.5547 | 2.1 | 512000 | 2.4909 |
|
| 113 |
+
| 2.5547 | 2.13 | 520000 | 2.4818 |
|
| 114 |
+
| 2.5617 | 2.16 | 528000 | 2.4855 |
|
| 115 |
+
| 2.5617 | 2.2 | 536000 | 2.4850 |
|
| 116 |
+
| 2.5569 | 2.23 | 544000 | 2.4803 |
|
| 117 |
+
| 2.5569 | 2.26 | 552000 | 2.4776 |
|
| 118 |
+
| 2.5535 | 2.29 | 560000 | 2.4824 |
|
| 119 |
+
| 2.5535 | 2.33 | 568000 | 2.4822 |
|
| 120 |
+
| 2.5534 | 2.36 | 576000 | 2.4763 |
|
| 121 |
+
| 2.5534 | 2.39 | 584000 | 2.4797 |
|
| 122 |
+
| 2.5583 | 2.43 | 592000 | 2.4872 |
|
| 123 |
+
| 2.5583 | 2.46 | 600000 | 2.4812 |
|
| 124 |
+
| 2.5545 | 2.49 | 608000 | 2.4748 |
|
| 125 |
+
| 2.5545 | 2.52 | 616000 | 2.4736 |
|
| 126 |
+
| 2.5561 | 2.56 | 624000 | 2.4714 |
|
| 127 |
+
| 2.5561 | 2.59 | 632000 | 2.4858 |
|
| 128 |
+
| 2.5384 | 2.62 | 640000 | 2.4829 |
|
| 129 |
+
| 2.5384 | 2.66 | 648000 | 2.4766 |
|
| 130 |
+
| 2.541 | 2.69 | 656000 | 2.4836 |
|
| 131 |
+
| 2.541 | 2.72 | 664000 | 2.4651 |
|
| 132 |
+
| 2.5439 | 2.75 | 672000 | 2.4797 |
|
| 133 |
+
| 2.5439 | 2.79 | 680000 | 2.4702 |
|
| 134 |
+
| 2.5597 | 2.82 | 688000 | 2.4751 |
|
| 135 |
+
| 2.5597 | 2.85 | 696000 | 2.4744 |
|
| 136 |
+
| 2.5491 | 2.88 | 704000 | 2.4756 |
|
| 137 |
+
| 2.5491 | 2.92 | 712000 | 2.4731 |
|
| 138 |
+
| 2.5505 | 2.95 | 720000 | 2.4756 |
|
| 139 |
+
| 2.5505 | 2.98 | 728000 | 2.4704 |
|
| 140 |
+
| 2.5432 | 3.02 | 736000 | 2.4763 |
|
| 141 |
+
| 2.5432 | 3.05 | 744000 | 2.4743 |
|
| 142 |
+
| 2.5485 | 3.08 | 752000 | 2.4627 |
|
| 143 |
+
| 2.5485 | 3.11 | 760000 | 2.4714 |
|
| 144 |
+
| 2.5482 | 3.15 | 768000 | 2.4685 |
|
| 145 |
+
| 2.5482 | 3.18 | 776000 | 2.4673 |
|
| 146 |
+
| 2.5411 | 3.21 | 784000 | 2.4726 |
|
| 147 |
+
| 2.5411 | 3.25 | 792000 | 2.4761 |
|
| 148 |
+
| 2.5407 | 3.28 | 800000 | 2.4612 |
|
| 149 |
+
| 2.5407 | 3.31 | 808000 | 2.4743 |
|
| 150 |
+
| 2.5307 | 3.34 | 816000 | 2.4699 |
|
| 151 |
+
| 2.5307 | 3.38 | 824000 | 2.4721 |
|
| 152 |
+
| 2.5391 | 3.41 | 832000 | 2.4614 |
|
| 153 |
+
| 2.5391 | 3.44 | 840000 | 2.4641 |
|
| 154 |
+
| 2.5378 | 3.47 | 848000 | 2.4652 |
|
| 155 |
+
| 2.5378 | 3.51 | 856000 | 2.4641 |
|
| 156 |
+
| 2.5399 | 3.54 | 864000 | 2.4691 |
|
| 157 |
+
| 2.5399 | 3.57 | 872000 | 2.4612 |
|
| 158 |
+
| 2.5412 | 3.61 | 880000 | 2.4696 |
|
| 159 |
+
| 2.5412 | 3.64 | 888000 | 2.4638 |
|
| 160 |
+
| 2.5389 | 3.67 | 896000 | 2.4658 |
|
| 161 |
+
| 2.5389 | 3.7 | 904000 | 2.4725 |
|
| 162 |
+
| 2.5325 | 3.74 | 912000 | 2.4642 |
|
| 163 |
+
| 2.5325 | 3.77 | 920000 | 2.4599 |
|
| 164 |
+
| 2.5351 | 3.8 | 928000 | 2.4617 |
|
| 165 |
+
| 2.5351 | 3.84 | 936000 | 2.4646 |
|
| 166 |
+
| 2.522 | 3.87 | 944000 | 2.4665 |
|
| 167 |
+
| 2.522 | 3.9 | 952000 | 2.4762 |
|
| 168 |
+
| 2.5331 | 3.93 | 960000 | 2.4669 |
|
| 169 |
+
| 2.5331 | 3.97 | 968000 | 2.4550 |
|
| 170 |
+
| 2.5276 | 4.0 | 976000 | 2.4662 |
|
| 171 |
+
| 2.5276 | 4.03 | 984000 | 2.4645 |
|
| 172 |
+
| 2.5206 | 4.06 | 992000 | 2.4587 |
|
| 173 |
+
| 2.5206 | 4.1 | 1000000 | 2.4725 |
|
| 174 |
+
| 2.5294 | 4.13 | 1008000 | 2.4588 |
|
| 175 |
+
| 2.5294 | 4.16 | 1016000 | 2.4591 |
|
| 176 |
+
| 2.5312 | 4.2 | 1024000 | 2.4681 |
|
| 177 |
+
| 2.5312 | 4.23 | 1032000 | 2.4625 |
|
| 178 |
+
| 2.525 | 4.26 | 1040000 | 2.4659 |
|
| 179 |
+
| 2.525 | 4.29 | 1048000 | 2.4609 |
|
| 180 |
+
| 2.5318 | 4.33 | 1056000 | 2.4571 |
|
| 181 |
+
| 2.5318 | 4.36 | 1064000 | 2.4582 |
|
| 182 |
+
| 2.5332 | 4.39 | 1072000 | 2.4566 |
|
| 183 |
+
| 2.5332 | 4.43 | 1080000 | 2.4588 |
|
| 184 |
+
| 2.5168 | 4.46 | 1088000 | 2.4606 |
|
| 185 |
+
| 2.5168 | 4.49 | 1096000 | 2.4598 |
|
| 186 |
+
| 2.5181 | 4.52 | 1104000 | 2.4543 |
|
| 187 |
+
| 2.5181 | 4.56 | 1112000 | 2.4620 |
|
| 188 |
+
| 2.5246 | 4.59 | 1120000 | 2.4639 |
|
| 189 |
+
| 2.5246 | 4.62 | 1128000 | 2.4556 |
|
| 190 |
+
| 2.5318 | 4.65 | 1136000 | 2.4571 |
|
| 191 |
+
| 2.5318 | 4.69 | 1144000 | 2.4636 |
|
| 192 |
+
| 2.512 | 4.72 | 1152000 | 2.4568 |
|
| 193 |
+
| 2.512 | 4.75 | 1160000 | 2.4644 |
|
| 194 |
+
| 2.5174 | 4.79 | 1168000 | 2.4529 |
|
| 195 |
+
| 2.5174 | 4.82 | 1176000 | 2.4614 |
|
| 196 |
+
| 2.5196 | 4.85 | 1184000 | 2.4638 |
|
| 197 |
+
| 2.5196 | 4.88 | 1192000 | 2.4534 |
|
| 198 |
+
| 2.5248 | 4.92 | 1200000 | 2.4553 |
|
| 199 |
+
| 2.5248 | 4.95 | 1208000 | 2.4537 |
|
| 200 |
+
| 2.5201 | 4.98 | 1216000 | 2.4579 |
|
| 201 |
+
| 2.5201 | 5.02 | 1224000 | 2.4525 |
|
| 202 |
+
| 2.5164 | 5.05 | 1232000 | 2.4645 |
|
| 203 |
+
| 2.5164 | 5.08 | 1240000 | 2.4480 |
|
| 204 |
+
| 2.5186 | 5.11 | 1248000 | 2.4606 |
|
| 205 |
+
| 2.5186 | 5.15 | 1256000 | 2.4623 |
|
| 206 |
+
| 2.5123 | 5.18 | 1264000 | 2.4566 |
|
| 207 |
+
| 2.5123 | 5.21 | 1272000 | 2.4644 |
|
| 208 |
+
| 2.5233 | 5.24 | 1280000 | 2.4576 |
|
| 209 |
+
| 2.5233 | 5.28 | 1288000 | 2.4519 |
|
| 210 |
+
| 2.513 | 5.31 | 1296000 | 2.4570 |
|
| 211 |
+
| 2.513 | 5.34 | 1304000 | 2.4627 |
|
| 212 |
+
| 2.5226 | 5.38 | 1312000 | 2.4500 |
|
| 213 |
+
| 2.5226 | 5.41 | 1320000 | 2.4563 |
|
| 214 |
+
| 2.5222 | 5.44 | 1328000 | 2.4521 |
|
| 215 |
+
| 2.5222 | 5.47 | 1336000 | 2.4591 |
|
| 216 |
+
| 2.5191 | 5.51 | 1344000 | 2.4509 |
|
| 217 |
+
| 2.5191 | 5.54 | 1352000 | 2.4559 |
|
| 218 |
+
| 2.5243 | 5.57 | 1360000 | 2.4502 |
|
| 219 |
+
| 2.5243 | 5.61 | 1368000 | 2.4515 |
|
| 220 |
+
| 2.5157 | 5.64 | 1376000 | 2.4563 |
|
| 221 |
+
| 2.5157 | 5.67 | 1384000 | 2.4526 |
|
| 222 |
+
| 2.5162 | 5.7 | 1392000 | 2.4586 |
|
| 223 |
+
| 2.5162 | 5.74 | 1400000 | 2.4584 |
|
| 224 |
+
| 2.5169 | 5.77 | 1408000 | 2.4542 |
|
| 225 |
+
| 2.5169 | 5.8 | 1416000 | 2.4602 |
|
| 226 |
+
| 2.5127 | 5.84 | 1424000 | 2.4587 |
|
| 227 |
+
| 2.5127 | 5.87 | 1432000 | 2.4529 |
|
| 228 |
+
| 2.5144 | 5.9 | 1440000 | 2.4620 |
|
| 229 |
+
| 2.5144 | 5.93 | 1448000 | 2.4509 |
|
| 230 |
+
| 2.5175 | 5.97 | 1456000 | 2.4503 |
|
| 231 |
+
| 2.5175 | 6.0 | 1464000 | 2.4545 |
|
| 232 |
+
| 2.5147 | 6.03 | 1472000 | 2.4440 |
|
| 233 |
+
| 2.5147 | 6.06 | 1480000 | 2.4577 |
|
| 234 |
+
| 2.5128 | 6.1 | 1488000 | 2.4566 |
|
| 235 |
+
| 2.5128 | 6.13 | 1496000 | 2.4499 |
|
| 236 |
+
| 2.5168 | 6.16 | 1504000 | 2.4480 |
|
| 237 |
+
| 2.5168 | 6.2 | 1512000 | 2.4436 |
|
| 238 |
+
| 2.5225 | 6.23 | 1520000 | 2.4467 |
|
| 239 |
+
| 2.5225 | 6.26 | 1528000 | 2.4520 |
|
| 240 |
+
| 2.5135 | 6.29 | 1536000 | 2.4535 |
|
| 241 |
+
| 2.5135 | 6.33 | 1544000 | 2.4463 |
|
| 242 |
+
| 2.5161 | 6.36 | 1552000 | 2.4556 |
|
| 243 |
+
| 2.5161 | 6.39 | 1560000 | 2.4605 |
|
| 244 |
+
| 2.5144 | 6.43 | 1568000 | 2.4516 |
|
| 245 |
+
| 2.5144 | 6.46 | 1576000 | 2.4488 |
|
| 246 |
+
| 2.5209 | 6.49 | 1584000 | 2.4525 |
|
| 247 |
+
| 2.5209 | 6.52 | 1592000 | 2.4502 |
|
| 248 |
+
| 2.5102 | 6.56 | 1600000 | 2.4538 |
|
| 249 |
+
| 2.5102 | 6.59 | 1608000 | 2.4491 |
|
| 250 |
+
| 2.5176 | 6.62 | 1616000 | 2.4528 |
|
| 251 |
+
| 2.5176 | 6.65 | 1624000 | 2.4460 |
|
| 252 |
+
| 2.5208 | 6.69 | 1632000 | 2.4485 |
|
| 253 |
+
| 2.5208 | 6.72 | 1640000 | 2.4513 |
|
| 254 |
+
| 2.5064 | 6.75 | 1648000 | 2.4519 |
|
| 255 |
+
| 2.5064 | 6.79 | 1656000 | 2.4493 |
|
| 256 |
+
| 2.5111 | 6.82 | 1664000 | 2.4505 |
|
| 257 |
+
| 2.5111 | 6.85 | 1672000 | 2.4502 |
|
| 258 |
+
| 2.5141 | 6.88 | 1680000 | 2.4560 |
|
| 259 |
+
| 2.5141 | 6.92 | 1688000 | 2.4500 |
|
| 260 |
+
| 2.5089 | 6.95 | 1696000 | 2.4513 |
|
| 261 |
+
| 2.5089 | 6.98 | 1704000 | 2.4418 |
|
| 262 |
+
| 2.5174 | 7.02 | 1712000 | 2.4477 |
|
| 263 |
+
| 2.5174 | 7.05 | 1720000 | 2.4508 |
|
| 264 |
+
| 2.5198 | 7.08 | 1728000 | 2.4486 |
|
| 265 |
+
| 2.5198 | 7.11 | 1736000 | 2.4577 |
|
| 266 |
+
| 2.4974 | 7.15 | 1744000 | 2.4416 |
|
| 267 |
+
| 2.4974 | 7.18 | 1752000 | 2.4549 |
|
| 268 |
+
| 2.5016 | 7.21 | 1760000 | 2.4557 |
|
| 269 |
+
| 2.5016 | 7.24 | 1768000 | 2.4532 |
|
| 270 |
+
| 2.5112 | 7.28 | 1776000 | 2.4451 |
|
| 271 |
+
| 2.5112 | 7.31 | 1784000 | 2.4607 |
|
| 272 |
+
| 2.5172 | 7.34 | 1792000 | 2.4452 |
|
| 273 |
+
| 2.5172 | 7.38 | 1800000 | 2.4427 |
|
| 274 |
+
| 2.5089 | 7.41 | 1808000 | 2.4511 |
|
| 275 |
+
| 2.5089 | 7.44 | 1816000 | 2.4441 |
|
| 276 |
+
| 2.5136 | 7.47 | 1824000 | 2.4492 |
|
| 277 |
+
| 2.5136 | 7.51 | 1832000 | 2.4524 |
|
| 278 |
+
| 2.509 | 7.54 | 1840000 | 2.4512 |
|
| 279 |
+
| 2.509 | 7.57 | 1848000 | 2.4528 |
|
| 280 |
+
| 2.5157 | 7.61 | 1856000 | 2.4440 |
|
| 281 |
+
| 2.5157 | 7.64 | 1864000 | 2.4402 |
|
| 282 |
+
| 2.5181 | 7.67 | 1872000 | 2.4538 |
|
| 283 |
+
| 2.5181 | 7.7 | 1880000 | 2.4481 |
|
| 284 |
+
| 2.5145 | 7.74 | 1888000 | 2.4417 |
|
| 285 |
+
| 2.5145 | 7.77 | 1896000 | 2.4512 |
|
| 286 |
+
| 2.5013 | 7.8 | 1904000 | 2.4560 |
|
| 287 |
+
| 2.5013 | 7.83 | 1912000 | 2.4509 |
|
| 288 |
+
| 2.5064 | 7.87 | 1920000 | 2.4473 |
|
| 289 |
+
| 2.5064 | 7.9 | 1928000 | 2.4576 |
|
| 290 |
+
| 2.5068 | 7.93 | 1936000 | 2.4461 |
|
| 291 |
+
| 2.5068 | 7.97 | 1944000 | 2.4451 |
|
| 292 |
+
| 2.5152 | 8.0 | 1952000 | 2.4421 |
|
| 293 |
+
| 2.5152 | 8.03 | 1960000 | 2.4458 |
|
| 294 |
+
| 2.5025 | 8.06 | 1968000 | 2.4532 |
|
| 295 |
+
| 2.5025 | 8.1 | 1976000 | 2.4541 |
|
| 296 |
+
| 2.5151 | 8.13 | 1984000 | 2.4499 |
|
| 297 |
+
| 2.5151 | 8.16 | 1992000 | 2.4501 |
|
| 298 |
+
| 2.5138 | 8.2 | 2000000 | 2.4448 |
|
| 299 |
+
| 2.5138 | 8.23 | 2008000 | 2.4562 |
|
| 300 |
+
| 2.5039 | 8.26 | 2016000 | 2.4613 |
|
| 301 |
+
| 2.5039 | 8.29 | 2024000 | 2.4471 |
|
| 302 |
+
| 2.5055 | 8.33 | 2032000 | 2.4450 |
|
| 303 |
+
| 2.5055 | 8.36 | 2040000 | 2.4493 |
|
| 304 |
+
| 2.5085 | 8.39 | 2048000 | 2.4482 |
|
| 305 |
+
| 2.5085 | 8.42 | 2056000 | 2.4572 |
|
| 306 |
+
| 2.5114 | 8.46 | 2064000 | 2.4443 |
|
| 307 |
+
| 2.5114 | 8.49 | 2072000 | 2.4456 |
|
| 308 |
+
| 2.5132 | 8.52 | 2080000 | 2.4528 |
|
| 309 |
+
| 2.5132 | 8.56 | 2088000 | 2.4497 |
|
| 310 |
+
| 2.5072 | 8.59 | 2096000 | 2.4548 |
|
| 311 |
+
| 2.5072 | 8.62 | 2104000 | 2.4548 |
|
| 312 |
+
| 2.504 | 8.65 | 2112000 | 2.4443 |
|
| 313 |
+
| 2.504 | 8.69 | 2120000 | 2.4452 |
|
| 314 |
+
| 2.5128 | 8.72 | 2128000 | 2.4510 |
|
| 315 |
+
| 2.5128 | 8.75 | 2136000 | 2.4480 |
|
| 316 |
+
| 2.5133 | 8.79 | 2144000 | 2.4470 |
|
| 317 |
+
| 2.5133 | 8.82 | 2152000 | 2.4437 |
|
| 318 |
+
| 2.5067 | 8.85 | 2160000 | 2.4447 |
|
| 319 |
+
| 2.5067 | 8.88 | 2168000 | 2.4531 |
|
| 320 |
+
| 2.4996 | 8.92 | 2176000 | 2.4475 |
|
| 321 |
+
| 2.4996 | 8.95 | 2184000 | 2.4438 |
|
| 322 |
+
| 2.5123 | 8.98 | 2192000 | 2.4552 |
|
| 323 |
+
| 2.5123 | 9.01 | 2200000 | 2.4441 |
|
| 324 |
+
| 2.5044 | 9.05 | 2208000 | 2.4438 |
|
| 325 |
+
| 2.5044 | 9.08 | 2216000 | 2.4534 |
|
| 326 |
+
| 2.5068 | 9.11 | 2224000 | 2.4497 |
|
| 327 |
+
| 2.5068 | 9.15 | 2232000 | 2.4440 |
|
| 328 |
+
| 2.5165 | 9.18 | 2240000 | 2.4577 |
|
| 329 |
+
| 2.5165 | 9.21 | 2248000 | 2.4507 |
|
| 330 |
+
| 2.5087 | 9.24 | 2256000 | 2.4494 |
|
| 331 |
+
| 2.5087 | 9.28 | 2264000 | 2.4393 |
|
| 332 |
+
| 2.5036 | 9.31 | 2272000 | 2.4487 |
|
| 333 |
+
| 2.5036 | 9.34 | 2280000 | 2.4423 |
|
| 334 |
+
| 2.5086 | 9.38 | 2288000 | 2.4456 |
|
| 335 |
+
| 2.5086 | 9.41 | 2296000 | 2.4496 |
|
| 336 |
+
| 2.5034 | 9.44 | 2304000 | 2.4499 |
|
| 337 |
+
| 2.5034 | 9.47 | 2312000 | 2.4433 |
|
| 338 |
+
| 2.5099 | 9.51 | 2320000 | 2.4534 |
|
| 339 |
+
| 2.5099 | 9.54 | 2328000 | 2.4495 |
|
| 340 |
+
| 2.5065 | 9.57 | 2336000 | 2.4510 |
|
| 341 |
+
| 2.5065 | 9.6 | 2344000 | 2.4513 |
|
| 342 |
+
| 2.502 | 9.64 | 2352000 | 2.4512 |
|
| 343 |
+
| 2.502 | 9.67 | 2360000 | 2.4469 |
|
| 344 |
+
| 2.5043 | 9.7 | 2368000 | 2.4544 |
|
| 345 |
+
| 2.5043 | 9.74 | 2376000 | 2.4493 |
|
| 346 |
+
| 2.5068 | 9.77 | 2384000 | 2.4537 |
|
| 347 |
+
| 2.5068 | 9.8 | 2392000 | 2.4387 |
|
| 348 |
+
| 2.5118 | 9.83 | 2400000 | 2.4494 |
|
| 349 |
+
|
| 350 |
+
|
| 351 |
+
### Framework versions
|
| 352 |
+
|
| 353 |
+
- Transformers 4.35.0.dev0
|
| 354 |
+
- Pytorch 2.0.1+cu117
|
| 355 |
+
- Datasets 2.14.5
|
| 356 |
+
- Tokenizers 0.14.0
|
added_tokens.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"</s>": 2,
|
| 3 |
+
"<mask>": 50264,
|
| 4 |
+
"<pad>": 1,
|
| 5 |
+
"<s>": 0,
|
| 6 |
+
"<unk>": 3
|
| 7 |
+
}
|
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.83,
|
| 3 |
+
"eval_loss": 2.4514307975769043,
|
| 4 |
+
"eval_runtime": 229.7941,
|
| 5 |
+
"eval_samples": 205510,
|
| 6 |
+
"eval_samples_per_second": 894.323,
|
| 7 |
+
"eval_steps_per_second": 55.898,
|
| 8 |
+
"perplexity": 11.604939165014551,
|
| 9 |
+
"train_loss": 2.5438934391276042,
|
| 10 |
+
"train_runtime": 220778.1092,
|
| 11 |
+
"train_samples": 3904699,
|
| 12 |
+
"train_samples_per_second": 173.93,
|
| 13 |
+
"train_steps_per_second": 10.871
|
| 14 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "cardiffnlp/twitter-roberta-base-2019-90m",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForMaskedLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"gradient_checkpointing": false,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.1,
|
| 13 |
+
"hidden_size": 768,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 3072,
|
| 16 |
+
"layer_norm_eps": 1e-05,
|
| 17 |
+
"max_position_embeddings": 514,
|
| 18 |
+
"model_type": "roberta",
|
| 19 |
+
"num_attention_heads": 12,
|
| 20 |
+
"num_hidden_layers": 12,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"position_embedding_type": "absolute",
|
| 23 |
+
"torch_dtype": "float32",
|
| 24 |
+
"transformers_version": "4.35.0.dev0",
|
| 25 |
+
"type_vocab_size": 1,
|
| 26 |
+
"use_cache": true,
|
| 27 |
+
"vocab_size": 50265
|
| 28 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.83,
|
| 3 |
+
"eval_loss": 2.4514307975769043,
|
| 4 |
+
"eval_runtime": 229.7941,
|
| 5 |
+
"eval_samples": 205510,
|
| 6 |
+
"eval_samples_per_second": 894.323,
|
| 7 |
+
"eval_steps_per_second": 55.898,
|
| 8 |
+
"perplexity": 11.604939165014551
|
| 9 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2c5392ea77766278aace58e2c3db8fc9a8f3441adad7da2c379b10938f3b553
|
| 3 |
+
size 498859189
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "<s>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "<pad>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "</s>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "<unk>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": false,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"50264": {
|
| 37 |
+
"content": "<mask>",
|
| 38 |
+
"lstrip": true,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"additional_special_tokens": [],
|
| 46 |
+
"bos_token": "<s>",
|
| 47 |
+
"clean_up_tokenization_spaces": true,
|
| 48 |
+
"cls_token": "<s>",
|
| 49 |
+
"eos_token": "</s>",
|
| 50 |
+
"errors": "replace",
|
| 51 |
+
"mask_token": "<mask>",
|
| 52 |
+
"max_length": 512,
|
| 53 |
+
"model_max_length": 512,
|
| 54 |
+
"pad_token": "<pad>",
|
| 55 |
+
"sep_token": "</s>",
|
| 56 |
+
"stride": 0,
|
| 57 |
+
"tokenizer_class": "RobertaTokenizer",
|
| 58 |
+
"trim_offsets": true,
|
| 59 |
+
"truncation_side": "right",
|
| 60 |
+
"truncation_strategy": "longest_first",
|
| 61 |
+
"unk_token": "<unk>"
|
| 62 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.83,
|
| 3 |
+
"train_loss": 2.5438934391276042,
|
| 4 |
+
"train_runtime": 220778.1092,
|
| 5 |
+
"train_samples": 3904699,
|
| 6 |
+
"train_samples_per_second": 173.93,
|
| 7 |
+
"train_steps_per_second": 10.871
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3328 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 2.4417428970336914,
|
| 3 |
+
"best_model_checkpoint": "./model_tweets_2020_Q1_50/checkpoint-1888000",
|
| 4 |
+
"epoch": 9.834292176820574,
|
| 5 |
+
"eval_steps": 8000,
|
| 6 |
+
"global_step": 2400000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.03,
|
| 13 |
+
"eval_loss": 2.89373779296875,
|
| 14 |
+
"eval_runtime": 220.7068,
|
| 15 |
+
"eval_samples_per_second": 931.145,
|
| 16 |
+
"eval_steps_per_second": 58.199,
|
| 17 |
+
"step": 8000
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.07,
|
| 21 |
+
"learning_rate": 4.0726666666666665e-07,
|
| 22 |
+
"loss": 3.073,
|
| 23 |
+
"step": 16000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.07,
|
| 27 |
+
"eval_loss": 2.76598858833313,
|
| 28 |
+
"eval_runtime": 221.0774,
|
| 29 |
+
"eval_samples_per_second": 929.584,
|
| 30 |
+
"eval_steps_per_second": 58.102,
|
| 31 |
+
"step": 16000
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.1,
|
| 35 |
+
"eval_loss": 2.7232513427734375,
|
| 36 |
+
"eval_runtime": 221.6288,
|
| 37 |
+
"eval_samples_per_second": 927.271,
|
| 38 |
+
"eval_steps_per_second": 57.957,
|
| 39 |
+
"step": 24000
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.13,
|
| 43 |
+
"learning_rate": 4.0453333333333336e-07,
|
| 44 |
+
"loss": 2.8244,
|
| 45 |
+
"step": 32000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.13,
|
| 49 |
+
"eval_loss": 2.687758207321167,
|
| 50 |
+
"eval_runtime": 220.1262,
|
| 51 |
+
"eval_samples_per_second": 933.601,
|
| 52 |
+
"eval_steps_per_second": 58.353,
|
| 53 |
+
"step": 32000
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.16,
|
| 57 |
+
"eval_loss": 2.6519503593444824,
|
| 58 |
+
"eval_runtime": 220.495,
|
| 59 |
+
"eval_samples_per_second": 932.039,
|
| 60 |
+
"eval_steps_per_second": 58.255,
|
| 61 |
+
"step": 40000
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.2,
|
| 65 |
+
"learning_rate": 4.018e-07,
|
| 66 |
+
"loss": 2.7542,
|
| 67 |
+
"step": 48000
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 0.2,
|
| 71 |
+
"eval_loss": 2.63004469871521,
|
| 72 |
+
"eval_runtime": 220.2803,
|
| 73 |
+
"eval_samples_per_second": 932.948,
|
| 74 |
+
"eval_steps_per_second": 58.312,
|
| 75 |
+
"step": 48000
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.23,
|
| 79 |
+
"eval_loss": 2.613522529602051,
|
| 80 |
+
"eval_runtime": 221.2317,
|
| 81 |
+
"eval_samples_per_second": 928.936,
|
| 82 |
+
"eval_steps_per_second": 58.061,
|
| 83 |
+
"step": 56000
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.26,
|
| 87 |
+
"learning_rate": 3.9906666666666667e-07,
|
| 88 |
+
"loss": 2.7083,
|
| 89 |
+
"step": 64000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.26,
|
| 93 |
+
"eval_loss": 2.6067709922790527,
|
| 94 |
+
"eval_runtime": 220.3177,
|
| 95 |
+
"eval_samples_per_second": 932.789,
|
| 96 |
+
"eval_steps_per_second": 58.302,
|
| 97 |
+
"step": 64000
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 0.3,
|
| 101 |
+
"eval_loss": 2.5854294300079346,
|
| 102 |
+
"eval_runtime": 220.8061,
|
| 103 |
+
"eval_samples_per_second": 930.726,
|
| 104 |
+
"eval_steps_per_second": 58.173,
|
| 105 |
+
"step": 72000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.33,
|
| 109 |
+
"learning_rate": 3.963333333333333e-07,
|
| 110 |
+
"loss": 2.6752,
|
| 111 |
+
"step": 80000
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.33,
|
| 115 |
+
"eval_loss": 2.575528860092163,
|
| 116 |
+
"eval_runtime": 221.8521,
|
| 117 |
+
"eval_samples_per_second": 926.338,
|
| 118 |
+
"eval_steps_per_second": 57.899,
|
| 119 |
+
"step": 80000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.36,
|
| 123 |
+
"eval_loss": 2.5720720291137695,
|
| 124 |
+
"eval_runtime": 221.4472,
|
| 125 |
+
"eval_samples_per_second": 928.032,
|
| 126 |
+
"eval_steps_per_second": 58.005,
|
| 127 |
+
"step": 88000
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.39,
|
| 131 |
+
"learning_rate": 3.936e-07,
|
| 132 |
+
"loss": 2.6657,
|
| 133 |
+
"step": 96000
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 0.39,
|
| 137 |
+
"eval_loss": 2.5709290504455566,
|
| 138 |
+
"eval_runtime": 220.9006,
|
| 139 |
+
"eval_samples_per_second": 930.328,
|
| 140 |
+
"eval_steps_per_second": 58.148,
|
| 141 |
+
"step": 96000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.43,
|
| 145 |
+
"eval_loss": 2.5656096935272217,
|
| 146 |
+
"eval_runtime": 220.5433,
|
| 147 |
+
"eval_samples_per_second": 931.835,
|
| 148 |
+
"eval_steps_per_second": 58.243,
|
| 149 |
+
"step": 104000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.46,
|
| 153 |
+
"learning_rate": 3.908666666666667e-07,
|
| 154 |
+
"loss": 2.6534,
|
| 155 |
+
"step": 112000
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 0.46,
|
| 159 |
+
"eval_loss": 2.5558407306671143,
|
| 160 |
+
"eval_runtime": 221.6371,
|
| 161 |
+
"eval_samples_per_second": 927.236,
|
| 162 |
+
"eval_steps_per_second": 57.955,
|
| 163 |
+
"step": 112000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.49,
|
| 167 |
+
"eval_loss": 2.5495829582214355,
|
| 168 |
+
"eval_runtime": 220.7733,
|
| 169 |
+
"eval_samples_per_second": 930.864,
|
| 170 |
+
"eval_steps_per_second": 58.182,
|
| 171 |
+
"step": 120000
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.52,
|
| 175 |
+
"learning_rate": 3.8813333333333334e-07,
|
| 176 |
+
"loss": 2.646,
|
| 177 |
+
"step": 128000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.52,
|
| 181 |
+
"eval_loss": 2.547106981277466,
|
| 182 |
+
"eval_runtime": 221.6448,
|
| 183 |
+
"eval_samples_per_second": 927.204,
|
| 184 |
+
"eval_steps_per_second": 57.953,
|
| 185 |
+
"step": 128000
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.56,
|
| 189 |
+
"eval_loss": 2.5408244132995605,
|
| 190 |
+
"eval_runtime": 221.6302,
|
| 191 |
+
"eval_samples_per_second": 927.265,
|
| 192 |
+
"eval_steps_per_second": 57.957,
|
| 193 |
+
"step": 136000
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.59,
|
| 197 |
+
"learning_rate": 3.854e-07,
|
| 198 |
+
"loss": 2.625,
|
| 199 |
+
"step": 144000
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.59,
|
| 203 |
+
"eval_loss": 2.531517744064331,
|
| 204 |
+
"eval_runtime": 223.6683,
|
| 205 |
+
"eval_samples_per_second": 918.816,
|
| 206 |
+
"eval_steps_per_second": 57.429,
|
| 207 |
+
"step": 144000
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.62,
|
| 211 |
+
"eval_loss": 2.5364675521850586,
|
| 212 |
+
"eval_runtime": 224.1465,
|
| 213 |
+
"eval_samples_per_second": 916.856,
|
| 214 |
+
"eval_steps_per_second": 57.306,
|
| 215 |
+
"step": 152000
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.66,
|
| 219 |
+
"learning_rate": 3.8266666666666665e-07,
|
| 220 |
+
"loss": 2.6222,
|
| 221 |
+
"step": 160000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.66,
|
| 225 |
+
"eval_loss": 2.5372273921966553,
|
| 226 |
+
"eval_runtime": 221.7325,
|
| 227 |
+
"eval_samples_per_second": 926.837,
|
| 228 |
+
"eval_steps_per_second": 57.93,
|
| 229 |
+
"step": 160000
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 0.69,
|
| 233 |
+
"eval_loss": 2.534186363220215,
|
| 234 |
+
"eval_runtime": 222.3525,
|
| 235 |
+
"eval_samples_per_second": 924.253,
|
| 236 |
+
"eval_steps_per_second": 57.769,
|
| 237 |
+
"step": 168000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.72,
|
| 241 |
+
"learning_rate": 3.799333333333333e-07,
|
| 242 |
+
"loss": 2.6256,
|
| 243 |
+
"step": 176000
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.72,
|
| 247 |
+
"eval_loss": 2.5308265686035156,
|
| 248 |
+
"eval_runtime": 221.538,
|
| 249 |
+
"eval_samples_per_second": 927.651,
|
| 250 |
+
"eval_steps_per_second": 57.981,
|
| 251 |
+
"step": 176000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.75,
|
| 255 |
+
"eval_loss": 2.5311617851257324,
|
| 256 |
+
"eval_runtime": 224.2919,
|
| 257 |
+
"eval_samples_per_second": 916.261,
|
| 258 |
+
"eval_steps_per_second": 57.269,
|
| 259 |
+
"step": 184000
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 0.79,
|
| 263 |
+
"learning_rate": 3.772e-07,
|
| 264 |
+
"loss": 2.6074,
|
| 265 |
+
"step": 192000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.79,
|
| 269 |
+
"eval_loss": 2.522848129272461,
|
| 270 |
+
"eval_runtime": 224.4507,
|
| 271 |
+
"eval_samples_per_second": 915.613,
|
| 272 |
+
"eval_steps_per_second": 57.229,
|
| 273 |
+
"step": 192000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.82,
|
| 277 |
+
"eval_loss": 2.529161214828491,
|
| 278 |
+
"eval_runtime": 222.7477,
|
| 279 |
+
"eval_samples_per_second": 922.613,
|
| 280 |
+
"eval_steps_per_second": 57.666,
|
| 281 |
+
"step": 200000
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.85,
|
| 285 |
+
"learning_rate": 3.7446666666666667e-07,
|
| 286 |
+
"loss": 2.6071,
|
| 287 |
+
"step": 208000
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.85,
|
| 291 |
+
"eval_loss": 2.5295047760009766,
|
| 292 |
+
"eval_runtime": 223.0891,
|
| 293 |
+
"eval_samples_per_second": 921.201,
|
| 294 |
+
"eval_steps_per_second": 57.578,
|
| 295 |
+
"step": 208000
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.89,
|
| 299 |
+
"eval_loss": 2.523491621017456,
|
| 300 |
+
"eval_runtime": 221.4007,
|
| 301 |
+
"eval_samples_per_second": 928.227,
|
| 302 |
+
"eval_steps_per_second": 58.017,
|
| 303 |
+
"step": 216000
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.92,
|
| 307 |
+
"learning_rate": 3.7173333333333333e-07,
|
| 308 |
+
"loss": 2.5955,
|
| 309 |
+
"step": 224000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.92,
|
| 313 |
+
"eval_loss": 2.5219199657440186,
|
| 314 |
+
"eval_runtime": 221.3605,
|
| 315 |
+
"eval_samples_per_second": 928.395,
|
| 316 |
+
"eval_steps_per_second": 58.028,
|
| 317 |
+
"step": 224000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.95,
|
| 321 |
+
"eval_loss": 2.5190882682800293,
|
| 322 |
+
"eval_runtime": 221.1449,
|
| 323 |
+
"eval_samples_per_second": 929.3,
|
| 324 |
+
"eval_steps_per_second": 58.084,
|
| 325 |
+
"step": 232000
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.98,
|
| 329 |
+
"learning_rate": 3.69e-07,
|
| 330 |
+
"loss": 2.6036,
|
| 331 |
+
"step": 240000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.98,
|
| 335 |
+
"eval_loss": 2.517120361328125,
|
| 336 |
+
"eval_runtime": 220.9198,
|
| 337 |
+
"eval_samples_per_second": 930.247,
|
| 338 |
+
"eval_steps_per_second": 58.143,
|
| 339 |
+
"step": 240000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.02,
|
| 343 |
+
"eval_loss": 2.5102434158325195,
|
| 344 |
+
"eval_runtime": 221.7647,
|
| 345 |
+
"eval_samples_per_second": 926.703,
|
| 346 |
+
"eval_steps_per_second": 57.922,
|
| 347 |
+
"step": 248000
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.05,
|
| 351 |
+
"learning_rate": 3.6626666666666664e-07,
|
| 352 |
+
"loss": 2.6046,
|
| 353 |
+
"step": 256000
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.05,
|
| 357 |
+
"eval_loss": 2.5070137977600098,
|
| 358 |
+
"eval_runtime": 221.6584,
|
| 359 |
+
"eval_samples_per_second": 927.147,
|
| 360 |
+
"eval_steps_per_second": 57.95,
|
| 361 |
+
"step": 256000
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 1.08,
|
| 365 |
+
"eval_loss": 2.5109376907348633,
|
| 366 |
+
"eval_runtime": 221.3382,
|
| 367 |
+
"eval_samples_per_second": 928.489,
|
| 368 |
+
"eval_steps_per_second": 58.033,
|
| 369 |
+
"step": 264000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.11,
|
| 373 |
+
"learning_rate": 3.6353333333333335e-07,
|
| 374 |
+
"loss": 2.5892,
|
| 375 |
+
"step": 272000
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.11,
|
| 379 |
+
"eval_loss": 2.5104565620422363,
|
| 380 |
+
"eval_runtime": 222.1683,
|
| 381 |
+
"eval_samples_per_second": 925.019,
|
| 382 |
+
"eval_steps_per_second": 57.817,
|
| 383 |
+
"step": 272000
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.15,
|
| 387 |
+
"eval_loss": 2.508704423904419,
|
| 388 |
+
"eval_runtime": 222.9629,
|
| 389 |
+
"eval_samples_per_second": 921.723,
|
| 390 |
+
"eval_steps_per_second": 57.61,
|
| 391 |
+
"step": 280000
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 1.18,
|
| 395 |
+
"learning_rate": 3.608e-07,
|
| 396 |
+
"loss": 2.5929,
|
| 397 |
+
"step": 288000
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 1.18,
|
| 401 |
+
"eval_loss": 2.509392738342285,
|
| 402 |
+
"eval_runtime": 223.3494,
|
| 403 |
+
"eval_samples_per_second": 920.128,
|
| 404 |
+
"eval_steps_per_second": 57.511,
|
| 405 |
+
"step": 288000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 1.21,
|
| 409 |
+
"eval_loss": 2.508585214614868,
|
| 410 |
+
"eval_runtime": 222.7314,
|
| 411 |
+
"eval_samples_per_second": 922.681,
|
| 412 |
+
"eval_steps_per_second": 57.67,
|
| 413 |
+
"step": 296000
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.25,
|
| 417 |
+
"learning_rate": 3.5806666666666666e-07,
|
| 418 |
+
"loss": 2.5857,
|
| 419 |
+
"step": 304000
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.25,
|
| 423 |
+
"eval_loss": 2.4991345405578613,
|
| 424 |
+
"eval_runtime": 223.3332,
|
| 425 |
+
"eval_samples_per_second": 920.195,
|
| 426 |
+
"eval_steps_per_second": 57.515,
|
| 427 |
+
"step": 304000
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 1.28,
|
| 431 |
+
"eval_loss": 2.508927822113037,
|
| 432 |
+
"eval_runtime": 224.1404,
|
| 433 |
+
"eval_samples_per_second": 916.881,
|
| 434 |
+
"eval_steps_per_second": 57.308,
|
| 435 |
+
"step": 312000
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.31,
|
| 439 |
+
"learning_rate": 3.553333333333333e-07,
|
| 440 |
+
"loss": 2.5828,
|
| 441 |
+
"step": 320000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.31,
|
| 445 |
+
"eval_loss": 2.501734972000122,
|
| 446 |
+
"eval_runtime": 223.1146,
|
| 447 |
+
"eval_samples_per_second": 921.096,
|
| 448 |
+
"eval_steps_per_second": 57.571,
|
| 449 |
+
"step": 320000
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 1.34,
|
| 453 |
+
"eval_loss": 2.503918409347534,
|
| 454 |
+
"eval_runtime": 223.3327,
|
| 455 |
+
"eval_samples_per_second": 920.196,
|
| 456 |
+
"eval_steps_per_second": 57.515,
|
| 457 |
+
"step": 328000
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.38,
|
| 461 |
+
"learning_rate": 3.5259999999999997e-07,
|
| 462 |
+
"loss": 2.5812,
|
| 463 |
+
"step": 336000
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 1.38,
|
| 467 |
+
"eval_loss": 2.5064587593078613,
|
| 468 |
+
"eval_runtime": 224.1587,
|
| 469 |
+
"eval_samples_per_second": 916.806,
|
| 470 |
+
"eval_steps_per_second": 57.303,
|
| 471 |
+
"step": 336000
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.41,
|
| 475 |
+
"eval_loss": 2.508263111114502,
|
| 476 |
+
"eval_runtime": 222.503,
|
| 477 |
+
"eval_samples_per_second": 923.628,
|
| 478 |
+
"eval_steps_per_second": 57.73,
|
| 479 |
+
"step": 344000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.44,
|
| 483 |
+
"learning_rate": 3.498666666666667e-07,
|
| 484 |
+
"loss": 2.5775,
|
| 485 |
+
"step": 352000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.44,
|
| 489 |
+
"eval_loss": 2.509936571121216,
|
| 490 |
+
"eval_runtime": 223.039,
|
| 491 |
+
"eval_samples_per_second": 921.408,
|
| 492 |
+
"eval_steps_per_second": 57.591,
|
| 493 |
+
"step": 352000
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 1.48,
|
| 497 |
+
"eval_loss": 2.5078811645507812,
|
| 498 |
+
"eval_runtime": 221.7646,
|
| 499 |
+
"eval_samples_per_second": 926.703,
|
| 500 |
+
"eval_steps_per_second": 57.922,
|
| 501 |
+
"step": 360000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 1.51,
|
| 505 |
+
"learning_rate": 3.4713333333333333e-07,
|
| 506 |
+
"loss": 2.5711,
|
| 507 |
+
"step": 368000
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.51,
|
| 511 |
+
"eval_loss": 2.4922046661376953,
|
| 512 |
+
"eval_runtime": 223.0544,
|
| 513 |
+
"eval_samples_per_second": 921.345,
|
| 514 |
+
"eval_steps_per_second": 57.587,
|
| 515 |
+
"step": 368000
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 1.54,
|
| 519 |
+
"eval_loss": 2.5012030601501465,
|
| 520 |
+
"eval_runtime": 222.0392,
|
| 521 |
+
"eval_samples_per_second": 925.557,
|
| 522 |
+
"eval_steps_per_second": 57.85,
|
| 523 |
+
"step": 376000
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 1.57,
|
| 527 |
+
"learning_rate": 3.444e-07,
|
| 528 |
+
"loss": 2.5797,
|
| 529 |
+
"step": 384000
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 1.57,
|
| 533 |
+
"eval_loss": 2.49989914894104,
|
| 534 |
+
"eval_runtime": 223.8829,
|
| 535 |
+
"eval_samples_per_second": 917.935,
|
| 536 |
+
"eval_steps_per_second": 57.374,
|
| 537 |
+
"step": 384000
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.61,
|
| 541 |
+
"eval_loss": 2.4881107807159424,
|
| 542 |
+
"eval_runtime": 222.4413,
|
| 543 |
+
"eval_samples_per_second": 923.884,
|
| 544 |
+
"eval_steps_per_second": 57.746,
|
| 545 |
+
"step": 392000
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 1.64,
|
| 549 |
+
"learning_rate": 3.416666666666667e-07,
|
| 550 |
+
"loss": 2.5718,
|
| 551 |
+
"step": 400000
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 1.64,
|
| 555 |
+
"eval_loss": 2.4960451126098633,
|
| 556 |
+
"eval_runtime": 222.8741,
|
| 557 |
+
"eval_samples_per_second": 922.09,
|
| 558 |
+
"eval_steps_per_second": 57.633,
|
| 559 |
+
"step": 400000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.67,
|
| 563 |
+
"eval_loss": 2.490837574005127,
|
| 564 |
+
"eval_runtime": 222.3679,
|
| 565 |
+
"eval_samples_per_second": 924.189,
|
| 566 |
+
"eval_steps_per_second": 57.765,
|
| 567 |
+
"step": 408000
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.7,
|
| 571 |
+
"learning_rate": 3.3893333333333335e-07,
|
| 572 |
+
"loss": 2.5627,
|
| 573 |
+
"step": 416000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.7,
|
| 577 |
+
"eval_loss": 2.4970648288726807,
|
| 578 |
+
"eval_runtime": 223.472,
|
| 579 |
+
"eval_samples_per_second": 919.623,
|
| 580 |
+
"eval_steps_per_second": 57.479,
|
| 581 |
+
"step": 416000
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 1.74,
|
| 585 |
+
"eval_loss": 2.4916465282440186,
|
| 586 |
+
"eval_runtime": 222.5109,
|
| 587 |
+
"eval_samples_per_second": 923.595,
|
| 588 |
+
"eval_steps_per_second": 57.728,
|
| 589 |
+
"step": 424000
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 1.77,
|
| 593 |
+
"learning_rate": 3.3619999999999995e-07,
|
| 594 |
+
"loss": 2.5641,
|
| 595 |
+
"step": 432000
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.77,
|
| 599 |
+
"eval_loss": 2.4971389770507812,
|
| 600 |
+
"eval_runtime": 222.1533,
|
| 601 |
+
"eval_samples_per_second": 925.082,
|
| 602 |
+
"eval_steps_per_second": 57.82,
|
| 603 |
+
"step": 432000
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.8,
|
| 607 |
+
"eval_loss": 2.495426654815674,
|
| 608 |
+
"eval_runtime": 223.2728,
|
| 609 |
+
"eval_samples_per_second": 920.444,
|
| 610 |
+
"eval_steps_per_second": 57.531,
|
| 611 |
+
"step": 440000
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 1.84,
|
| 615 |
+
"learning_rate": 3.3346666666666666e-07,
|
| 616 |
+
"loss": 2.5633,
|
| 617 |
+
"step": 448000
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 1.84,
|
| 621 |
+
"eval_loss": 2.485994815826416,
|
| 622 |
+
"eval_runtime": 222.7264,
|
| 623 |
+
"eval_samples_per_second": 922.702,
|
| 624 |
+
"eval_steps_per_second": 57.672,
|
| 625 |
+
"step": 448000
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.87,
|
| 629 |
+
"eval_loss": 2.4893651008605957,
|
| 630 |
+
"eval_runtime": 223.4251,
|
| 631 |
+
"eval_samples_per_second": 919.816,
|
| 632 |
+
"eval_steps_per_second": 57.491,
|
| 633 |
+
"step": 456000
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.9,
|
| 637 |
+
"learning_rate": 3.307333333333333e-07,
|
| 638 |
+
"loss": 2.5676,
|
| 639 |
+
"step": 464000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 1.9,
|
| 643 |
+
"eval_loss": 2.489337205886841,
|
| 644 |
+
"eval_runtime": 222.9423,
|
| 645 |
+
"eval_samples_per_second": 921.808,
|
| 646 |
+
"eval_steps_per_second": 57.616,
|
| 647 |
+
"step": 464000
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.93,
|
| 651 |
+
"eval_loss": 2.4883553981781006,
|
| 652 |
+
"eval_runtime": 223.2404,
|
| 653 |
+
"eval_samples_per_second": 920.577,
|
| 654 |
+
"eval_steps_per_second": 57.539,
|
| 655 |
+
"step": 472000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 1.97,
|
| 659 |
+
"learning_rate": 3.28e-07,
|
| 660 |
+
"loss": 2.5687,
|
| 661 |
+
"step": 480000
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.97,
|
| 665 |
+
"eval_loss": 2.4921038150787354,
|
| 666 |
+
"eval_runtime": 223.8809,
|
| 667 |
+
"eval_samples_per_second": 917.943,
|
| 668 |
+
"eval_steps_per_second": 57.374,
|
| 669 |
+
"step": 480000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 2.0,
|
| 673 |
+
"eval_loss": 2.4873294830322266,
|
| 674 |
+
"eval_runtime": 222.8771,
|
| 675 |
+
"eval_samples_per_second": 922.078,
|
| 676 |
+
"eval_steps_per_second": 57.633,
|
| 677 |
+
"step": 488000
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 2.03,
|
| 681 |
+
"learning_rate": 3.252666666666667e-07,
|
| 682 |
+
"loss": 2.5633,
|
| 683 |
+
"step": 496000
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 2.03,
|
| 687 |
+
"eval_loss": 2.4919497966766357,
|
| 688 |
+
"eval_runtime": 222.6439,
|
| 689 |
+
"eval_samples_per_second": 923.043,
|
| 690 |
+
"eval_steps_per_second": 57.693,
|
| 691 |
+
"step": 496000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 2.07,
|
| 695 |
+
"eval_loss": 2.482137441635132,
|
| 696 |
+
"eval_runtime": 222.747,
|
| 697 |
+
"eval_samples_per_second": 922.616,
|
| 698 |
+
"eval_steps_per_second": 57.666,
|
| 699 |
+
"step": 504000
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.1,
|
| 703 |
+
"learning_rate": 3.2253333333333334e-07,
|
| 704 |
+
"loss": 2.5547,
|
| 705 |
+
"step": 512000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 2.1,
|
| 709 |
+
"eval_loss": 2.490872621536255,
|
| 710 |
+
"eval_runtime": 222.6765,
|
| 711 |
+
"eval_samples_per_second": 922.908,
|
| 712 |
+
"eval_steps_per_second": 57.685,
|
| 713 |
+
"step": 512000
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 2.13,
|
| 717 |
+
"eval_loss": 2.4818356037139893,
|
| 718 |
+
"eval_runtime": 223.7166,
|
| 719 |
+
"eval_samples_per_second": 918.617,
|
| 720 |
+
"eval_steps_per_second": 57.416,
|
| 721 |
+
"step": 520000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 2.16,
|
| 725 |
+
"learning_rate": 3.198e-07,
|
| 726 |
+
"loss": 2.5617,
|
| 727 |
+
"step": 528000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 2.16,
|
| 731 |
+
"eval_loss": 2.4854869842529297,
|
| 732 |
+
"eval_runtime": 223.7715,
|
| 733 |
+
"eval_samples_per_second": 918.392,
|
| 734 |
+
"eval_steps_per_second": 57.402,
|
| 735 |
+
"step": 528000
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 2.2,
|
| 739 |
+
"eval_loss": 2.48504638671875,
|
| 740 |
+
"eval_runtime": 223.6654,
|
| 741 |
+
"eval_samples_per_second": 918.828,
|
| 742 |
+
"eval_steps_per_second": 57.43,
|
| 743 |
+
"step": 536000
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 2.23,
|
| 747 |
+
"learning_rate": 3.1706666666666665e-07,
|
| 748 |
+
"loss": 2.5569,
|
| 749 |
+
"step": 544000
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 2.23,
|
| 753 |
+
"eval_loss": 2.480282783508301,
|
| 754 |
+
"eval_runtime": 222.7744,
|
| 755 |
+
"eval_samples_per_second": 922.503,
|
| 756 |
+
"eval_steps_per_second": 57.659,
|
| 757 |
+
"step": 544000
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 2.26,
|
| 761 |
+
"eval_loss": 2.4775896072387695,
|
| 762 |
+
"eval_runtime": 223.0018,
|
| 763 |
+
"eval_samples_per_second": 921.562,
|
| 764 |
+
"eval_steps_per_second": 57.6,
|
| 765 |
+
"step": 552000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.29,
|
| 769 |
+
"learning_rate": 3.1433333333333336e-07,
|
| 770 |
+
"loss": 2.5535,
|
| 771 |
+
"step": 560000
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 2.29,
|
| 775 |
+
"eval_loss": 2.4824471473693848,
|
| 776 |
+
"eval_runtime": 223.1733,
|
| 777 |
+
"eval_samples_per_second": 920.854,
|
| 778 |
+
"eval_steps_per_second": 57.556,
|
| 779 |
+
"step": 560000
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 2.33,
|
| 783 |
+
"eval_loss": 2.4821510314941406,
|
| 784 |
+
"eval_runtime": 224.1586,
|
| 785 |
+
"eval_samples_per_second": 916.806,
|
| 786 |
+
"eval_steps_per_second": 57.303,
|
| 787 |
+
"step": 568000
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 2.36,
|
| 791 |
+
"learning_rate": 3.116e-07,
|
| 792 |
+
"loss": 2.5534,
|
| 793 |
+
"step": 576000
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 2.36,
|
| 797 |
+
"eval_loss": 2.476337432861328,
|
| 798 |
+
"eval_runtime": 223.4733,
|
| 799 |
+
"eval_samples_per_second": 919.618,
|
| 800 |
+
"eval_steps_per_second": 57.479,
|
| 801 |
+
"step": 576000
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 2.39,
|
| 805 |
+
"eval_loss": 2.47969388961792,
|
| 806 |
+
"eval_runtime": 224.2217,
|
| 807 |
+
"eval_samples_per_second": 916.548,
|
| 808 |
+
"eval_steps_per_second": 57.287,
|
| 809 |
+
"step": 584000
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 2.43,
|
| 813 |
+
"learning_rate": 3.0886666666666667e-07,
|
| 814 |
+
"loss": 2.5583,
|
| 815 |
+
"step": 592000
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 2.43,
|
| 819 |
+
"eval_loss": 2.4872305393218994,
|
| 820 |
+
"eval_runtime": 224.237,
|
| 821 |
+
"eval_samples_per_second": 916.486,
|
| 822 |
+
"eval_steps_per_second": 57.283,
|
| 823 |
+
"step": 592000
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 2.46,
|
| 827 |
+
"eval_loss": 2.4812192916870117,
|
| 828 |
+
"eval_runtime": 222.6272,
|
| 829 |
+
"eval_samples_per_second": 923.113,
|
| 830 |
+
"eval_steps_per_second": 57.697,
|
| 831 |
+
"step": 600000
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 2.49,
|
| 835 |
+
"learning_rate": 3.061333333333333e-07,
|
| 836 |
+
"loss": 2.5545,
|
| 837 |
+
"step": 608000
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.49,
|
| 841 |
+
"eval_loss": 2.474827527999878,
|
| 842 |
+
"eval_runtime": 223.4042,
|
| 843 |
+
"eval_samples_per_second": 919.902,
|
| 844 |
+
"eval_steps_per_second": 57.497,
|
| 845 |
+
"step": 608000
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 2.52,
|
| 849 |
+
"eval_loss": 2.4735865592956543,
|
| 850 |
+
"eval_runtime": 224.1504,
|
| 851 |
+
"eval_samples_per_second": 916.84,
|
| 852 |
+
"eval_steps_per_second": 57.305,
|
| 853 |
+
"step": 616000
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 2.56,
|
| 857 |
+
"learning_rate": 3.034e-07,
|
| 858 |
+
"loss": 2.5561,
|
| 859 |
+
"step": 624000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 2.56,
|
| 863 |
+
"eval_loss": 2.4714128971099854,
|
| 864 |
+
"eval_runtime": 223.2085,
|
| 865 |
+
"eval_samples_per_second": 920.709,
|
| 866 |
+
"eval_steps_per_second": 57.547,
|
| 867 |
+
"step": 624000
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 2.59,
|
| 871 |
+
"eval_loss": 2.485759973526001,
|
| 872 |
+
"eval_runtime": 222.8361,
|
| 873 |
+
"eval_samples_per_second": 922.247,
|
| 874 |
+
"eval_steps_per_second": 57.643,
|
| 875 |
+
"step": 632000
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 2.62,
|
| 879 |
+
"learning_rate": 3.0066666666666663e-07,
|
| 880 |
+
"loss": 2.5384,
|
| 881 |
+
"step": 640000
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 2.62,
|
| 885 |
+
"eval_loss": 2.482938289642334,
|
| 886 |
+
"eval_runtime": 223.4494,
|
| 887 |
+
"eval_samples_per_second": 919.716,
|
| 888 |
+
"eval_steps_per_second": 57.485,
|
| 889 |
+
"step": 640000
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 2.66,
|
| 893 |
+
"eval_loss": 2.47662091255188,
|
| 894 |
+
"eval_runtime": 222.9171,
|
| 895 |
+
"eval_samples_per_second": 921.912,
|
| 896 |
+
"eval_steps_per_second": 57.622,
|
| 897 |
+
"step": 648000
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 2.69,
|
| 901 |
+
"learning_rate": 2.9793333333333334e-07,
|
| 902 |
+
"loss": 2.541,
|
| 903 |
+
"step": 656000
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 2.69,
|
| 907 |
+
"eval_loss": 2.4835963249206543,
|
| 908 |
+
"eval_runtime": 223.4062,
|
| 909 |
+
"eval_samples_per_second": 919.894,
|
| 910 |
+
"eval_steps_per_second": 57.496,
|
| 911 |
+
"step": 656000
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 2.72,
|
| 915 |
+
"eval_loss": 2.465118408203125,
|
| 916 |
+
"eval_runtime": 226.1239,
|
| 917 |
+
"eval_samples_per_second": 908.838,
|
| 918 |
+
"eval_steps_per_second": 56.805,
|
| 919 |
+
"step": 664000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 2.75,
|
| 923 |
+
"learning_rate": 2.952e-07,
|
| 924 |
+
"loss": 2.5439,
|
| 925 |
+
"step": 672000
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 2.75,
|
| 929 |
+
"eval_loss": 2.4797005653381348,
|
| 930 |
+
"eval_runtime": 224.1173,
|
| 931 |
+
"eval_samples_per_second": 916.975,
|
| 932 |
+
"eval_steps_per_second": 57.314,
|
| 933 |
+
"step": 672000
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.79,
|
| 937 |
+
"eval_loss": 2.4702000617980957,
|
| 938 |
+
"eval_runtime": 223.8532,
|
| 939 |
+
"eval_samples_per_second": 918.057,
|
| 940 |
+
"eval_steps_per_second": 57.381,
|
| 941 |
+
"step": 680000
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 2.82,
|
| 945 |
+
"learning_rate": 2.9246666666666665e-07,
|
| 946 |
+
"loss": 2.5597,
|
| 947 |
+
"step": 688000
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 2.82,
|
| 951 |
+
"eval_loss": 2.475144386291504,
|
| 952 |
+
"eval_runtime": 223.4589,
|
| 953 |
+
"eval_samples_per_second": 919.677,
|
| 954 |
+
"eval_steps_per_second": 57.483,
|
| 955 |
+
"step": 688000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 2.85,
|
| 959 |
+
"eval_loss": 2.474367618560791,
|
| 960 |
+
"eval_runtime": 222.9092,
|
| 961 |
+
"eval_samples_per_second": 921.945,
|
| 962 |
+
"eval_steps_per_second": 57.624,
|
| 963 |
+
"step": 696000
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 2.88,
|
| 967 |
+
"learning_rate": 2.897333333333333e-07,
|
| 968 |
+
"loss": 2.5491,
|
| 969 |
+
"step": 704000
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 2.88,
|
| 973 |
+
"eval_loss": 2.4756221771240234,
|
| 974 |
+
"eval_runtime": 223.5443,
|
| 975 |
+
"eval_samples_per_second": 919.325,
|
| 976 |
+
"eval_steps_per_second": 57.461,
|
| 977 |
+
"step": 704000
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 2.92,
|
| 981 |
+
"eval_loss": 2.4731247425079346,
|
| 982 |
+
"eval_runtime": 223.5397,
|
| 983 |
+
"eval_samples_per_second": 919.345,
|
| 984 |
+
"eval_steps_per_second": 57.462,
|
| 985 |
+
"step": 712000
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 2.95,
|
| 989 |
+
"learning_rate": 2.8699999999999996e-07,
|
| 990 |
+
"loss": 2.5505,
|
| 991 |
+
"step": 720000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 2.95,
|
| 995 |
+
"eval_loss": 2.475615978240967,
|
| 996 |
+
"eval_runtime": 223.941,
|
| 997 |
+
"eval_samples_per_second": 917.697,
|
| 998 |
+
"eval_steps_per_second": 57.359,
|
| 999 |
+
"step": 720000
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 2.98,
|
| 1003 |
+
"eval_loss": 2.4703986644744873,
|
| 1004 |
+
"eval_runtime": 224.1288,
|
| 1005 |
+
"eval_samples_per_second": 916.928,
|
| 1006 |
+
"eval_steps_per_second": 57.311,
|
| 1007 |
+
"step": 728000
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 3.02,
|
| 1011 |
+
"learning_rate": 2.8426666666666667e-07,
|
| 1012 |
+
"loss": 2.5432,
|
| 1013 |
+
"step": 736000
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 3.02,
|
| 1017 |
+
"eval_loss": 2.4762611389160156,
|
| 1018 |
+
"eval_runtime": 223.7009,
|
| 1019 |
+
"eval_samples_per_second": 918.682,
|
| 1020 |
+
"eval_steps_per_second": 57.42,
|
| 1021 |
+
"step": 736000
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 3.05,
|
| 1025 |
+
"eval_loss": 2.4743261337280273,
|
| 1026 |
+
"eval_runtime": 224.407,
|
| 1027 |
+
"eval_samples_per_second": 915.791,
|
| 1028 |
+
"eval_steps_per_second": 57.24,
|
| 1029 |
+
"step": 744000
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 3.08,
|
| 1033 |
+
"learning_rate": 2.815333333333333e-07,
|
| 1034 |
+
"loss": 2.5485,
|
| 1035 |
+
"step": 752000
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 3.08,
|
| 1039 |
+
"eval_loss": 2.4626660346984863,
|
| 1040 |
+
"eval_runtime": 224.0612,
|
| 1041 |
+
"eval_samples_per_second": 917.205,
|
| 1042 |
+
"eval_steps_per_second": 57.328,
|
| 1043 |
+
"step": 752000
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 3.11,
|
| 1047 |
+
"eval_loss": 2.471444606781006,
|
| 1048 |
+
"eval_runtime": 223.4318,
|
| 1049 |
+
"eval_samples_per_second": 919.788,
|
| 1050 |
+
"eval_steps_per_second": 57.49,
|
| 1051 |
+
"step": 760000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 3.15,
|
| 1055 |
+
"learning_rate": 2.7880000000000003e-07,
|
| 1056 |
+
"loss": 2.5482,
|
| 1057 |
+
"step": 768000
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 3.15,
|
| 1061 |
+
"eval_loss": 2.4684672355651855,
|
| 1062 |
+
"eval_runtime": 224.1026,
|
| 1063 |
+
"eval_samples_per_second": 917.035,
|
| 1064 |
+
"eval_steps_per_second": 57.317,
|
| 1065 |
+
"step": 768000
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 3.18,
|
| 1069 |
+
"eval_loss": 2.4672694206237793,
|
| 1070 |
+
"eval_runtime": 224.9545,
|
| 1071 |
+
"eval_samples_per_second": 913.562,
|
| 1072 |
+
"eval_steps_per_second": 57.1,
|
| 1073 |
+
"step": 776000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 3.21,
|
| 1077 |
+
"learning_rate": 2.7606666666666664e-07,
|
| 1078 |
+
"loss": 2.5411,
|
| 1079 |
+
"step": 784000
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 3.21,
|
| 1083 |
+
"eval_loss": 2.4726006984710693,
|
| 1084 |
+
"eval_runtime": 224.2901,
|
| 1085 |
+
"eval_samples_per_second": 916.269,
|
| 1086 |
+
"eval_steps_per_second": 57.27,
|
| 1087 |
+
"step": 784000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 3.25,
|
| 1091 |
+
"eval_loss": 2.476133108139038,
|
| 1092 |
+
"eval_runtime": 224.2831,
|
| 1093 |
+
"eval_samples_per_second": 916.297,
|
| 1094 |
+
"eval_steps_per_second": 57.271,
|
| 1095 |
+
"step": 792000
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 3.28,
|
| 1099 |
+
"learning_rate": 2.733333333333333e-07,
|
| 1100 |
+
"loss": 2.5407,
|
| 1101 |
+
"step": 800000
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 3.28,
|
| 1105 |
+
"eval_loss": 2.4611737728118896,
|
| 1106 |
+
"eval_runtime": 223.8119,
|
| 1107 |
+
"eval_samples_per_second": 918.226,
|
| 1108 |
+
"eval_steps_per_second": 57.392,
|
| 1109 |
+
"step": 800000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 3.31,
|
| 1113 |
+
"eval_loss": 2.4742894172668457,
|
| 1114 |
+
"eval_runtime": 224.8606,
|
| 1115 |
+
"eval_samples_per_second": 913.944,
|
| 1116 |
+
"eval_steps_per_second": 57.124,
|
| 1117 |
+
"step": 808000
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 3.34,
|
| 1121 |
+
"learning_rate": 2.706e-07,
|
| 1122 |
+
"loss": 2.5307,
|
| 1123 |
+
"step": 816000
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 3.34,
|
| 1127 |
+
"eval_loss": 2.469853401184082,
|
| 1128 |
+
"eval_runtime": 224.177,
|
| 1129 |
+
"eval_samples_per_second": 916.731,
|
| 1130 |
+
"eval_steps_per_second": 57.298,
|
| 1131 |
+
"step": 816000
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 3.38,
|
| 1135 |
+
"eval_loss": 2.4721498489379883,
|
| 1136 |
+
"eval_runtime": 223.5611,
|
| 1137 |
+
"eval_samples_per_second": 919.256,
|
| 1138 |
+
"eval_steps_per_second": 57.456,
|
| 1139 |
+
"step": 824000
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 3.41,
|
| 1143 |
+
"learning_rate": 2.6786666666666666e-07,
|
| 1144 |
+
"loss": 2.5391,
|
| 1145 |
+
"step": 832000
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 3.41,
|
| 1149 |
+
"eval_loss": 2.461381435394287,
|
| 1150 |
+
"eval_runtime": 224.3794,
|
| 1151 |
+
"eval_samples_per_second": 915.904,
|
| 1152 |
+
"eval_steps_per_second": 57.247,
|
| 1153 |
+
"step": 832000
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 3.44,
|
| 1157 |
+
"eval_loss": 2.4641225337982178,
|
| 1158 |
+
"eval_runtime": 224.9887,
|
| 1159 |
+
"eval_samples_per_second": 913.423,
|
| 1160 |
+
"eval_steps_per_second": 57.092,
|
| 1161 |
+
"step": 840000
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 3.47,
|
| 1165 |
+
"learning_rate": 2.651333333333333e-07,
|
| 1166 |
+
"loss": 2.5378,
|
| 1167 |
+
"step": 848000
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 3.47,
|
| 1171 |
+
"eval_loss": 2.4652435779571533,
|
| 1172 |
+
"eval_runtime": 225.299,
|
| 1173 |
+
"eval_samples_per_second": 912.166,
|
| 1174 |
+
"eval_steps_per_second": 57.013,
|
| 1175 |
+
"step": 848000
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 3.51,
|
| 1179 |
+
"eval_loss": 2.4640512466430664,
|
| 1180 |
+
"eval_runtime": 224.7841,
|
| 1181 |
+
"eval_samples_per_second": 914.255,
|
| 1182 |
+
"eval_steps_per_second": 57.144,
|
| 1183 |
+
"step": 856000
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 3.54,
|
| 1187 |
+
"learning_rate": 2.624e-07,
|
| 1188 |
+
"loss": 2.5399,
|
| 1189 |
+
"step": 864000
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 3.54,
|
| 1193 |
+
"eval_loss": 2.469067096710205,
|
| 1194 |
+
"eval_runtime": 225.0567,
|
| 1195 |
+
"eval_samples_per_second": 913.148,
|
| 1196 |
+
"eval_steps_per_second": 57.075,
|
| 1197 |
+
"step": 864000
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 3.57,
|
| 1201 |
+
"eval_loss": 2.4611856937408447,
|
| 1202 |
+
"eval_runtime": 224.6227,
|
| 1203 |
+
"eval_samples_per_second": 914.912,
|
| 1204 |
+
"eval_steps_per_second": 57.185,
|
| 1205 |
+
"step": 872000
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 3.61,
|
| 1209 |
+
"learning_rate": 2.596666666666667e-07,
|
| 1210 |
+
"loss": 2.5412,
|
| 1211 |
+
"step": 880000
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 3.61,
|
| 1215 |
+
"eval_loss": 2.469621419906616,
|
| 1216 |
+
"eval_runtime": 225.3239,
|
| 1217 |
+
"eval_samples_per_second": 912.065,
|
| 1218 |
+
"eval_steps_per_second": 57.007,
|
| 1219 |
+
"step": 880000
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 3.64,
|
| 1223 |
+
"eval_loss": 2.4638073444366455,
|
| 1224 |
+
"eval_runtime": 224.7878,
|
| 1225 |
+
"eval_samples_per_second": 914.24,
|
| 1226 |
+
"eval_steps_per_second": 57.143,
|
| 1227 |
+
"step": 888000
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 3.67,
|
| 1231 |
+
"learning_rate": 2.5693333333333333e-07,
|
| 1232 |
+
"loss": 2.5389,
|
| 1233 |
+
"step": 896000
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 3.67,
|
| 1237 |
+
"eval_loss": 2.4658303260803223,
|
| 1238 |
+
"eval_runtime": 224.4987,
|
| 1239 |
+
"eval_samples_per_second": 915.417,
|
| 1240 |
+
"eval_steps_per_second": 57.216,
|
| 1241 |
+
"step": 896000
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 3.7,
|
| 1245 |
+
"eval_loss": 2.4725189208984375,
|
| 1246 |
+
"eval_runtime": 226.711,
|
| 1247 |
+
"eval_samples_per_second": 906.484,
|
| 1248 |
+
"eval_steps_per_second": 56.658,
|
| 1249 |
+
"step": 904000
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 3.74,
|
| 1253 |
+
"learning_rate": 2.542e-07,
|
| 1254 |
+
"loss": 2.5325,
|
| 1255 |
+
"step": 912000
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 3.74,
|
| 1259 |
+
"eval_loss": 2.46415114402771,
|
| 1260 |
+
"eval_runtime": 225.3655,
|
| 1261 |
+
"eval_samples_per_second": 911.896,
|
| 1262 |
+
"eval_steps_per_second": 56.996,
|
| 1263 |
+
"step": 912000
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 3.77,
|
| 1267 |
+
"eval_loss": 2.4599404335021973,
|
| 1268 |
+
"eval_runtime": 224.7087,
|
| 1269 |
+
"eval_samples_per_second": 914.562,
|
| 1270 |
+
"eval_steps_per_second": 57.163,
|
| 1271 |
+
"step": 920000
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 3.8,
|
| 1275 |
+
"learning_rate": 2.5146666666666664e-07,
|
| 1276 |
+
"loss": 2.5351,
|
| 1277 |
+
"step": 928000
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 3.8,
|
| 1281 |
+
"eval_loss": 2.4616599082946777,
|
| 1282 |
+
"eval_runtime": 226.6966,
|
| 1283 |
+
"eval_samples_per_second": 906.542,
|
| 1284 |
+
"eval_steps_per_second": 56.662,
|
| 1285 |
+
"step": 928000
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 3.84,
|
| 1289 |
+
"eval_loss": 2.464627265930176,
|
| 1290 |
+
"eval_runtime": 224.9933,
|
| 1291 |
+
"eval_samples_per_second": 913.405,
|
| 1292 |
+
"eval_steps_per_second": 57.091,
|
| 1293 |
+
"step": 936000
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 3.87,
|
| 1297 |
+
"learning_rate": 2.4873333333333335e-07,
|
| 1298 |
+
"loss": 2.522,
|
| 1299 |
+
"step": 944000
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 3.87,
|
| 1303 |
+
"eval_loss": 2.4665021896362305,
|
| 1304 |
+
"eval_runtime": 225.0978,
|
| 1305 |
+
"eval_samples_per_second": 912.981,
|
| 1306 |
+
"eval_steps_per_second": 57.064,
|
| 1307 |
+
"step": 944000
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 3.9,
|
| 1311 |
+
"eval_loss": 2.4761972427368164,
|
| 1312 |
+
"eval_runtime": 224.2641,
|
| 1313 |
+
"eval_samples_per_second": 916.375,
|
| 1314 |
+
"eval_steps_per_second": 57.276,
|
| 1315 |
+
"step": 952000
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 3.93,
|
| 1319 |
+
"learning_rate": 2.46e-07,
|
| 1320 |
+
"loss": 2.5331,
|
| 1321 |
+
"step": 960000
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 3.93,
|
| 1325 |
+
"eval_loss": 2.4668779373168945,
|
| 1326 |
+
"eval_runtime": 225.069,
|
| 1327 |
+
"eval_samples_per_second": 913.098,
|
| 1328 |
+
"eval_steps_per_second": 57.071,
|
| 1329 |
+
"step": 960000
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 3.97,
|
| 1333 |
+
"eval_loss": 2.4549825191497803,
|
| 1334 |
+
"eval_runtime": 224.677,
|
| 1335 |
+
"eval_samples_per_second": 914.691,
|
| 1336 |
+
"eval_steps_per_second": 57.171,
|
| 1337 |
+
"step": 968000
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 4.0,
|
| 1341 |
+
"learning_rate": 2.4326666666666666e-07,
|
| 1342 |
+
"loss": 2.5276,
|
| 1343 |
+
"step": 976000
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 4.0,
|
| 1347 |
+
"eval_loss": 2.466226577758789,
|
| 1348 |
+
"eval_runtime": 224.6009,
|
| 1349 |
+
"eval_samples_per_second": 915.001,
|
| 1350 |
+
"eval_steps_per_second": 57.19,
|
| 1351 |
+
"step": 976000
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 4.03,
|
| 1355 |
+
"eval_loss": 2.464536190032959,
|
| 1356 |
+
"eval_runtime": 224.76,
|
| 1357 |
+
"eval_samples_per_second": 914.353,
|
| 1358 |
+
"eval_steps_per_second": 57.15,
|
| 1359 |
+
"step": 984000
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 4.06,
|
| 1363 |
+
"learning_rate": 2.405333333333333e-07,
|
| 1364 |
+
"loss": 2.5206,
|
| 1365 |
+
"step": 992000
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 4.06,
|
| 1369 |
+
"eval_loss": 2.4586591720581055,
|
| 1370 |
+
"eval_runtime": 225.6548,
|
| 1371 |
+
"eval_samples_per_second": 910.727,
|
| 1372 |
+
"eval_steps_per_second": 56.923,
|
| 1373 |
+
"step": 992000
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 4.1,
|
| 1377 |
+
"eval_loss": 2.47253680229187,
|
| 1378 |
+
"eval_runtime": 225.9081,
|
| 1379 |
+
"eval_samples_per_second": 909.706,
|
| 1380 |
+
"eval_steps_per_second": 56.859,
|
| 1381 |
+
"step": 1000000
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 4.13,
|
| 1385 |
+
"learning_rate": 2.3779999999999997e-07,
|
| 1386 |
+
"loss": 2.5294,
|
| 1387 |
+
"step": 1008000
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 4.13,
|
| 1391 |
+
"eval_loss": 2.458824634552002,
|
| 1392 |
+
"eval_runtime": 224.8489,
|
| 1393 |
+
"eval_samples_per_second": 913.991,
|
| 1394 |
+
"eval_steps_per_second": 57.127,
|
| 1395 |
+
"step": 1008000
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 4.16,
|
| 1399 |
+
"eval_loss": 2.4591076374053955,
|
| 1400 |
+
"eval_runtime": 225.8271,
|
| 1401 |
+
"eval_samples_per_second": 910.032,
|
| 1402 |
+
"eval_steps_per_second": 56.88,
|
| 1403 |
+
"step": 1016000
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 4.2,
|
| 1407 |
+
"learning_rate": 2.3506666666666668e-07,
|
| 1408 |
+
"loss": 2.5312,
|
| 1409 |
+
"step": 1024000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 4.2,
|
| 1413 |
+
"eval_loss": 2.4680891036987305,
|
| 1414 |
+
"eval_runtime": 226.319,
|
| 1415 |
+
"eval_samples_per_second": 908.055,
|
| 1416 |
+
"eval_steps_per_second": 56.756,
|
| 1417 |
+
"step": 1024000
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 4.23,
|
| 1421 |
+
"eval_loss": 2.4624712467193604,
|
| 1422 |
+
"eval_runtime": 226.6472,
|
| 1423 |
+
"eval_samples_per_second": 906.74,
|
| 1424 |
+
"eval_steps_per_second": 56.674,
|
| 1425 |
+
"step": 1032000
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 4.26,
|
| 1429 |
+
"learning_rate": 2.3233333333333334e-07,
|
| 1430 |
+
"loss": 2.525,
|
| 1431 |
+
"step": 1040000
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 4.26,
|
| 1435 |
+
"eval_loss": 2.4659371376037598,
|
| 1436 |
+
"eval_runtime": 225.5287,
|
| 1437 |
+
"eval_samples_per_second": 911.237,
|
| 1438 |
+
"eval_steps_per_second": 56.955,
|
| 1439 |
+
"step": 1040000
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 4.29,
|
| 1443 |
+
"eval_loss": 2.460909843444824,
|
| 1444 |
+
"eval_runtime": 225.5374,
|
| 1445 |
+
"eval_samples_per_second": 911.201,
|
| 1446 |
+
"eval_steps_per_second": 56.953,
|
| 1447 |
+
"step": 1048000
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 4.33,
|
| 1451 |
+
"learning_rate": 2.2960000000000002e-07,
|
| 1452 |
+
"loss": 2.5318,
|
| 1453 |
+
"step": 1056000
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 4.33,
|
| 1457 |
+
"eval_loss": 2.4571011066436768,
|
| 1458 |
+
"eval_runtime": 225.1138,
|
| 1459 |
+
"eval_samples_per_second": 912.916,
|
| 1460 |
+
"eval_steps_per_second": 57.06,
|
| 1461 |
+
"step": 1056000
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 4.36,
|
| 1465 |
+
"eval_loss": 2.4581968784332275,
|
| 1466 |
+
"eval_runtime": 226.7154,
|
| 1467 |
+
"eval_samples_per_second": 906.467,
|
| 1468 |
+
"eval_steps_per_second": 56.657,
|
| 1469 |
+
"step": 1064000
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 4.39,
|
| 1473 |
+
"learning_rate": 2.2686666666666667e-07,
|
| 1474 |
+
"loss": 2.5332,
|
| 1475 |
+
"step": 1072000
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 4.39,
|
| 1479 |
+
"eval_loss": 2.456618547439575,
|
| 1480 |
+
"eval_runtime": 225.422,
|
| 1481 |
+
"eval_samples_per_second": 911.668,
|
| 1482 |
+
"eval_steps_per_second": 56.982,
|
| 1483 |
+
"step": 1072000
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 4.43,
|
| 1487 |
+
"eval_loss": 2.4587738513946533,
|
| 1488 |
+
"eval_runtime": 226.533,
|
| 1489 |
+
"eval_samples_per_second": 907.197,
|
| 1490 |
+
"eval_steps_per_second": 56.703,
|
| 1491 |
+
"step": 1080000
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 4.46,
|
| 1495 |
+
"learning_rate": 2.2413333333333333e-07,
|
| 1496 |
+
"loss": 2.5168,
|
| 1497 |
+
"step": 1088000
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 4.46,
|
| 1501 |
+
"eval_loss": 2.4606146812438965,
|
| 1502 |
+
"eval_runtime": 226.9924,
|
| 1503 |
+
"eval_samples_per_second": 905.361,
|
| 1504 |
+
"eval_steps_per_second": 56.588,
|
| 1505 |
+
"step": 1088000
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 4.49,
|
| 1509 |
+
"eval_loss": 2.4597506523132324,
|
| 1510 |
+
"eval_runtime": 228.0241,
|
| 1511 |
+
"eval_samples_per_second": 901.264,
|
| 1512 |
+
"eval_steps_per_second": 56.332,
|
| 1513 |
+
"step": 1096000
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 4.52,
|
| 1517 |
+
"learning_rate": 2.214e-07,
|
| 1518 |
+
"loss": 2.5181,
|
| 1519 |
+
"step": 1104000
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 4.52,
|
| 1523 |
+
"eval_loss": 2.454252004623413,
|
| 1524 |
+
"eval_runtime": 228.6537,
|
| 1525 |
+
"eval_samples_per_second": 898.783,
|
| 1526 |
+
"eval_steps_per_second": 56.177,
|
| 1527 |
+
"step": 1104000
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 4.56,
|
| 1531 |
+
"eval_loss": 2.4619953632354736,
|
| 1532 |
+
"eval_runtime": 226.293,
|
| 1533 |
+
"eval_samples_per_second": 908.159,
|
| 1534 |
+
"eval_steps_per_second": 56.763,
|
| 1535 |
+
"step": 1112000
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 4.59,
|
| 1539 |
+
"learning_rate": 2.1866666666666667e-07,
|
| 1540 |
+
"loss": 2.5246,
|
| 1541 |
+
"step": 1120000
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 4.59,
|
| 1545 |
+
"eval_loss": 2.4638657569885254,
|
| 1546 |
+
"eval_runtime": 228.163,
|
| 1547 |
+
"eval_samples_per_second": 900.716,
|
| 1548 |
+
"eval_steps_per_second": 56.297,
|
| 1549 |
+
"step": 1120000
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 4.62,
|
| 1553 |
+
"eval_loss": 2.4556171894073486,
|
| 1554 |
+
"eval_runtime": 228.3656,
|
| 1555 |
+
"eval_samples_per_second": 899.917,
|
| 1556 |
+
"eval_steps_per_second": 56.248,
|
| 1557 |
+
"step": 1128000
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 4.65,
|
| 1561 |
+
"learning_rate": 2.1593333333333332e-07,
|
| 1562 |
+
"loss": 2.5318,
|
| 1563 |
+
"step": 1136000
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 4.65,
|
| 1567 |
+
"eval_loss": 2.457075595855713,
|
| 1568 |
+
"eval_runtime": 227.349,
|
| 1569 |
+
"eval_samples_per_second": 903.941,
|
| 1570 |
+
"eval_steps_per_second": 56.499,
|
| 1571 |
+
"step": 1136000
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 4.69,
|
| 1575 |
+
"eval_loss": 2.4636013507843018,
|
| 1576 |
+
"eval_runtime": 228.1517,
|
| 1577 |
+
"eval_samples_per_second": 900.76,
|
| 1578 |
+
"eval_steps_per_second": 56.3,
|
| 1579 |
+
"step": 1144000
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 4.72,
|
| 1583 |
+
"learning_rate": 2.132e-07,
|
| 1584 |
+
"loss": 2.512,
|
| 1585 |
+
"step": 1152000
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 4.72,
|
| 1589 |
+
"eval_loss": 2.4567556381225586,
|
| 1590 |
+
"eval_runtime": 228.1066,
|
| 1591 |
+
"eval_samples_per_second": 900.938,
|
| 1592 |
+
"eval_steps_per_second": 56.311,
|
| 1593 |
+
"step": 1152000
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 4.75,
|
| 1597 |
+
"eval_loss": 2.4644010066986084,
|
| 1598 |
+
"eval_runtime": 228.916,
|
| 1599 |
+
"eval_samples_per_second": 897.753,
|
| 1600 |
+
"eval_steps_per_second": 56.112,
|
| 1601 |
+
"step": 1160000
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 4.79,
|
| 1605 |
+
"learning_rate": 2.1046666666666666e-07,
|
| 1606 |
+
"loss": 2.5174,
|
| 1607 |
+
"step": 1168000
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 4.79,
|
| 1611 |
+
"eval_loss": 2.4528720378875732,
|
| 1612 |
+
"eval_runtime": 228.2634,
|
| 1613 |
+
"eval_samples_per_second": 900.32,
|
| 1614 |
+
"eval_steps_per_second": 56.273,
|
| 1615 |
+
"step": 1168000
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 4.82,
|
| 1619 |
+
"eval_loss": 2.4613921642303467,
|
| 1620 |
+
"eval_runtime": 228.3765,
|
| 1621 |
+
"eval_samples_per_second": 899.874,
|
| 1622 |
+
"eval_steps_per_second": 56.245,
|
| 1623 |
+
"step": 1176000
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 4.85,
|
| 1627 |
+
"learning_rate": 2.0773333333333334e-07,
|
| 1628 |
+
"loss": 2.5196,
|
| 1629 |
+
"step": 1184000
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 4.85,
|
| 1633 |
+
"eval_loss": 2.463758707046509,
|
| 1634 |
+
"eval_runtime": 227.0091,
|
| 1635 |
+
"eval_samples_per_second": 905.294,
|
| 1636 |
+
"eval_steps_per_second": 56.584,
|
| 1637 |
+
"step": 1184000
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 4.88,
|
| 1641 |
+
"eval_loss": 2.453406572341919,
|
| 1642 |
+
"eval_runtime": 227.3538,
|
| 1643 |
+
"eval_samples_per_second": 903.921,
|
| 1644 |
+
"eval_steps_per_second": 56.498,
|
| 1645 |
+
"step": 1192000
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 4.92,
|
| 1649 |
+
"learning_rate": 2.05e-07,
|
| 1650 |
+
"loss": 2.5248,
|
| 1651 |
+
"step": 1200000
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 4.92,
|
| 1655 |
+
"eval_loss": 2.4553115367889404,
|
| 1656 |
+
"eval_runtime": 227.0142,
|
| 1657 |
+
"eval_samples_per_second": 905.274,
|
| 1658 |
+
"eval_steps_per_second": 56.582,
|
| 1659 |
+
"step": 1200000
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 4.95,
|
| 1663 |
+
"eval_loss": 2.453683853149414,
|
| 1664 |
+
"eval_runtime": 226.8684,
|
| 1665 |
+
"eval_samples_per_second": 905.855,
|
| 1666 |
+
"eval_steps_per_second": 56.619,
|
| 1667 |
+
"step": 1208000
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 4.98,
|
| 1671 |
+
"learning_rate": 2.0226666666666668e-07,
|
| 1672 |
+
"loss": 2.5201,
|
| 1673 |
+
"step": 1216000
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 4.98,
|
| 1677 |
+
"eval_loss": 2.4578709602355957,
|
| 1678 |
+
"eval_runtime": 226.9695,
|
| 1679 |
+
"eval_samples_per_second": 905.452,
|
| 1680 |
+
"eval_steps_per_second": 56.594,
|
| 1681 |
+
"step": 1216000
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 5.02,
|
| 1685 |
+
"eval_loss": 2.4524765014648438,
|
| 1686 |
+
"eval_runtime": 226.5657,
|
| 1687 |
+
"eval_samples_per_second": 907.066,
|
| 1688 |
+
"eval_steps_per_second": 56.694,
|
| 1689 |
+
"step": 1224000
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 5.05,
|
| 1693 |
+
"learning_rate": 1.9953333333333333e-07,
|
| 1694 |
+
"loss": 2.5164,
|
| 1695 |
+
"step": 1232000
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 5.05,
|
| 1699 |
+
"eval_loss": 2.4645235538482666,
|
| 1700 |
+
"eval_runtime": 227.3689,
|
| 1701 |
+
"eval_samples_per_second": 903.861,
|
| 1702 |
+
"eval_steps_per_second": 56.494,
|
| 1703 |
+
"step": 1232000
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 5.08,
|
| 1707 |
+
"eval_loss": 2.447993040084839,
|
| 1708 |
+
"eval_runtime": 228.8072,
|
| 1709 |
+
"eval_samples_per_second": 898.18,
|
| 1710 |
+
"eval_steps_per_second": 56.139,
|
| 1711 |
+
"step": 1240000
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 5.11,
|
| 1715 |
+
"learning_rate": 1.968e-07,
|
| 1716 |
+
"loss": 2.5186,
|
| 1717 |
+
"step": 1248000
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 5.11,
|
| 1721 |
+
"eval_loss": 2.4605581760406494,
|
| 1722 |
+
"eval_runtime": 229.2086,
|
| 1723 |
+
"eval_samples_per_second": 896.607,
|
| 1724 |
+
"eval_steps_per_second": 56.041,
|
| 1725 |
+
"step": 1248000
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 5.15,
|
| 1729 |
+
"eval_loss": 2.4623043537139893,
|
| 1730 |
+
"eval_runtime": 229.8264,
|
| 1731 |
+
"eval_samples_per_second": 894.197,
|
| 1732 |
+
"eval_steps_per_second": 55.89,
|
| 1733 |
+
"step": 1256000
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 5.18,
|
| 1737 |
+
"learning_rate": 1.9406666666666667e-07,
|
| 1738 |
+
"loss": 2.5123,
|
| 1739 |
+
"step": 1264000
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 5.18,
|
| 1743 |
+
"eval_loss": 2.456583261489868,
|
| 1744 |
+
"eval_runtime": 230.427,
|
| 1745 |
+
"eval_samples_per_second": 891.866,
|
| 1746 |
+
"eval_steps_per_second": 55.744,
|
| 1747 |
+
"step": 1264000
|
| 1748 |
+
},
|
| 1749 |
+
{
|
| 1750 |
+
"epoch": 5.21,
|
| 1751 |
+
"eval_loss": 2.464402437210083,
|
| 1752 |
+
"eval_runtime": 229.1803,
|
| 1753 |
+
"eval_samples_per_second": 896.717,
|
| 1754 |
+
"eval_steps_per_second": 56.048,
|
| 1755 |
+
"step": 1272000
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 5.24,
|
| 1759 |
+
"learning_rate": 1.9133333333333333e-07,
|
| 1760 |
+
"loss": 2.5233,
|
| 1761 |
+
"step": 1280000
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 5.24,
|
| 1765 |
+
"eval_loss": 2.457606792449951,
|
| 1766 |
+
"eval_runtime": 227.9824,
|
| 1767 |
+
"eval_samples_per_second": 901.429,
|
| 1768 |
+
"eval_steps_per_second": 56.342,
|
| 1769 |
+
"step": 1280000
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 5.28,
|
| 1773 |
+
"eval_loss": 2.451943874359131,
|
| 1774 |
+
"eval_runtime": 229.374,
|
| 1775 |
+
"eval_samples_per_second": 895.96,
|
| 1776 |
+
"eval_steps_per_second": 56.0,
|
| 1777 |
+
"step": 1288000
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 5.31,
|
| 1781 |
+
"learning_rate": 1.886e-07,
|
| 1782 |
+
"loss": 2.513,
|
| 1783 |
+
"step": 1296000
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 5.31,
|
| 1787 |
+
"eval_loss": 2.456979513168335,
|
| 1788 |
+
"eval_runtime": 228.051,
|
| 1789 |
+
"eval_samples_per_second": 901.158,
|
| 1790 |
+
"eval_steps_per_second": 56.325,
|
| 1791 |
+
"step": 1296000
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 5.34,
|
| 1795 |
+
"eval_loss": 2.462719202041626,
|
| 1796 |
+
"eval_runtime": 228.489,
|
| 1797 |
+
"eval_samples_per_second": 899.43,
|
| 1798 |
+
"eval_steps_per_second": 56.217,
|
| 1799 |
+
"step": 1304000
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 5.38,
|
| 1803 |
+
"learning_rate": 1.8586666666666666e-07,
|
| 1804 |
+
"loss": 2.5226,
|
| 1805 |
+
"step": 1312000
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 5.38,
|
| 1809 |
+
"eval_loss": 2.449977397918701,
|
| 1810 |
+
"eval_runtime": 227.9952,
|
| 1811 |
+
"eval_samples_per_second": 901.379,
|
| 1812 |
+
"eval_steps_per_second": 56.339,
|
| 1813 |
+
"step": 1312000
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 5.41,
|
| 1817 |
+
"eval_loss": 2.4563188552856445,
|
| 1818 |
+
"eval_runtime": 227.2759,
|
| 1819 |
+
"eval_samples_per_second": 904.231,
|
| 1820 |
+
"eval_steps_per_second": 56.517,
|
| 1821 |
+
"step": 1320000
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 5.44,
|
| 1825 |
+
"learning_rate": 1.8313333333333332e-07,
|
| 1826 |
+
"loss": 2.5222,
|
| 1827 |
+
"step": 1328000
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 5.44,
|
| 1831 |
+
"eval_loss": 2.4521265029907227,
|
| 1832 |
+
"eval_runtime": 226.8418,
|
| 1833 |
+
"eval_samples_per_second": 905.962,
|
| 1834 |
+
"eval_steps_per_second": 56.625,
|
| 1835 |
+
"step": 1328000
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 5.47,
|
| 1839 |
+
"eval_loss": 2.4591453075408936,
|
| 1840 |
+
"eval_runtime": 226.8374,
|
| 1841 |
+
"eval_samples_per_second": 905.98,
|
| 1842 |
+
"eval_steps_per_second": 56.626,
|
| 1843 |
+
"step": 1336000
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 5.51,
|
| 1847 |
+
"learning_rate": 1.804e-07,
|
| 1848 |
+
"loss": 2.5191,
|
| 1849 |
+
"step": 1344000
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 5.51,
|
| 1853 |
+
"eval_loss": 2.4508602619171143,
|
| 1854 |
+
"eval_runtime": 228.6931,
|
| 1855 |
+
"eval_samples_per_second": 898.628,
|
| 1856 |
+
"eval_steps_per_second": 56.167,
|
| 1857 |
+
"step": 1344000
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 5.54,
|
| 1861 |
+
"eval_loss": 2.455850124359131,
|
| 1862 |
+
"eval_runtime": 228.2295,
|
| 1863 |
+
"eval_samples_per_second": 900.453,
|
| 1864 |
+
"eval_steps_per_second": 56.281,
|
| 1865 |
+
"step": 1352000
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 5.57,
|
| 1869 |
+
"learning_rate": 1.7766666666666666e-07,
|
| 1870 |
+
"loss": 2.5243,
|
| 1871 |
+
"step": 1360000
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 5.57,
|
| 1875 |
+
"eval_loss": 2.4501898288726807,
|
| 1876 |
+
"eval_runtime": 228.7596,
|
| 1877 |
+
"eval_samples_per_second": 898.367,
|
| 1878 |
+
"eval_steps_per_second": 56.151,
|
| 1879 |
+
"step": 1360000
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 5.61,
|
| 1883 |
+
"eval_loss": 2.4514639377593994,
|
| 1884 |
+
"eval_runtime": 227.2719,
|
| 1885 |
+
"eval_samples_per_second": 904.247,
|
| 1886 |
+
"eval_steps_per_second": 56.518,
|
| 1887 |
+
"step": 1368000
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 5.64,
|
| 1891 |
+
"learning_rate": 1.7493333333333334e-07,
|
| 1892 |
+
"loss": 2.5157,
|
| 1893 |
+
"step": 1376000
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 5.64,
|
| 1897 |
+
"eval_loss": 2.4562854766845703,
|
| 1898 |
+
"eval_runtime": 227.9532,
|
| 1899 |
+
"eval_samples_per_second": 901.545,
|
| 1900 |
+
"eval_steps_per_second": 56.349,
|
| 1901 |
+
"step": 1376000
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 5.67,
|
| 1905 |
+
"eval_loss": 2.452606678009033,
|
| 1906 |
+
"eval_runtime": 227.4532,
|
| 1907 |
+
"eval_samples_per_second": 903.527,
|
| 1908 |
+
"eval_steps_per_second": 56.473,
|
| 1909 |
+
"step": 1384000
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 5.7,
|
| 1913 |
+
"learning_rate": 1.722e-07,
|
| 1914 |
+
"loss": 2.5162,
|
| 1915 |
+
"step": 1392000
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 5.7,
|
| 1919 |
+
"eval_loss": 2.458620071411133,
|
| 1920 |
+
"eval_runtime": 228.2374,
|
| 1921 |
+
"eval_samples_per_second": 900.422,
|
| 1922 |
+
"eval_steps_per_second": 56.279,
|
| 1923 |
+
"step": 1392000
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 5.74,
|
| 1927 |
+
"eval_loss": 2.458387613296509,
|
| 1928 |
+
"eval_runtime": 228.0105,
|
| 1929 |
+
"eval_samples_per_second": 901.318,
|
| 1930 |
+
"eval_steps_per_second": 56.335,
|
| 1931 |
+
"step": 1400000
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 5.77,
|
| 1935 |
+
"learning_rate": 1.6946666666666668e-07,
|
| 1936 |
+
"loss": 2.5169,
|
| 1937 |
+
"step": 1408000
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 5.77,
|
| 1941 |
+
"eval_loss": 2.454158067703247,
|
| 1942 |
+
"eval_runtime": 227.4312,
|
| 1943 |
+
"eval_samples_per_second": 903.614,
|
| 1944 |
+
"eval_steps_per_second": 56.479,
|
| 1945 |
+
"step": 1408000
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 5.8,
|
| 1949 |
+
"eval_loss": 2.460242986679077,
|
| 1950 |
+
"eval_runtime": 228.5958,
|
| 1951 |
+
"eval_samples_per_second": 899.01,
|
| 1952 |
+
"eval_steps_per_second": 56.191,
|
| 1953 |
+
"step": 1416000
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 5.84,
|
| 1957 |
+
"learning_rate": 1.6673333333333333e-07,
|
| 1958 |
+
"loss": 2.5127,
|
| 1959 |
+
"step": 1424000
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 5.84,
|
| 1963 |
+
"eval_loss": 2.458707809448242,
|
| 1964 |
+
"eval_runtime": 228.0452,
|
| 1965 |
+
"eval_samples_per_second": 901.181,
|
| 1966 |
+
"eval_steps_per_second": 56.327,
|
| 1967 |
+
"step": 1424000
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 5.87,
|
| 1971 |
+
"eval_loss": 2.452913284301758,
|
| 1972 |
+
"eval_runtime": 227.4908,
|
| 1973 |
+
"eval_samples_per_second": 903.377,
|
| 1974 |
+
"eval_steps_per_second": 56.464,
|
| 1975 |
+
"step": 1432000
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 5.9,
|
| 1979 |
+
"learning_rate": 1.64e-07,
|
| 1980 |
+
"loss": 2.5144,
|
| 1981 |
+
"step": 1440000
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 5.9,
|
| 1985 |
+
"eval_loss": 2.462021827697754,
|
| 1986 |
+
"eval_runtime": 229.4885,
|
| 1987 |
+
"eval_samples_per_second": 895.513,
|
| 1988 |
+
"eval_steps_per_second": 55.972,
|
| 1989 |
+
"step": 1440000
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 5.93,
|
| 1993 |
+
"eval_loss": 2.450927972793579,
|
| 1994 |
+
"eval_runtime": 227.9748,
|
| 1995 |
+
"eval_samples_per_second": 901.459,
|
| 1996 |
+
"eval_steps_per_second": 56.344,
|
| 1997 |
+
"step": 1448000
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 5.97,
|
| 2001 |
+
"learning_rate": 1.6126666666666667e-07,
|
| 2002 |
+
"loss": 2.5175,
|
| 2003 |
+
"step": 1456000
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 5.97,
|
| 2007 |
+
"eval_loss": 2.4503204822540283,
|
| 2008 |
+
"eval_runtime": 227.5178,
|
| 2009 |
+
"eval_samples_per_second": 903.27,
|
| 2010 |
+
"eval_steps_per_second": 56.457,
|
| 2011 |
+
"step": 1456000
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 6.0,
|
| 2015 |
+
"eval_loss": 2.4545462131500244,
|
| 2016 |
+
"eval_runtime": 227.7963,
|
| 2017 |
+
"eval_samples_per_second": 902.165,
|
| 2018 |
+
"eval_steps_per_second": 56.388,
|
| 2019 |
+
"step": 1464000
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 6.03,
|
| 2023 |
+
"learning_rate": 1.5853333333333332e-07,
|
| 2024 |
+
"loss": 2.5147,
|
| 2025 |
+
"step": 1472000
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 6.03,
|
| 2029 |
+
"eval_loss": 2.4440090656280518,
|
| 2030 |
+
"eval_runtime": 227.8162,
|
| 2031 |
+
"eval_samples_per_second": 902.087,
|
| 2032 |
+
"eval_steps_per_second": 56.383,
|
| 2033 |
+
"step": 1472000
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 6.06,
|
| 2037 |
+
"eval_loss": 2.457670211791992,
|
| 2038 |
+
"eval_runtime": 228.5245,
|
| 2039 |
+
"eval_samples_per_second": 899.291,
|
| 2040 |
+
"eval_steps_per_second": 56.208,
|
| 2041 |
+
"step": 1480000
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 6.1,
|
| 2045 |
+
"learning_rate": 1.558e-07,
|
| 2046 |
+
"loss": 2.5128,
|
| 2047 |
+
"step": 1488000
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 6.1,
|
| 2051 |
+
"eval_loss": 2.456602096557617,
|
| 2052 |
+
"eval_runtime": 230.1502,
|
| 2053 |
+
"eval_samples_per_second": 892.939,
|
| 2054 |
+
"eval_steps_per_second": 55.811,
|
| 2055 |
+
"step": 1488000
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 6.13,
|
| 2059 |
+
"eval_loss": 2.449889659881592,
|
| 2060 |
+
"eval_runtime": 228.3041,
|
| 2061 |
+
"eval_samples_per_second": 900.159,
|
| 2062 |
+
"eval_steps_per_second": 56.263,
|
| 2063 |
+
"step": 1496000
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 6.16,
|
| 2067 |
+
"learning_rate": 1.5306666666666666e-07,
|
| 2068 |
+
"loss": 2.5168,
|
| 2069 |
+
"step": 1504000
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 6.16,
|
| 2073 |
+
"eval_loss": 2.4480044841766357,
|
| 2074 |
+
"eval_runtime": 228.2508,
|
| 2075 |
+
"eval_samples_per_second": 900.369,
|
| 2076 |
+
"eval_steps_per_second": 56.276,
|
| 2077 |
+
"step": 1504000
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 6.2,
|
| 2081 |
+
"eval_loss": 2.4436299800872803,
|
| 2082 |
+
"eval_runtime": 229.3638,
|
| 2083 |
+
"eval_samples_per_second": 896.0,
|
| 2084 |
+
"eval_steps_per_second": 56.003,
|
| 2085 |
+
"step": 1512000
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 6.23,
|
| 2089 |
+
"learning_rate": 1.5033333333333332e-07,
|
| 2090 |
+
"loss": 2.5225,
|
| 2091 |
+
"step": 1520000
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 6.23,
|
| 2095 |
+
"eval_loss": 2.446739912033081,
|
| 2096 |
+
"eval_runtime": 228.4899,
|
| 2097 |
+
"eval_samples_per_second": 899.427,
|
| 2098 |
+
"eval_steps_per_second": 56.217,
|
| 2099 |
+
"step": 1520000
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 6.26,
|
| 2103 |
+
"eval_loss": 2.4519920349121094,
|
| 2104 |
+
"eval_runtime": 228.2075,
|
| 2105 |
+
"eval_samples_per_second": 900.54,
|
| 2106 |
+
"eval_steps_per_second": 56.286,
|
| 2107 |
+
"step": 1528000
|
| 2108 |
+
},
|
| 2109 |
+
{
|
| 2110 |
+
"epoch": 6.29,
|
| 2111 |
+
"learning_rate": 1.476e-07,
|
| 2112 |
+
"loss": 2.5135,
|
| 2113 |
+
"step": 1536000
|
| 2114 |
+
},
|
| 2115 |
+
{
|
| 2116 |
+
"epoch": 6.29,
|
| 2117 |
+
"eval_loss": 2.4535210132598877,
|
| 2118 |
+
"eval_runtime": 228.7342,
|
| 2119 |
+
"eval_samples_per_second": 898.466,
|
| 2120 |
+
"eval_steps_per_second": 56.157,
|
| 2121 |
+
"step": 1536000
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 6.33,
|
| 2125 |
+
"eval_loss": 2.4462831020355225,
|
| 2126 |
+
"eval_runtime": 229.9473,
|
| 2127 |
+
"eval_samples_per_second": 893.727,
|
| 2128 |
+
"eval_steps_per_second": 55.861,
|
| 2129 |
+
"step": 1544000
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 6.36,
|
| 2133 |
+
"learning_rate": 1.4486666666666665e-07,
|
| 2134 |
+
"loss": 2.5161,
|
| 2135 |
+
"step": 1552000
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 6.36,
|
| 2139 |
+
"eval_loss": 2.4556400775909424,
|
| 2140 |
+
"eval_runtime": 228.5872,
|
| 2141 |
+
"eval_samples_per_second": 899.044,
|
| 2142 |
+
"eval_steps_per_second": 56.193,
|
| 2143 |
+
"step": 1552000
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 6.39,
|
| 2147 |
+
"eval_loss": 2.4604580402374268,
|
| 2148 |
+
"eval_runtime": 229.1233,
|
| 2149 |
+
"eval_samples_per_second": 896.941,
|
| 2150 |
+
"eval_steps_per_second": 56.062,
|
| 2151 |
+
"step": 1560000
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 6.43,
|
| 2155 |
+
"learning_rate": 1.4213333333333334e-07,
|
| 2156 |
+
"loss": 2.5144,
|
| 2157 |
+
"step": 1568000
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 6.43,
|
| 2161 |
+
"eval_loss": 2.4516451358795166,
|
| 2162 |
+
"eval_runtime": 229.9726,
|
| 2163 |
+
"eval_samples_per_second": 893.628,
|
| 2164 |
+
"eval_steps_per_second": 55.854,
|
| 2165 |
+
"step": 1568000
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 6.46,
|
| 2169 |
+
"eval_loss": 2.4487648010253906,
|
| 2170 |
+
"eval_runtime": 229.4253,
|
| 2171 |
+
"eval_samples_per_second": 895.76,
|
| 2172 |
+
"eval_steps_per_second": 55.988,
|
| 2173 |
+
"step": 1576000
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 6.49,
|
| 2177 |
+
"learning_rate": 1.3940000000000002e-07,
|
| 2178 |
+
"loss": 2.5209,
|
| 2179 |
+
"step": 1584000
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 6.49,
|
| 2183 |
+
"eval_loss": 2.4525067806243896,
|
| 2184 |
+
"eval_runtime": 228.8527,
|
| 2185 |
+
"eval_samples_per_second": 898.001,
|
| 2186 |
+
"eval_steps_per_second": 56.128,
|
| 2187 |
+
"step": 1584000
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 6.52,
|
| 2191 |
+
"eval_loss": 2.450185537338257,
|
| 2192 |
+
"eval_runtime": 230.8087,
|
| 2193 |
+
"eval_samples_per_second": 890.391,
|
| 2194 |
+
"eval_steps_per_second": 55.652,
|
| 2195 |
+
"step": 1592000
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 6.56,
|
| 2199 |
+
"learning_rate": 1.3666666666666665e-07,
|
| 2200 |
+
"loss": 2.5102,
|
| 2201 |
+
"step": 1600000
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 6.56,
|
| 2205 |
+
"eval_loss": 2.453780174255371,
|
| 2206 |
+
"eval_runtime": 229.4733,
|
| 2207 |
+
"eval_samples_per_second": 895.573,
|
| 2208 |
+
"eval_steps_per_second": 55.976,
|
| 2209 |
+
"step": 1600000
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 6.59,
|
| 2213 |
+
"eval_loss": 2.4490787982940674,
|
| 2214 |
+
"eval_runtime": 229.059,
|
| 2215 |
+
"eval_samples_per_second": 897.192,
|
| 2216 |
+
"eval_steps_per_second": 56.077,
|
| 2217 |
+
"step": 1608000
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 6.62,
|
| 2221 |
+
"learning_rate": 1.3393333333333333e-07,
|
| 2222 |
+
"loss": 2.5176,
|
| 2223 |
+
"step": 1616000
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 6.62,
|
| 2227 |
+
"eval_loss": 2.452752113342285,
|
| 2228 |
+
"eval_runtime": 228.4962,
|
| 2229 |
+
"eval_samples_per_second": 899.402,
|
| 2230 |
+
"eval_steps_per_second": 56.215,
|
| 2231 |
+
"step": 1616000
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 6.65,
|
| 2235 |
+
"eval_loss": 2.44599986076355,
|
| 2236 |
+
"eval_runtime": 228.5114,
|
| 2237 |
+
"eval_samples_per_second": 899.342,
|
| 2238 |
+
"eval_steps_per_second": 56.212,
|
| 2239 |
+
"step": 1624000
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 6.69,
|
| 2243 |
+
"learning_rate": 1.312e-07,
|
| 2244 |
+
"loss": 2.5208,
|
| 2245 |
+
"step": 1632000
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 6.69,
|
| 2249 |
+
"eval_loss": 2.4484992027282715,
|
| 2250 |
+
"eval_runtime": 230.1605,
|
| 2251 |
+
"eval_samples_per_second": 892.899,
|
| 2252 |
+
"eval_steps_per_second": 55.809,
|
| 2253 |
+
"step": 1632000
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 6.72,
|
| 2257 |
+
"eval_loss": 2.451284646987915,
|
| 2258 |
+
"eval_runtime": 229.1401,
|
| 2259 |
+
"eval_samples_per_second": 896.875,
|
| 2260 |
+
"eval_steps_per_second": 56.057,
|
| 2261 |
+
"step": 1640000
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 6.75,
|
| 2265 |
+
"learning_rate": 1.2846666666666667e-07,
|
| 2266 |
+
"loss": 2.5064,
|
| 2267 |
+
"step": 1648000
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 6.75,
|
| 2271 |
+
"eval_loss": 2.451927900314331,
|
| 2272 |
+
"eval_runtime": 229.3071,
|
| 2273 |
+
"eval_samples_per_second": 896.222,
|
| 2274 |
+
"eval_steps_per_second": 56.017,
|
| 2275 |
+
"step": 1648000
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 6.79,
|
| 2279 |
+
"eval_loss": 2.449305295944214,
|
| 2280 |
+
"eval_runtime": 231.2204,
|
| 2281 |
+
"eval_samples_per_second": 888.806,
|
| 2282 |
+
"eval_steps_per_second": 55.553,
|
| 2283 |
+
"step": 1656000
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 6.82,
|
| 2287 |
+
"learning_rate": 1.2573333333333332e-07,
|
| 2288 |
+
"loss": 2.5111,
|
| 2289 |
+
"step": 1664000
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 6.82,
|
| 2293 |
+
"eval_loss": 2.4505178928375244,
|
| 2294 |
+
"eval_runtime": 230.2462,
|
| 2295 |
+
"eval_samples_per_second": 892.566,
|
| 2296 |
+
"eval_steps_per_second": 55.788,
|
| 2297 |
+
"step": 1664000
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 6.85,
|
| 2301 |
+
"eval_loss": 2.4501988887786865,
|
| 2302 |
+
"eval_runtime": 229.973,
|
| 2303 |
+
"eval_samples_per_second": 893.627,
|
| 2304 |
+
"eval_steps_per_second": 55.854,
|
| 2305 |
+
"step": 1672000
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 6.88,
|
| 2309 |
+
"learning_rate": 1.23e-07,
|
| 2310 |
+
"loss": 2.5141,
|
| 2311 |
+
"step": 1680000
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 6.88,
|
| 2315 |
+
"eval_loss": 2.4560253620147705,
|
| 2316 |
+
"eval_runtime": 229.6465,
|
| 2317 |
+
"eval_samples_per_second": 894.897,
|
| 2318 |
+
"eval_steps_per_second": 55.934,
|
| 2319 |
+
"step": 1680000
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 6.92,
|
| 2323 |
+
"eval_loss": 2.4499940872192383,
|
| 2324 |
+
"eval_runtime": 229.0726,
|
| 2325 |
+
"eval_samples_per_second": 897.139,
|
| 2326 |
+
"eval_steps_per_second": 56.074,
|
| 2327 |
+
"step": 1688000
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 6.95,
|
| 2331 |
+
"learning_rate": 1.2026666666666666e-07,
|
| 2332 |
+
"loss": 2.5089,
|
| 2333 |
+
"step": 1696000
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 6.95,
|
| 2337 |
+
"eval_loss": 2.4512550830841064,
|
| 2338 |
+
"eval_runtime": 228.4897,
|
| 2339 |
+
"eval_samples_per_second": 899.428,
|
| 2340 |
+
"eval_steps_per_second": 56.217,
|
| 2341 |
+
"step": 1696000
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 6.98,
|
| 2345 |
+
"eval_loss": 2.4418201446533203,
|
| 2346 |
+
"eval_runtime": 229.377,
|
| 2347 |
+
"eval_samples_per_second": 895.949,
|
| 2348 |
+
"eval_steps_per_second": 56.0,
|
| 2349 |
+
"step": 1704000
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 7.02,
|
| 2353 |
+
"learning_rate": 1.1753333333333334e-07,
|
| 2354 |
+
"loss": 2.5174,
|
| 2355 |
+
"step": 1712000
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 7.02,
|
| 2359 |
+
"eval_loss": 2.447690010070801,
|
| 2360 |
+
"eval_runtime": 231.0137,
|
| 2361 |
+
"eval_samples_per_second": 889.601,
|
| 2362 |
+
"eval_steps_per_second": 55.603,
|
| 2363 |
+
"step": 1712000
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 7.05,
|
| 2367 |
+
"eval_loss": 2.450817584991455,
|
| 2368 |
+
"eval_runtime": 231.6212,
|
| 2369 |
+
"eval_samples_per_second": 887.268,
|
| 2370 |
+
"eval_steps_per_second": 55.457,
|
| 2371 |
+
"step": 1720000
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 7.08,
|
| 2375 |
+
"learning_rate": 1.1480000000000001e-07,
|
| 2376 |
+
"loss": 2.5198,
|
| 2377 |
+
"step": 1728000
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 7.08,
|
| 2381 |
+
"eval_loss": 2.448648691177368,
|
| 2382 |
+
"eval_runtime": 230.9308,
|
| 2383 |
+
"eval_samples_per_second": 889.92,
|
| 2384 |
+
"eval_steps_per_second": 55.623,
|
| 2385 |
+
"step": 1728000
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 7.11,
|
| 2389 |
+
"eval_loss": 2.4577322006225586,
|
| 2390 |
+
"eval_runtime": 230.4865,
|
| 2391 |
+
"eval_samples_per_second": 891.636,
|
| 2392 |
+
"eval_steps_per_second": 55.73,
|
| 2393 |
+
"step": 1736000
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 7.15,
|
| 2397 |
+
"learning_rate": 1.1206666666666666e-07,
|
| 2398 |
+
"loss": 2.4974,
|
| 2399 |
+
"step": 1744000
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 7.15,
|
| 2403 |
+
"eval_loss": 2.4416255950927734,
|
| 2404 |
+
"eval_runtime": 229.8237,
|
| 2405 |
+
"eval_samples_per_second": 894.207,
|
| 2406 |
+
"eval_steps_per_second": 55.891,
|
| 2407 |
+
"step": 1744000
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 7.18,
|
| 2411 |
+
"eval_loss": 2.4549336433410645,
|
| 2412 |
+
"eval_runtime": 229.7571,
|
| 2413 |
+
"eval_samples_per_second": 894.466,
|
| 2414 |
+
"eval_steps_per_second": 55.907,
|
| 2415 |
+
"step": 1752000
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 7.21,
|
| 2419 |
+
"learning_rate": 1.0933333333333333e-07,
|
| 2420 |
+
"loss": 2.5016,
|
| 2421 |
+
"step": 1760000
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 7.21,
|
| 2425 |
+
"eval_loss": 2.455679416656494,
|
| 2426 |
+
"eval_runtime": 230.1335,
|
| 2427 |
+
"eval_samples_per_second": 893.003,
|
| 2428 |
+
"eval_steps_per_second": 55.815,
|
| 2429 |
+
"step": 1760000
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 7.24,
|
| 2433 |
+
"eval_loss": 2.4531571865081787,
|
| 2434 |
+
"eval_runtime": 231.3847,
|
| 2435 |
+
"eval_samples_per_second": 888.175,
|
| 2436 |
+
"eval_steps_per_second": 55.514,
|
| 2437 |
+
"step": 1768000
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 7.28,
|
| 2441 |
+
"learning_rate": 1.066e-07,
|
| 2442 |
+
"loss": 2.5112,
|
| 2443 |
+
"step": 1776000
|
| 2444 |
+
},
|
| 2445 |
+
{
|
| 2446 |
+
"epoch": 7.28,
|
| 2447 |
+
"eval_loss": 2.445054531097412,
|
| 2448 |
+
"eval_runtime": 231.2999,
|
| 2449 |
+
"eval_samples_per_second": 888.5,
|
| 2450 |
+
"eval_steps_per_second": 55.534,
|
| 2451 |
+
"step": 1776000
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 7.31,
|
| 2455 |
+
"eval_loss": 2.460723638534546,
|
| 2456 |
+
"eval_runtime": 230.196,
|
| 2457 |
+
"eval_samples_per_second": 892.761,
|
| 2458 |
+
"eval_steps_per_second": 55.8,
|
| 2459 |
+
"step": 1784000
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 7.34,
|
| 2463 |
+
"learning_rate": 1.0386666666666667e-07,
|
| 2464 |
+
"loss": 2.5172,
|
| 2465 |
+
"step": 1792000
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 7.34,
|
| 2469 |
+
"eval_loss": 2.4451537132263184,
|
| 2470 |
+
"eval_runtime": 231.1406,
|
| 2471 |
+
"eval_samples_per_second": 889.112,
|
| 2472 |
+
"eval_steps_per_second": 55.572,
|
| 2473 |
+
"step": 1792000
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 7.38,
|
| 2477 |
+
"eval_loss": 2.4426777362823486,
|
| 2478 |
+
"eval_runtime": 230.7159,
|
| 2479 |
+
"eval_samples_per_second": 890.749,
|
| 2480 |
+
"eval_steps_per_second": 55.675,
|
| 2481 |
+
"step": 1800000
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 7.41,
|
| 2485 |
+
"learning_rate": 1.0113333333333334e-07,
|
| 2486 |
+
"loss": 2.5089,
|
| 2487 |
+
"step": 1808000
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 7.41,
|
| 2491 |
+
"eval_loss": 2.4511077404022217,
|
| 2492 |
+
"eval_runtime": 231.5975,
|
| 2493 |
+
"eval_samples_per_second": 887.359,
|
| 2494 |
+
"eval_steps_per_second": 55.463,
|
| 2495 |
+
"step": 1808000
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 7.44,
|
| 2499 |
+
"eval_loss": 2.4440526962280273,
|
| 2500 |
+
"eval_runtime": 231.4447,
|
| 2501 |
+
"eval_samples_per_second": 887.944,
|
| 2502 |
+
"eval_steps_per_second": 55.499,
|
| 2503 |
+
"step": 1816000
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 7.47,
|
| 2507 |
+
"learning_rate": 9.84e-08,
|
| 2508 |
+
"loss": 2.5136,
|
| 2509 |
+
"step": 1824000
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 7.47,
|
| 2513 |
+
"eval_loss": 2.4492361545562744,
|
| 2514 |
+
"eval_runtime": 231.7181,
|
| 2515 |
+
"eval_samples_per_second": 886.896,
|
| 2516 |
+
"eval_steps_per_second": 55.434,
|
| 2517 |
+
"step": 1824000
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 7.51,
|
| 2521 |
+
"eval_loss": 2.4523823261260986,
|
| 2522 |
+
"eval_runtime": 231.3659,
|
| 2523 |
+
"eval_samples_per_second": 888.247,
|
| 2524 |
+
"eval_steps_per_second": 55.518,
|
| 2525 |
+
"step": 1832000
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 7.54,
|
| 2529 |
+
"learning_rate": 9.566666666666666e-08,
|
| 2530 |
+
"loss": 2.509,
|
| 2531 |
+
"step": 1840000
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 7.54,
|
| 2535 |
+
"eval_loss": 2.451181411743164,
|
| 2536 |
+
"eval_runtime": 230.8127,
|
| 2537 |
+
"eval_samples_per_second": 890.376,
|
| 2538 |
+
"eval_steps_per_second": 55.651,
|
| 2539 |
+
"step": 1840000
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 7.57,
|
| 2543 |
+
"eval_loss": 2.4528069496154785,
|
| 2544 |
+
"eval_runtime": 230.6096,
|
| 2545 |
+
"eval_samples_per_second": 891.16,
|
| 2546 |
+
"eval_steps_per_second": 55.7,
|
| 2547 |
+
"step": 1848000
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 7.61,
|
| 2551 |
+
"learning_rate": 9.293333333333333e-08,
|
| 2552 |
+
"loss": 2.5157,
|
| 2553 |
+
"step": 1856000
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 7.61,
|
| 2557 |
+
"eval_loss": 2.4439537525177,
|
| 2558 |
+
"eval_runtime": 233.382,
|
| 2559 |
+
"eval_samples_per_second": 880.573,
|
| 2560 |
+
"eval_steps_per_second": 55.039,
|
| 2561 |
+
"step": 1856000
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 7.64,
|
| 2565 |
+
"eval_loss": 2.4401602745056152,
|
| 2566 |
+
"eval_runtime": 231.584,
|
| 2567 |
+
"eval_samples_per_second": 887.41,
|
| 2568 |
+
"eval_steps_per_second": 55.466,
|
| 2569 |
+
"step": 1864000
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 7.67,
|
| 2573 |
+
"learning_rate": 9.02e-08,
|
| 2574 |
+
"loss": 2.5181,
|
| 2575 |
+
"step": 1872000
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 7.67,
|
| 2579 |
+
"eval_loss": 2.4537830352783203,
|
| 2580 |
+
"eval_runtime": 230.4518,
|
| 2581 |
+
"eval_samples_per_second": 891.77,
|
| 2582 |
+
"eval_steps_per_second": 55.738,
|
| 2583 |
+
"step": 1872000
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 7.7,
|
| 2587 |
+
"eval_loss": 2.4480724334716797,
|
| 2588 |
+
"eval_runtime": 229.9532,
|
| 2589 |
+
"eval_samples_per_second": 893.703,
|
| 2590 |
+
"eval_steps_per_second": 55.859,
|
| 2591 |
+
"step": 1880000
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 7.74,
|
| 2595 |
+
"learning_rate": 8.746666666666667e-08,
|
| 2596 |
+
"loss": 2.5145,
|
| 2597 |
+
"step": 1888000
|
| 2598 |
+
},
|
| 2599 |
+
{
|
| 2600 |
+
"epoch": 7.74,
|
| 2601 |
+
"eval_loss": 2.4417428970336914,
|
| 2602 |
+
"eval_runtime": 231.464,
|
| 2603 |
+
"eval_samples_per_second": 887.87,
|
| 2604 |
+
"eval_steps_per_second": 55.495,
|
| 2605 |
+
"step": 1888000
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 7.77,
|
| 2609 |
+
"eval_loss": 2.4512147903442383,
|
| 2610 |
+
"eval_runtime": 231.0711,
|
| 2611 |
+
"eval_samples_per_second": 889.38,
|
| 2612 |
+
"eval_steps_per_second": 55.589,
|
| 2613 |
+
"step": 1896000
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 7.8,
|
| 2617 |
+
"learning_rate": 8.473333333333334e-08,
|
| 2618 |
+
"loss": 2.5013,
|
| 2619 |
+
"step": 1904000
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 7.8,
|
| 2623 |
+
"eval_loss": 2.45603084564209,
|
| 2624 |
+
"eval_runtime": 231.877,
|
| 2625 |
+
"eval_samples_per_second": 886.289,
|
| 2626 |
+
"eval_steps_per_second": 55.396,
|
| 2627 |
+
"step": 1904000
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 7.83,
|
| 2631 |
+
"eval_loss": 2.4508955478668213,
|
| 2632 |
+
"eval_runtime": 230.4147,
|
| 2633 |
+
"eval_samples_per_second": 891.913,
|
| 2634 |
+
"eval_steps_per_second": 55.747,
|
| 2635 |
+
"step": 1912000
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 7.87,
|
| 2639 |
+
"learning_rate": 8.2e-08,
|
| 2640 |
+
"loss": 2.5064,
|
| 2641 |
+
"step": 1920000
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 7.87,
|
| 2645 |
+
"eval_loss": 2.447256565093994,
|
| 2646 |
+
"eval_runtime": 231.4505,
|
| 2647 |
+
"eval_samples_per_second": 887.922,
|
| 2648 |
+
"eval_steps_per_second": 55.498,
|
| 2649 |
+
"step": 1920000
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 7.9,
|
| 2653 |
+
"eval_loss": 2.457575559616089,
|
| 2654 |
+
"eval_runtime": 232.2387,
|
| 2655 |
+
"eval_samples_per_second": 884.908,
|
| 2656 |
+
"eval_steps_per_second": 55.309,
|
| 2657 |
+
"step": 1928000
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 7.93,
|
| 2661 |
+
"learning_rate": 7.926666666666666e-08,
|
| 2662 |
+
"loss": 2.5068,
|
| 2663 |
+
"step": 1936000
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 7.93,
|
| 2667 |
+
"eval_loss": 2.4460949897766113,
|
| 2668 |
+
"eval_runtime": 230.6448,
|
| 2669 |
+
"eval_samples_per_second": 891.024,
|
| 2670 |
+
"eval_steps_per_second": 55.692,
|
| 2671 |
+
"step": 1936000
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 7.97,
|
| 2675 |
+
"eval_loss": 2.4451067447662354,
|
| 2676 |
+
"eval_runtime": 231.6713,
|
| 2677 |
+
"eval_samples_per_second": 887.076,
|
| 2678 |
+
"eval_steps_per_second": 55.445,
|
| 2679 |
+
"step": 1944000
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 8.0,
|
| 2683 |
+
"learning_rate": 7.653333333333333e-08,
|
| 2684 |
+
"loss": 2.5152,
|
| 2685 |
+
"step": 1952000
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 8.0,
|
| 2689 |
+
"eval_loss": 2.442117214202881,
|
| 2690 |
+
"eval_runtime": 231.4315,
|
| 2691 |
+
"eval_samples_per_second": 887.995,
|
| 2692 |
+
"eval_steps_per_second": 55.502,
|
| 2693 |
+
"step": 1952000
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 8.03,
|
| 2697 |
+
"eval_loss": 2.4458179473876953,
|
| 2698 |
+
"eval_runtime": 230.6413,
|
| 2699 |
+
"eval_samples_per_second": 891.037,
|
| 2700 |
+
"eval_steps_per_second": 55.693,
|
| 2701 |
+
"step": 1960000
|
| 2702 |
+
},
|
| 2703 |
+
{
|
| 2704 |
+
"epoch": 8.06,
|
| 2705 |
+
"learning_rate": 7.38e-08,
|
| 2706 |
+
"loss": 2.5025,
|
| 2707 |
+
"step": 1968000
|
| 2708 |
+
},
|
| 2709 |
+
{
|
| 2710 |
+
"epoch": 8.06,
|
| 2711 |
+
"eval_loss": 2.4532368183135986,
|
| 2712 |
+
"eval_runtime": 230.9812,
|
| 2713 |
+
"eval_samples_per_second": 889.726,
|
| 2714 |
+
"eval_steps_per_second": 55.611,
|
| 2715 |
+
"step": 1968000
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 8.1,
|
| 2719 |
+
"eval_loss": 2.4541139602661133,
|
| 2720 |
+
"eval_runtime": 231.1965,
|
| 2721 |
+
"eval_samples_per_second": 888.898,
|
| 2722 |
+
"eval_steps_per_second": 55.559,
|
| 2723 |
+
"step": 1976000
|
| 2724 |
+
},
|
| 2725 |
+
{
|
| 2726 |
+
"epoch": 8.13,
|
| 2727 |
+
"learning_rate": 7.106666666666667e-08,
|
| 2728 |
+
"loss": 2.5151,
|
| 2729 |
+
"step": 1984000
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 8.13,
|
| 2733 |
+
"eval_loss": 2.4499058723449707,
|
| 2734 |
+
"eval_runtime": 231.2124,
|
| 2735 |
+
"eval_samples_per_second": 888.836,
|
| 2736 |
+
"eval_steps_per_second": 55.555,
|
| 2737 |
+
"step": 1984000
|
| 2738 |
+
},
|
| 2739 |
+
{
|
| 2740 |
+
"epoch": 8.16,
|
| 2741 |
+
"eval_loss": 2.4501264095306396,
|
| 2742 |
+
"eval_runtime": 231.2241,
|
| 2743 |
+
"eval_samples_per_second": 888.791,
|
| 2744 |
+
"eval_steps_per_second": 55.552,
|
| 2745 |
+
"step": 1992000
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 8.2,
|
| 2749 |
+
"learning_rate": 6.833333333333332e-08,
|
| 2750 |
+
"loss": 2.5138,
|
| 2751 |
+
"step": 2000000
|
| 2752 |
+
},
|
| 2753 |
+
{
|
| 2754 |
+
"epoch": 8.2,
|
| 2755 |
+
"eval_loss": 2.444784641265869,
|
| 2756 |
+
"eval_runtime": 231.6831,
|
| 2757 |
+
"eval_samples_per_second": 887.031,
|
| 2758 |
+
"eval_steps_per_second": 55.442,
|
| 2759 |
+
"step": 2000000
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 8.23,
|
| 2763 |
+
"eval_loss": 2.4562456607818604,
|
| 2764 |
+
"eval_runtime": 231.974,
|
| 2765 |
+
"eval_samples_per_second": 885.918,
|
| 2766 |
+
"eval_steps_per_second": 55.373,
|
| 2767 |
+
"step": 2008000
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 8.26,
|
| 2771 |
+
"learning_rate": 6.56e-08,
|
| 2772 |
+
"loss": 2.5039,
|
| 2773 |
+
"step": 2016000
|
| 2774 |
+
},
|
| 2775 |
+
{
|
| 2776 |
+
"epoch": 8.26,
|
| 2777 |
+
"eval_loss": 2.4612646102905273,
|
| 2778 |
+
"eval_runtime": 234.4229,
|
| 2779 |
+
"eval_samples_per_second": 876.663,
|
| 2780 |
+
"eval_steps_per_second": 54.794,
|
| 2781 |
+
"step": 2016000
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 8.29,
|
| 2785 |
+
"eval_loss": 2.4471163749694824,
|
| 2786 |
+
"eval_runtime": 233.3806,
|
| 2787 |
+
"eval_samples_per_second": 880.579,
|
| 2788 |
+
"eval_steps_per_second": 55.039,
|
| 2789 |
+
"step": 2024000
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 8.33,
|
| 2793 |
+
"learning_rate": 6.286666666666666e-08,
|
| 2794 |
+
"loss": 2.5055,
|
| 2795 |
+
"step": 2032000
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 8.33,
|
| 2799 |
+
"eval_loss": 2.445026159286499,
|
| 2800 |
+
"eval_runtime": 233.3418,
|
| 2801 |
+
"eval_samples_per_second": 880.725,
|
| 2802 |
+
"eval_steps_per_second": 55.048,
|
| 2803 |
+
"step": 2032000
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 8.36,
|
| 2807 |
+
"eval_loss": 2.4492921829223633,
|
| 2808 |
+
"eval_runtime": 232.3875,
|
| 2809 |
+
"eval_samples_per_second": 884.342,
|
| 2810 |
+
"eval_steps_per_second": 55.274,
|
| 2811 |
+
"step": 2040000
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 8.39,
|
| 2815 |
+
"learning_rate": 6.013333333333333e-08,
|
| 2816 |
+
"loss": 2.5085,
|
| 2817 |
+
"step": 2048000
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 8.39,
|
| 2821 |
+
"eval_loss": 2.448164224624634,
|
| 2822 |
+
"eval_runtime": 233.4578,
|
| 2823 |
+
"eval_samples_per_second": 880.288,
|
| 2824 |
+
"eval_steps_per_second": 55.021,
|
| 2825 |
+
"step": 2048000
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 8.42,
|
| 2829 |
+
"eval_loss": 2.4571895599365234,
|
| 2830 |
+
"eval_runtime": 235.4355,
|
| 2831 |
+
"eval_samples_per_second": 872.893,
|
| 2832 |
+
"eval_steps_per_second": 54.558,
|
| 2833 |
+
"step": 2056000
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 8.46,
|
| 2837 |
+
"learning_rate": 5.7400000000000004e-08,
|
| 2838 |
+
"loss": 2.5114,
|
| 2839 |
+
"step": 2064000
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 8.46,
|
| 2843 |
+
"eval_loss": 2.444307804107666,
|
| 2844 |
+
"eval_runtime": 234.4924,
|
| 2845 |
+
"eval_samples_per_second": 876.404,
|
| 2846 |
+
"eval_steps_per_second": 54.778,
|
| 2847 |
+
"step": 2064000
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 8.49,
|
| 2851 |
+
"eval_loss": 2.445603132247925,
|
| 2852 |
+
"eval_runtime": 234.6223,
|
| 2853 |
+
"eval_samples_per_second": 875.919,
|
| 2854 |
+
"eval_steps_per_second": 54.748,
|
| 2855 |
+
"step": 2072000
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 8.52,
|
| 2859 |
+
"learning_rate": 5.4666666666666666e-08,
|
| 2860 |
+
"loss": 2.5132,
|
| 2861 |
+
"step": 2080000
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 8.52,
|
| 2865 |
+
"eval_loss": 2.4528441429138184,
|
| 2866 |
+
"eval_runtime": 234.3887,
|
| 2867 |
+
"eval_samples_per_second": 876.791,
|
| 2868 |
+
"eval_steps_per_second": 54.802,
|
| 2869 |
+
"step": 2080000
|
| 2870 |
+
},
|
| 2871 |
+
{
|
| 2872 |
+
"epoch": 8.56,
|
| 2873 |
+
"eval_loss": 2.449744939804077,
|
| 2874 |
+
"eval_runtime": 233.1003,
|
| 2875 |
+
"eval_samples_per_second": 881.638,
|
| 2876 |
+
"eval_steps_per_second": 55.105,
|
| 2877 |
+
"step": 2088000
|
| 2878 |
+
},
|
| 2879 |
+
{
|
| 2880 |
+
"epoch": 8.59,
|
| 2881 |
+
"learning_rate": 5.1933333333333335e-08,
|
| 2882 |
+
"loss": 2.5072,
|
| 2883 |
+
"step": 2096000
|
| 2884 |
+
},
|
| 2885 |
+
{
|
| 2886 |
+
"epoch": 8.59,
|
| 2887 |
+
"eval_loss": 2.4547877311706543,
|
| 2888 |
+
"eval_runtime": 232.2237,
|
| 2889 |
+
"eval_samples_per_second": 884.966,
|
| 2890 |
+
"eval_steps_per_second": 55.313,
|
| 2891 |
+
"step": 2096000
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 8.62,
|
| 2895 |
+
"eval_loss": 2.4547617435455322,
|
| 2896 |
+
"eval_runtime": 232.0067,
|
| 2897 |
+
"eval_samples_per_second": 885.794,
|
| 2898 |
+
"eval_steps_per_second": 55.365,
|
| 2899 |
+
"step": 2104000
|
| 2900 |
+
},
|
| 2901 |
+
{
|
| 2902 |
+
"epoch": 8.65,
|
| 2903 |
+
"learning_rate": 4.92e-08,
|
| 2904 |
+
"loss": 2.504,
|
| 2905 |
+
"step": 2112000
|
| 2906 |
+
},
|
| 2907 |
+
{
|
| 2908 |
+
"epoch": 8.65,
|
| 2909 |
+
"eval_loss": 2.444261312484741,
|
| 2910 |
+
"eval_runtime": 232.4079,
|
| 2911 |
+
"eval_samples_per_second": 884.264,
|
| 2912 |
+
"eval_steps_per_second": 55.269,
|
| 2913 |
+
"step": 2112000
|
| 2914 |
+
},
|
| 2915 |
+
{
|
| 2916 |
+
"epoch": 8.69,
|
| 2917 |
+
"eval_loss": 2.445204734802246,
|
| 2918 |
+
"eval_runtime": 233.2645,
|
| 2919 |
+
"eval_samples_per_second": 881.017,
|
| 2920 |
+
"eval_steps_per_second": 55.066,
|
| 2921 |
+
"step": 2120000
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 8.72,
|
| 2925 |
+
"learning_rate": 4.6466666666666666e-08,
|
| 2926 |
+
"loss": 2.5128,
|
| 2927 |
+
"step": 2128000
|
| 2928 |
+
},
|
| 2929 |
+
{
|
| 2930 |
+
"epoch": 8.72,
|
| 2931 |
+
"eval_loss": 2.4509565830230713,
|
| 2932 |
+
"eval_runtime": 233.1857,
|
| 2933 |
+
"eval_samples_per_second": 881.315,
|
| 2934 |
+
"eval_steps_per_second": 55.085,
|
| 2935 |
+
"step": 2128000
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 8.75,
|
| 2939 |
+
"eval_loss": 2.447999954223633,
|
| 2940 |
+
"eval_runtime": 233.2452,
|
| 2941 |
+
"eval_samples_per_second": 881.09,
|
| 2942 |
+
"eval_steps_per_second": 55.071,
|
| 2943 |
+
"step": 2136000
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 8.79,
|
| 2947 |
+
"learning_rate": 4.3733333333333335e-08,
|
| 2948 |
+
"loss": 2.5133,
|
| 2949 |
+
"step": 2144000
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 8.79,
|
| 2953 |
+
"eval_loss": 2.4470479488372803,
|
| 2954 |
+
"eval_runtime": 234.9529,
|
| 2955 |
+
"eval_samples_per_second": 874.686,
|
| 2956 |
+
"eval_steps_per_second": 54.671,
|
| 2957 |
+
"step": 2144000
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 8.82,
|
| 2961 |
+
"eval_loss": 2.4436631202697754,
|
| 2962 |
+
"eval_runtime": 234.9836,
|
| 2963 |
+
"eval_samples_per_second": 874.572,
|
| 2964 |
+
"eval_steps_per_second": 54.663,
|
| 2965 |
+
"step": 2152000
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 8.85,
|
| 2969 |
+
"learning_rate": 4.1e-08,
|
| 2970 |
+
"loss": 2.5067,
|
| 2971 |
+
"step": 2160000
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 8.85,
|
| 2975 |
+
"eval_loss": 2.444672107696533,
|
| 2976 |
+
"eval_runtime": 234.3233,
|
| 2977 |
+
"eval_samples_per_second": 877.036,
|
| 2978 |
+
"eval_steps_per_second": 54.817,
|
| 2979 |
+
"step": 2160000
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 8.88,
|
| 2983 |
+
"eval_loss": 2.453118085861206,
|
| 2984 |
+
"eval_runtime": 233.5384,
|
| 2985 |
+
"eval_samples_per_second": 879.984,
|
| 2986 |
+
"eval_steps_per_second": 55.002,
|
| 2987 |
+
"step": 2168000
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 8.92,
|
| 2991 |
+
"learning_rate": 3.8266666666666665e-08,
|
| 2992 |
+
"loss": 2.4996,
|
| 2993 |
+
"step": 2176000
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 8.92,
|
| 2997 |
+
"eval_loss": 2.447479009628296,
|
| 2998 |
+
"eval_runtime": 235.7844,
|
| 2999 |
+
"eval_samples_per_second": 871.601,
|
| 3000 |
+
"eval_steps_per_second": 54.478,
|
| 3001 |
+
"step": 2176000
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 8.95,
|
| 3005 |
+
"eval_loss": 2.4438347816467285,
|
| 3006 |
+
"eval_runtime": 233.6193,
|
| 3007 |
+
"eval_samples_per_second": 879.679,
|
| 3008 |
+
"eval_steps_per_second": 54.983,
|
| 3009 |
+
"step": 2184000
|
| 3010 |
+
},
|
| 3011 |
+
{
|
| 3012 |
+
"epoch": 8.98,
|
| 3013 |
+
"learning_rate": 3.5533333333333334e-08,
|
| 3014 |
+
"loss": 2.5123,
|
| 3015 |
+
"step": 2192000
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 8.98,
|
| 3019 |
+
"eval_loss": 2.4552195072174072,
|
| 3020 |
+
"eval_runtime": 235.201,
|
| 3021 |
+
"eval_samples_per_second": 873.763,
|
| 3022 |
+
"eval_steps_per_second": 54.613,
|
| 3023 |
+
"step": 2192000
|
| 3024 |
+
},
|
| 3025 |
+
{
|
| 3026 |
+
"epoch": 9.01,
|
| 3027 |
+
"eval_loss": 2.4441311359405518,
|
| 3028 |
+
"eval_runtime": 234.6948,
|
| 3029 |
+
"eval_samples_per_second": 875.648,
|
| 3030 |
+
"eval_steps_per_second": 54.731,
|
| 3031 |
+
"step": 2200000
|
| 3032 |
+
},
|
| 3033 |
+
{
|
| 3034 |
+
"epoch": 9.05,
|
| 3035 |
+
"learning_rate": 3.28e-08,
|
| 3036 |
+
"loss": 2.5044,
|
| 3037 |
+
"step": 2208000
|
| 3038 |
+
},
|
| 3039 |
+
{
|
| 3040 |
+
"epoch": 9.05,
|
| 3041 |
+
"eval_loss": 2.4438366889953613,
|
| 3042 |
+
"eval_runtime": 233.1145,
|
| 3043 |
+
"eval_samples_per_second": 881.584,
|
| 3044 |
+
"eval_steps_per_second": 55.102,
|
| 3045 |
+
"step": 2208000
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 9.08,
|
| 3049 |
+
"eval_loss": 2.453371286392212,
|
| 3050 |
+
"eval_runtime": 234.9783,
|
| 3051 |
+
"eval_samples_per_second": 874.592,
|
| 3052 |
+
"eval_steps_per_second": 54.665,
|
| 3053 |
+
"step": 2216000
|
| 3054 |
+
},
|
| 3055 |
+
{
|
| 3056 |
+
"epoch": 9.11,
|
| 3057 |
+
"learning_rate": 3.0066666666666665e-08,
|
| 3058 |
+
"loss": 2.5068,
|
| 3059 |
+
"step": 2224000
|
| 3060 |
+
},
|
| 3061 |
+
{
|
| 3062 |
+
"epoch": 9.11,
|
| 3063 |
+
"eval_loss": 2.449671745300293,
|
| 3064 |
+
"eval_runtime": 232.7881,
|
| 3065 |
+
"eval_samples_per_second": 882.82,
|
| 3066 |
+
"eval_steps_per_second": 55.179,
|
| 3067 |
+
"step": 2224000
|
| 3068 |
+
},
|
| 3069 |
+
{
|
| 3070 |
+
"epoch": 9.15,
|
| 3071 |
+
"eval_loss": 2.444044828414917,
|
| 3072 |
+
"eval_runtime": 233.2255,
|
| 3073 |
+
"eval_samples_per_second": 881.164,
|
| 3074 |
+
"eval_steps_per_second": 55.075,
|
| 3075 |
+
"step": 2232000
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 9.18,
|
| 3079 |
+
"learning_rate": 2.7333333333333333e-08,
|
| 3080 |
+
"loss": 2.5165,
|
| 3081 |
+
"step": 2240000
|
| 3082 |
+
},
|
| 3083 |
+
{
|
| 3084 |
+
"epoch": 9.18,
|
| 3085 |
+
"eval_loss": 2.457695722579956,
|
| 3086 |
+
"eval_runtime": 234.2621,
|
| 3087 |
+
"eval_samples_per_second": 877.265,
|
| 3088 |
+
"eval_steps_per_second": 54.832,
|
| 3089 |
+
"step": 2240000
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 9.21,
|
| 3093 |
+
"eval_loss": 2.4506990909576416,
|
| 3094 |
+
"eval_runtime": 232.4001,
|
| 3095 |
+
"eval_samples_per_second": 884.294,
|
| 3096 |
+
"eval_steps_per_second": 55.271,
|
| 3097 |
+
"step": 2248000
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 9.24,
|
| 3101 |
+
"learning_rate": 2.46e-08,
|
| 3102 |
+
"loss": 2.5087,
|
| 3103 |
+
"step": 2256000
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 9.24,
|
| 3107 |
+
"eval_loss": 2.4494166374206543,
|
| 3108 |
+
"eval_runtime": 233.0606,
|
| 3109 |
+
"eval_samples_per_second": 881.788,
|
| 3110 |
+
"eval_steps_per_second": 55.114,
|
| 3111 |
+
"step": 2256000
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 9.28,
|
| 3115 |
+
"eval_loss": 2.4393150806427,
|
| 3116 |
+
"eval_runtime": 234.1923,
|
| 3117 |
+
"eval_samples_per_second": 877.527,
|
| 3118 |
+
"eval_steps_per_second": 54.848,
|
| 3119 |
+
"step": 2264000
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 9.31,
|
| 3123 |
+
"learning_rate": 2.1866666666666667e-08,
|
| 3124 |
+
"loss": 2.5036,
|
| 3125 |
+
"step": 2272000
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 9.31,
|
| 3129 |
+
"eval_loss": 2.4486756324768066,
|
| 3130 |
+
"eval_runtime": 233.5876,
|
| 3131 |
+
"eval_samples_per_second": 879.798,
|
| 3132 |
+
"eval_steps_per_second": 54.99,
|
| 3133 |
+
"step": 2272000
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 9.34,
|
| 3137 |
+
"eval_loss": 2.442298173904419,
|
| 3138 |
+
"eval_runtime": 233.1053,
|
| 3139 |
+
"eval_samples_per_second": 881.619,
|
| 3140 |
+
"eval_steps_per_second": 55.104,
|
| 3141 |
+
"step": 2280000
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 9.38,
|
| 3145 |
+
"learning_rate": 1.9133333333333333e-08,
|
| 3146 |
+
"loss": 2.5086,
|
| 3147 |
+
"step": 2288000
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 9.38,
|
| 3151 |
+
"eval_loss": 2.4455623626708984,
|
| 3152 |
+
"eval_runtime": 232.7856,
|
| 3153 |
+
"eval_samples_per_second": 882.83,
|
| 3154 |
+
"eval_steps_per_second": 55.18,
|
| 3155 |
+
"step": 2288000
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 9.41,
|
| 3159 |
+
"eval_loss": 2.449575185775757,
|
| 3160 |
+
"eval_runtime": 234.5471,
|
| 3161 |
+
"eval_samples_per_second": 876.199,
|
| 3162 |
+
"eval_steps_per_second": 54.765,
|
| 3163 |
+
"step": 2296000
|
| 3164 |
+
},
|
| 3165 |
+
{
|
| 3166 |
+
"epoch": 9.44,
|
| 3167 |
+
"learning_rate": 1.64e-08,
|
| 3168 |
+
"loss": 2.5034,
|
| 3169 |
+
"step": 2304000
|
| 3170 |
+
},
|
| 3171 |
+
{
|
| 3172 |
+
"epoch": 9.44,
|
| 3173 |
+
"eval_loss": 2.4498891830444336,
|
| 3174 |
+
"eval_runtime": 232.7935,
|
| 3175 |
+
"eval_samples_per_second": 882.8,
|
| 3176 |
+
"eval_steps_per_second": 55.178,
|
| 3177 |
+
"step": 2304000
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 9.47,
|
| 3181 |
+
"eval_loss": 2.4432790279388428,
|
| 3182 |
+
"eval_runtime": 233.6332,
|
| 3183 |
+
"eval_samples_per_second": 879.627,
|
| 3184 |
+
"eval_steps_per_second": 54.979,
|
| 3185 |
+
"step": 2312000
|
| 3186 |
+
},
|
| 3187 |
+
{
|
| 3188 |
+
"epoch": 9.51,
|
| 3189 |
+
"learning_rate": 1.3666666666666667e-08,
|
| 3190 |
+
"loss": 2.5099,
|
| 3191 |
+
"step": 2320000
|
| 3192 |
+
},
|
| 3193 |
+
{
|
| 3194 |
+
"epoch": 9.51,
|
| 3195 |
+
"eval_loss": 2.4534084796905518,
|
| 3196 |
+
"eval_runtime": 233.6601,
|
| 3197 |
+
"eval_samples_per_second": 879.525,
|
| 3198 |
+
"eval_steps_per_second": 54.973,
|
| 3199 |
+
"step": 2320000
|
| 3200 |
+
},
|
| 3201 |
+
{
|
| 3202 |
+
"epoch": 9.54,
|
| 3203 |
+
"eval_loss": 2.4494857788085938,
|
| 3204 |
+
"eval_runtime": 233.4959,
|
| 3205 |
+
"eval_samples_per_second": 880.144,
|
| 3206 |
+
"eval_steps_per_second": 55.012,
|
| 3207 |
+
"step": 2328000
|
| 3208 |
+
},
|
| 3209 |
+
{
|
| 3210 |
+
"epoch": 9.57,
|
| 3211 |
+
"learning_rate": 1.0933333333333334e-08,
|
| 3212 |
+
"loss": 2.5065,
|
| 3213 |
+
"step": 2336000
|
| 3214 |
+
},
|
| 3215 |
+
{
|
| 3216 |
+
"epoch": 9.57,
|
| 3217 |
+
"eval_loss": 2.4510202407836914,
|
| 3218 |
+
"eval_runtime": 233.7041,
|
| 3219 |
+
"eval_samples_per_second": 879.36,
|
| 3220 |
+
"eval_steps_per_second": 54.963,
|
| 3221 |
+
"step": 2336000
|
| 3222 |
+
},
|
| 3223 |
+
{
|
| 3224 |
+
"epoch": 9.6,
|
| 3225 |
+
"eval_loss": 2.4512877464294434,
|
| 3226 |
+
"eval_runtime": 237.5888,
|
| 3227 |
+
"eval_samples_per_second": 864.982,
|
| 3228 |
+
"eval_steps_per_second": 54.064,
|
| 3229 |
+
"step": 2344000
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 9.64,
|
| 3233 |
+
"learning_rate": 8.2e-09,
|
| 3234 |
+
"loss": 2.502,
|
| 3235 |
+
"step": 2352000
|
| 3236 |
+
},
|
| 3237 |
+
{
|
| 3238 |
+
"epoch": 9.64,
|
| 3239 |
+
"eval_loss": 2.451225996017456,
|
| 3240 |
+
"eval_runtime": 233.4207,
|
| 3241 |
+
"eval_samples_per_second": 880.427,
|
| 3242 |
+
"eval_steps_per_second": 55.029,
|
| 3243 |
+
"step": 2352000
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 9.67,
|
| 3247 |
+
"eval_loss": 2.4469268321990967,
|
| 3248 |
+
"eval_runtime": 234.7938,
|
| 3249 |
+
"eval_samples_per_second": 875.279,
|
| 3250 |
+
"eval_steps_per_second": 54.708,
|
| 3251 |
+
"step": 2360000
|
| 3252 |
+
},
|
| 3253 |
+
{
|
| 3254 |
+
"epoch": 9.7,
|
| 3255 |
+
"learning_rate": 5.466666666666667e-09,
|
| 3256 |
+
"loss": 2.5043,
|
| 3257 |
+
"step": 2368000
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 9.7,
|
| 3261 |
+
"eval_loss": 2.4544479846954346,
|
| 3262 |
+
"eval_runtime": 234.2869,
|
| 3263 |
+
"eval_samples_per_second": 877.173,
|
| 3264 |
+
"eval_steps_per_second": 54.826,
|
| 3265 |
+
"step": 2368000
|
| 3266 |
+
},
|
| 3267 |
+
{
|
| 3268 |
+
"epoch": 9.74,
|
| 3269 |
+
"eval_loss": 2.4492740631103516,
|
| 3270 |
+
"eval_runtime": 234.1805,
|
| 3271 |
+
"eval_samples_per_second": 877.571,
|
| 3272 |
+
"eval_steps_per_second": 54.851,
|
| 3273 |
+
"step": 2376000
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 9.77,
|
| 3277 |
+
"learning_rate": 2.7333333333333334e-09,
|
| 3278 |
+
"loss": 2.5068,
|
| 3279 |
+
"step": 2384000
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 9.77,
|
| 3283 |
+
"eval_loss": 2.453711748123169,
|
| 3284 |
+
"eval_runtime": 233.0608,
|
| 3285 |
+
"eval_samples_per_second": 881.787,
|
| 3286 |
+
"eval_steps_per_second": 55.114,
|
| 3287 |
+
"step": 2384000
|
| 3288 |
+
},
|
| 3289 |
+
{
|
| 3290 |
+
"epoch": 9.8,
|
| 3291 |
+
"eval_loss": 2.4386837482452393,
|
| 3292 |
+
"eval_runtime": 234.1662,
|
| 3293 |
+
"eval_samples_per_second": 877.625,
|
| 3294 |
+
"eval_steps_per_second": 54.854,
|
| 3295 |
+
"step": 2392000
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 9.83,
|
| 3299 |
+
"learning_rate": 0.0,
|
| 3300 |
+
"loss": 2.5118,
|
| 3301 |
+
"step": 2400000
|
| 3302 |
+
},
|
| 3303 |
+
{
|
| 3304 |
+
"epoch": 9.83,
|
| 3305 |
+
"eval_loss": 2.4494030475616455,
|
| 3306 |
+
"eval_runtime": 234.0187,
|
| 3307 |
+
"eval_samples_per_second": 878.178,
|
| 3308 |
+
"eval_steps_per_second": 54.889,
|
| 3309 |
+
"step": 2400000
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 9.83,
|
| 3313 |
+
"step": 2400000,
|
| 3314 |
+
"total_flos": 7.305293129309786e+17,
|
| 3315 |
+
"train_loss": 2.5438934391276042,
|
| 3316 |
+
"train_runtime": 220778.1092,
|
| 3317 |
+
"train_samples_per_second": 173.93,
|
| 3318 |
+
"train_steps_per_second": 10.871
|
| 3319 |
+
}
|
| 3320 |
+
],
|
| 3321 |
+
"logging_steps": 16000,
|
| 3322 |
+
"max_steps": 2400000,
|
| 3323 |
+
"num_train_epochs": 10,
|
| 3324 |
+
"save_steps": 32000,
|
| 3325 |
+
"total_flos": 7.305293129309786e+17,
|
| 3326 |
+
"trial_name": null,
|
| 3327 |
+
"trial_params": null
|
| 3328 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2802db4c811c667007fed18e8662dd6e460e87647e53480984e9ddd11481bb7b
|
| 3 |
+
size 4219
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|