Commit
•
46eb42d
1
Parent(s):
5c37bb3
Training in progress, step 32000
Browse files- README.md +357 -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 +1,357 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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-Q1-90p-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-Q1-90p-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: 3.2574
|
| 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: 1e-05
|
| 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 |
+
- lr_scheduler_warmup_steps: 1400
|
| 44 |
+
- training_steps: 2400000
|
| 45 |
+
|
| 46 |
+
### Training results
|
| 47 |
+
|
| 48 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 49 |
+
|:-------------:|:-----:|:-------:|:---------------:|
|
| 50 |
+
| No log | 0.16 | 8000 | 3.4495 |
|
| 51 |
+
| 3.5684 | 0.33 | 16000 | 3.4166 |
|
| 52 |
+
| 3.5684 | 0.49 | 24000 | 3.3847 |
|
| 53 |
+
| 3.3755 | 0.66 | 32000 | 3.3665 |
|
| 54 |
+
| 3.3755 | 0.82 | 40000 | 3.3654 |
|
| 55 |
+
| 3.3533 | 0.98 | 48000 | 3.3654 |
|
| 56 |
+
| 3.3533 | 1.15 | 56000 | 3.3328 |
|
| 57 |
+
| 3.3014 | 1.31 | 64000 | 3.3210 |
|
| 58 |
+
| 3.3014 | 1.48 | 72000 | 3.3492 |
|
| 59 |
+
| 3.2888 | 1.64 | 80000 | 3.3213 |
|
| 60 |
+
| 3.2888 | 1.8 | 88000 | 3.2708 |
|
| 61 |
+
| 3.2609 | 1.97 | 96000 | 3.2908 |
|
| 62 |
+
| 3.2609 | 2.13 | 104000 | 3.2767 |
|
| 63 |
+
| 3.2159 | 2.29 | 112000 | 3.2592 |
|
| 64 |
+
| 3.2159 | 2.46 | 120000 | 3.2411 |
|
| 65 |
+
| 3.2167 | 2.62 | 128000 | 3.2377 |
|
| 66 |
+
| 3.2167 | 2.79 | 136000 | 3.2485 |
|
| 67 |
+
| 3.199 | 2.95 | 144000 | 3.2609 |
|
| 68 |
+
| 3.199 | 3.11 | 152000 | 3.2553 |
|
| 69 |
+
| 3.1905 | 3.28 | 160000 | 3.2425 |
|
| 70 |
+
| 3.1905 | 3.44 | 168000 | 3.2422 |
|
| 71 |
+
| 3.1822 | 3.61 | 176000 | 3.2624 |
|
| 72 |
+
| 3.1822 | 3.77 | 184000 | 3.2507 |
|
| 73 |
+
| 3.1852 | 3.93 | 192000 | 3.2483 |
|
| 74 |
+
| 3.1852 | 4.1 | 200000 | 3.2514 |
|
| 75 |
+
| 3.1767 | 4.26 | 208000 | 3.2426 |
|
| 76 |
+
| 3.1767 | 4.43 | 216000 | 3.2348 |
|
| 77 |
+
| 3.1767 | 4.59 | 224000 | 3.2735 |
|
| 78 |
+
| 3.1767 | 4.75 | 232000 | 3.2472 |
|
| 79 |
+
| 3.1973 | 4.92 | 240000 | 3.2596 |
|
| 80 |
+
| 3.1973 | 5.08 | 248000 | 3.2606 |
|
| 81 |
+
| 3.1781 | 5.24 | 256000 | 3.2815 |
|
| 82 |
+
| 3.1781 | 5.41 | 264000 | 3.2734 |
|
| 83 |
+
| 3.1803 | 5.57 | 272000 | 3.2739 |
|
| 84 |
+
| 3.1803 | 5.74 | 280000 | 3.2712 |
|
| 85 |
+
| 3.1989 | 5.9 | 288000 | 3.2734 |
|
| 86 |
+
| 3.1989 | 6.06 | 296000 | 3.2939 |
|
| 87 |
+
| 3.1929 | 6.23 | 304000 | 3.2880 |
|
| 88 |
+
| 3.1929 | 6.39 | 312000 | 3.2894 |
|
| 89 |
+
| 3.2083 | 6.56 | 320000 | 3.3086 |
|
| 90 |
+
| 3.2083 | 6.72 | 328000 | 3.3067 |
|
| 91 |
+
| 3.2013 | 6.88 | 336000 | 3.2787 |
|
| 92 |
+
| 3.2013 | 7.05 | 344000 | 3.3153 |
|
| 93 |
+
| 3.2111 | 7.21 | 352000 | 3.3246 |
|
| 94 |
+
| 3.2111 | 7.38 | 360000 | 3.3323 |
|
| 95 |
+
| 3.2186 | 7.54 | 368000 | 3.2938 |
|
| 96 |
+
| 3.2186 | 7.7 | 376000 | 3.3500 |
|
| 97 |
+
| 3.2268 | 7.87 | 384000 | 3.3180 |
|
| 98 |
+
| 3.2268 | 8.03 | 392000 | 3.3171 |
|
| 99 |
+
| 3.233 | 8.2 | 400000 | 3.3462 |
|
| 100 |
+
| 3.233 | 8.36 | 408000 | 3.3413 |
|
| 101 |
+
| 3.2432 | 8.52 | 416000 | 3.3281 |
|
| 102 |
+
| 3.2432 | 8.69 | 424000 | 3.3420 |
|
| 103 |
+
| 3.2586 | 8.85 | 432000 | 3.3609 |
|
| 104 |
+
| 3.2586 | 9.01 | 440000 | 3.3527 |
|
| 105 |
+
| 3.2567 | 9.18 | 448000 | 3.3594 |
|
| 106 |
+
| 3.2567 | 9.34 | 456000 | 3.3497 |
|
| 107 |
+
| 3.2592 | 9.51 | 464000 | 3.3607 |
|
| 108 |
+
| 3.2592 | 9.67 | 472000 | 3.3840 |
|
| 109 |
+
| 3.2793 | 9.83 | 480000 | 3.3668 |
|
| 110 |
+
| 3.2793 | 10.0 | 488000 | 3.3609 |
|
| 111 |
+
| 3.257 | 10.16 | 496000 | 3.3682 |
|
| 112 |
+
| 3.257 | 10.33 | 504000 | 3.4006 |
|
| 113 |
+
| 3.2656 | 10.49 | 512000 | 3.3588 |
|
| 114 |
+
| 3.2656 | 10.65 | 520000 | 3.3799 |
|
| 115 |
+
| 3.2727 | 10.82 | 528000 | 3.3833 |
|
| 116 |
+
| 3.2727 | 10.98 | 536000 | 3.3566 |
|
| 117 |
+
| 3.2705 | 11.15 | 544000 | 3.3794 |
|
| 118 |
+
| 3.2705 | 11.31 | 552000 | 3.3838 |
|
| 119 |
+
| 3.2676 | 11.47 | 560000 | 3.3660 |
|
| 120 |
+
| 3.2676 | 11.64 | 568000 | 3.3938 |
|
| 121 |
+
| 3.258 | 11.8 | 576000 | 3.3661 |
|
| 122 |
+
| 3.258 | 11.97 | 584000 | 3.3490 |
|
| 123 |
+
| 3.2646 | 12.13 | 592000 | 3.3716 |
|
| 124 |
+
| 3.2646 | 12.29 | 600000 | 3.3877 |
|
| 125 |
+
| 3.2578 | 12.46 | 608000 | 3.3930 |
|
| 126 |
+
| 3.2578 | 12.62 | 616000 | 3.3921 |
|
| 127 |
+
| 3.2719 | 12.78 | 624000 | 3.3957 |
|
| 128 |
+
| 3.2719 | 12.95 | 632000 | 3.4196 |
|
| 129 |
+
| 3.2828 | 13.11 | 640000 | 3.4078 |
|
| 130 |
+
| 3.2828 | 13.28 | 648000 | 3.4203 |
|
| 131 |
+
| 3.2805 | 13.44 | 656000 | 3.3900 |
|
| 132 |
+
| 3.2805 | 13.6 | 664000 | 3.4038 |
|
| 133 |
+
| 3.2975 | 13.77 | 672000 | 3.4056 |
|
| 134 |
+
| 3.2975 | 13.93 | 680000 | 3.4284 |
|
| 135 |
+
| 3.2965 | 14.1 | 688000 | 3.4180 |
|
| 136 |
+
| 3.2965 | 14.26 | 696000 | 3.4196 |
|
| 137 |
+
| 3.3069 | 14.42 | 704000 | 3.4257 |
|
| 138 |
+
| 3.3069 | 14.59 | 712000 | 3.4299 |
|
| 139 |
+
| 3.3152 | 14.75 | 720000 | 3.4788 |
|
| 140 |
+
| 3.3152 | 14.92 | 728000 | 3.4425 |
|
| 141 |
+
| 3.3125 | 15.08 | 736000 | 3.4301 |
|
| 142 |
+
| 3.3125 | 15.24 | 744000 | 3.4441 |
|
| 143 |
+
| 3.3174 | 15.41 | 752000 | 3.4396 |
|
| 144 |
+
| 3.3174 | 15.57 | 760000 | 3.4639 |
|
| 145 |
+
| 3.3242 | 15.73 | 768000 | 3.4524 |
|
| 146 |
+
| 3.3242 | 15.9 | 776000 | 3.4560 |
|
| 147 |
+
| 3.3385 | 16.06 | 784000 | 3.4780 |
|
| 148 |
+
| 3.3385 | 16.23 | 792000 | 3.4774 |
|
| 149 |
+
| 3.3371 | 16.39 | 800000 | 3.4772 |
|
| 150 |
+
| 3.3371 | 16.55 | 808000 | 3.4955 |
|
| 151 |
+
| 3.3633 | 16.72 | 816000 | 3.4861 |
|
| 152 |
+
| 3.3633 | 16.88 | 824000 | 3.5063 |
|
| 153 |
+
| 3.3678 | 17.05 | 832000 | 3.5044 |
|
| 154 |
+
| 3.3678 | 17.21 | 840000 | 3.5202 |
|
| 155 |
+
| 3.3634 | 17.37 | 848000 | 3.4941 |
|
| 156 |
+
| 3.3634 | 17.54 | 856000 | 3.5223 |
|
| 157 |
+
| 3.3797 | 17.7 | 864000 | 3.5028 |
|
| 158 |
+
| 3.3797 | 17.87 | 872000 | 3.5264 |
|
| 159 |
+
| 3.3802 | 18.03 | 880000 | 3.5313 |
|
| 160 |
+
| 3.3802 | 18.19 | 888000 | 3.4963 |
|
| 161 |
+
| 3.357 | 18.36 | 896000 | 3.5171 |
|
| 162 |
+
| 3.357 | 18.52 | 904000 | 3.5307 |
|
| 163 |
+
| 3.3866 | 18.69 | 912000 | 3.5222 |
|
| 164 |
+
| 3.3866 | 18.85 | 920000 | 3.5319 |
|
| 165 |
+
| 3.3818 | 19.01 | 928000 | 3.5326 |
|
| 166 |
+
| 3.3818 | 19.18 | 936000 | 3.5116 |
|
| 167 |
+
| 3.3754 | 19.34 | 944000 | 3.5229 |
|
| 168 |
+
| 3.3754 | 19.5 | 952000 | 3.5383 |
|
| 169 |
+
| 3.3893 | 19.67 | 960000 | 3.5445 |
|
| 170 |
+
| 3.3893 | 19.83 | 968000 | 3.5231 |
|
| 171 |
+
| 3.3899 | 20.0 | 976000 | 3.5310 |
|
| 172 |
+
| 3.3899 | 20.16 | 984000 | 3.5329 |
|
| 173 |
+
| 3.3918 | 20.32 | 992000 | 3.5159 |
|
| 174 |
+
| 3.3918 | 20.49 | 1000000 | 3.5628 |
|
| 175 |
+
| 3.3786 | 20.65 | 1008000 | 3.5291 |
|
| 176 |
+
| 3.3786 | 20.82 | 1016000 | 3.5163 |
|
| 177 |
+
| 3.3862 | 20.98 | 1024000 | 3.5312 |
|
| 178 |
+
| 3.3862 | 21.14 | 1032000 | 3.5140 |
|
| 179 |
+
| 3.3855 | 21.31 | 1040000 | 3.5617 |
|
| 180 |
+
| 3.3855 | 21.47 | 1048000 | 3.5375 |
|
| 181 |
+
| 3.3872 | 21.64 | 1056000 | 3.5328 |
|
| 182 |
+
| 3.3872 | 21.8 | 1064000 | 3.5616 |
|
| 183 |
+
| 3.3931 | 21.96 | 1072000 | 3.5648 |
|
| 184 |
+
| 3.3931 | 22.13 | 1080000 | 3.5443 |
|
| 185 |
+
| 3.3708 | 22.29 | 1088000 | 3.5401 |
|
| 186 |
+
| 3.3708 | 22.45 | 1096000 | 3.5529 |
|
| 187 |
+
| 3.4099 | 22.62 | 1104000 | 3.5334 |
|
| 188 |
+
| 3.4099 | 22.78 | 1112000 | 3.5325 |
|
| 189 |
+
| 3.4027 | 22.95 | 1120000 | 3.5819 |
|
| 190 |
+
| 3.4027 | 23.11 | 1128000 | 3.5471 |
|
| 191 |
+
| 3.4035 | 23.27 | 1136000 | 3.5486 |
|
| 192 |
+
| 3.4035 | 23.44 | 1144000 | 3.5470 |
|
| 193 |
+
| 3.3964 | 23.6 | 1152000 | 3.5722 |
|
| 194 |
+
| 3.3964 | 23.77 | 1160000 | 3.5510 |
|
| 195 |
+
| 3.4115 | 23.93 | 1168000 | 3.5610 |
|
| 196 |
+
| 3.4115 | 24.09 | 1176000 | 3.5757 |
|
| 197 |
+
| 3.4173 | 24.26 | 1184000 | 3.5541 |
|
| 198 |
+
| 3.4173 | 24.42 | 1192000 | 3.5777 |
|
| 199 |
+
| 3.4169 | 24.59 | 1200000 | 3.5638 |
|
| 200 |
+
| 3.4169 | 24.75 | 1208000 | 3.5463 |
|
| 201 |
+
| 3.4031 | 24.91 | 1216000 | 3.5300 |
|
| 202 |
+
| 3.4031 | 25.08 | 1224000 | 3.5584 |
|
| 203 |
+
| 3.4094 | 25.24 | 1232000 | 3.5682 |
|
| 204 |
+
| 3.4094 | 25.41 | 1240000 | 3.5558 |
|
| 205 |
+
| 3.4116 | 25.57 | 1248000 | 3.5629 |
|
| 206 |
+
| 3.4116 | 25.73 | 1256000 | 3.5490 |
|
| 207 |
+
| 3.4199 | 25.9 | 1264000 | 3.5679 |
|
| 208 |
+
| 3.4199 | 26.06 | 1272000 | 3.5885 |
|
| 209 |
+
| 3.412 | 26.22 | 1280000 | 3.5579 |
|
| 210 |
+
| 3.412 | 26.39 | 1288000 | 3.5465 |
|
| 211 |
+
| 3.4123 | 26.55 | 1296000 | 3.5726 |
|
| 212 |
+
| 3.4123 | 26.72 | 1304000 | 3.5775 |
|
| 213 |
+
| 3.4132 | 26.88 | 1312000 | 3.5478 |
|
| 214 |
+
| 3.4132 | 27.04 | 1320000 | 3.5589 |
|
| 215 |
+
| 3.4161 | 27.21 | 1328000 | 3.5662 |
|
| 216 |
+
| 3.4161 | 27.37 | 1336000 | 3.5895 |
|
| 217 |
+
| 3.4097 | 27.54 | 1344000 | 3.5941 |
|
| 218 |
+
| 3.4097 | 27.7 | 1352000 | 3.5912 |
|
| 219 |
+
| 3.415 | 27.86 | 1360000 | 3.5658 |
|
| 220 |
+
| 3.415 | 28.03 | 1368000 | 3.5554 |
|
| 221 |
+
| 3.4193 | 28.19 | 1376000 | 3.5899 |
|
| 222 |
+
| 3.4193 | 28.36 | 1384000 | 3.5652 |
|
| 223 |
+
| 3.4136 | 28.52 | 1392000 | 3.5832 |
|
| 224 |
+
| 3.4136 | 28.68 | 1400000 | 3.5885 |
|
| 225 |
+
| 3.4294 | 28.85 | 1408000 | 3.5832 |
|
| 226 |
+
| 3.4294 | 29.01 | 1416000 | 3.6025 |
|
| 227 |
+
| 3.4243 | 29.17 | 1424000 | 3.6040 |
|
| 228 |
+
| 3.4243 | 29.34 | 1432000 | 3.5890 |
|
| 229 |
+
| 3.4427 | 29.5 | 1440000 | 3.5835 |
|
| 230 |
+
| 3.4427 | 29.67 | 1448000 | 3.6185 |
|
| 231 |
+
| 3.4293 | 29.83 | 1456000 | 3.6029 |
|
| 232 |
+
| 3.4293 | 29.99 | 1464000 | 3.6162 |
|
| 233 |
+
| 3.4363 | 30.16 | 1472000 | 3.6258 |
|
| 234 |
+
| 3.4363 | 30.32 | 1480000 | 3.6038 |
|
| 235 |
+
| 3.4532 | 30.49 | 1488000 | 3.6039 |
|
| 236 |
+
| 3.4532 | 30.65 | 1496000 | 3.6054 |
|
| 237 |
+
| 3.4401 | 30.81 | 1504000 | 3.6269 |
|
| 238 |
+
| 3.4401 | 30.98 | 1512000 | 3.6004 |
|
| 239 |
+
| 3.4491 | 31.14 | 1520000 | 3.6096 |
|
| 240 |
+
| 3.4491 | 31.31 | 1528000 | 3.6217 |
|
| 241 |
+
| 3.4438 | 31.47 | 1536000 | 3.6081 |
|
| 242 |
+
| 3.4438 | 31.63 | 1544000 | 3.6190 |
|
| 243 |
+
| 3.4337 | 31.8 | 1552000 | 3.6120 |
|
| 244 |
+
| 3.4337 | 31.96 | 1560000 | 3.5861 |
|
| 245 |
+
| 3.4475 | 32.13 | 1568000 | 3.6209 |
|
| 246 |
+
| 3.4475 | 32.29 | 1576000 | 3.6302 |
|
| 247 |
+
| 3.4406 | 32.45 | 1584000 | 3.6053 |
|
| 248 |
+
| 3.4406 | 32.62 | 1592000 | 3.5934 |
|
| 249 |
+
| 3.4392 | 32.78 | 1600000 | 3.5942 |
|
| 250 |
+
| 3.4392 | 32.94 | 1608000 | 3.6013 |
|
| 251 |
+
| 3.4514 | 33.11 | 1616000 | 3.6506 |
|
| 252 |
+
| 3.4514 | 33.27 | 1624000 | 3.6049 |
|
| 253 |
+
| 3.4406 | 33.44 | 1632000 | 3.6285 |
|
| 254 |
+
| 3.4406 | 33.6 | 1640000 | 3.6107 |
|
| 255 |
+
| 3.4522 | 33.76 | 1648000 | 3.6081 |
|
| 256 |
+
| 3.4522 | 33.93 | 1656000 | 3.6121 |
|
| 257 |
+
| 3.4592 | 34.09 | 1664000 | 3.6396 |
|
| 258 |
+
| 3.4592 | 34.26 | 1672000 | 3.6284 |
|
| 259 |
+
| 3.4587 | 34.42 | 1680000 | 3.6195 |
|
| 260 |
+
| 3.4587 | 34.58 | 1688000 | 3.6168 |
|
| 261 |
+
| 3.4589 | 34.75 | 1696000 | 3.6315 |
|
| 262 |
+
| 3.4589 | 34.91 | 1704000 | 3.6045 |
|
| 263 |
+
| 3.4703 | 35.08 | 1712000 | 3.6251 |
|
| 264 |
+
| 3.4703 | 35.24 | 1720000 | 3.6252 |
|
| 265 |
+
| 3.4565 | 35.4 | 1728000 | 3.6254 |
|
| 266 |
+
| 3.4565 | 35.57 | 1736000 | 3.6544 |
|
| 267 |
+
| 3.4634 | 35.73 | 1744000 | 3.6290 |
|
| 268 |
+
| 3.4634 | 35.9 | 1752000 | 3.6124 |
|
| 269 |
+
| 3.4625 | 36.06 | 1760000 | 3.6262 |
|
| 270 |
+
| 3.4625 | 36.22 | 1768000 | 3.6318 |
|
| 271 |
+
| 3.457 | 36.39 | 1776000 | 3.6408 |
|
| 272 |
+
| 3.457 | 36.55 | 1784000 | 3.6433 |
|
| 273 |
+
| 3.4618 | 36.71 | 1792000 | 3.6276 |
|
| 274 |
+
| 3.4618 | 36.88 | 1800000 | 3.6314 |
|
| 275 |
+
| 3.4611 | 37.04 | 1808000 | 3.6416 |
|
| 276 |
+
| 3.4611 | 37.21 | 1816000 | 3.6658 |
|
| 277 |
+
| 3.4651 | 37.37 | 1824000 | 3.6382 |
|
| 278 |
+
| 3.4651 | 37.53 | 1832000 | 3.6562 |
|
| 279 |
+
| 3.4625 | 37.7 | 1840000 | 3.6376 |
|
| 280 |
+
| 3.4625 | 37.86 | 1848000 | 3.6520 |
|
| 281 |
+
| 3.4561 | 38.03 | 1856000 | 3.6301 |
|
| 282 |
+
| 3.4561 | 38.19 | 1864000 | 3.6195 |
|
| 283 |
+
| 3.4655 | 38.35 | 1872000 | 3.6279 |
|
| 284 |
+
| 3.4655 | 38.52 | 1880000 | 3.6365 |
|
| 285 |
+
| 3.4637 | 38.68 | 1888000 | 3.6386 |
|
| 286 |
+
| 3.4637 | 38.85 | 1896000 | 3.6434 |
|
| 287 |
+
| 3.458 | 39.01 | 1904000 | 3.6519 |
|
| 288 |
+
| 3.458 | 39.17 | 1912000 | 3.6438 |
|
| 289 |
+
| 3.4523 | 39.34 | 1920000 | 3.6408 |
|
| 290 |
+
| 3.4523 | 39.5 | 1928000 | 3.6513 |
|
| 291 |
+
| 3.4743 | 39.66 | 1936000 | 3.6178 |
|
| 292 |
+
| 3.4743 | 39.83 | 1944000 | 3.6399 |
|
| 293 |
+
| 3.4626 | 39.99 | 1952000 | 3.6243 |
|
| 294 |
+
| 3.4626 | 40.16 | 1960000 | 3.6326 |
|
| 295 |
+
| 3.4692 | 40.32 | 1968000 | 3.6723 |
|
| 296 |
+
| 3.4692 | 40.48 | 1976000 | 3.6456 |
|
| 297 |
+
| 3.4765 | 40.65 | 1984000 | 3.6437 |
|
| 298 |
+
| 3.4765 | 40.81 | 1992000 | 3.6477 |
|
| 299 |
+
| 3.4747 | 40.98 | 2000000 | 3.6384 |
|
| 300 |
+
| 3.4747 | 41.14 | 2008000 | 3.6370 |
|
| 301 |
+
| 3.4683 | 41.3 | 2016000 | 3.6625 |
|
| 302 |
+
| 3.4683 | 41.47 | 2024000 | 3.6453 |
|
| 303 |
+
| 3.4599 | 41.63 | 2032000 | 3.6489 |
|
| 304 |
+
| 3.4599 | 41.8 | 2040000 | 3.6311 |
|
| 305 |
+
| 3.4713 | 41.96 | 2048000 | 3.6192 |
|
| 306 |
+
| 3.4713 | 42.12 | 2056000 | 3.6511 |
|
| 307 |
+
| 3.4677 | 42.29 | 2064000 | 3.6426 |
|
| 308 |
+
| 3.4677 | 42.45 | 2072000 | 3.6363 |
|
| 309 |
+
| 3.4689 | 42.62 | 2080000 | 3.6378 |
|
| 310 |
+
| 3.4689 | 42.78 | 2088000 | 3.6450 |
|
| 311 |
+
| 3.4598 | 42.94 | 2096000 | 3.6481 |
|
| 312 |
+
| 3.4598 | 43.11 | 2104000 | 3.6675 |
|
| 313 |
+
| 3.4487 | 43.27 | 2112000 | 3.6558 |
|
| 314 |
+
| 3.4487 | 43.43 | 2120000 | 3.6451 |
|
| 315 |
+
| 3.4555 | 43.6 | 2128000 | 3.6431 |
|
| 316 |
+
| 3.4555 | 43.76 | 2136000 | 3.6470 |
|
| 317 |
+
| 3.4727 | 43.93 | 2144000 | 3.6265 |
|
| 318 |
+
| 3.4727 | 44.09 | 2152000 | 3.6335 |
|
| 319 |
+
| 3.4626 | 44.25 | 2160000 | 3.6396 |
|
| 320 |
+
| 3.4626 | 44.42 | 2168000 | 3.6537 |
|
| 321 |
+
| 3.4724 | 44.58 | 2176000 | 3.6168 |
|
| 322 |
+
| 3.4724 | 44.75 | 2184000 | 3.6444 |
|
| 323 |
+
| 3.4545 | 44.91 | 2192000 | 3.6440 |
|
| 324 |
+
| 3.4545 | 45.07 | 2200000 | 3.6327 |
|
| 325 |
+
| 3.461 | 45.24 | 2208000 | 3.6363 |
|
| 326 |
+
| 3.461 | 45.4 | 2216000 | 3.6537 |
|
| 327 |
+
| 3.4702 | 45.57 | 2224000 | 3.6123 |
|
| 328 |
+
| 3.4702 | 45.73 | 2232000 | 3.6554 |
|
| 329 |
+
| 3.4565 | 45.89 | 2240000 | 3.6523 |
|
| 330 |
+
| 3.4565 | 46.06 | 2248000 | 3.6340 |
|
| 331 |
+
| 3.4517 | 46.22 | 2256000 | 3.6459 |
|
| 332 |
+
| 3.4517 | 46.38 | 2264000 | 3.6561 |
|
| 333 |
+
| 3.4631 | 46.55 | 2272000 | 3.6548 |
|
| 334 |
+
| 3.4631 | 46.71 | 2280000 | 3.6229 |
|
| 335 |
+
| 3.4518 | 46.88 | 2288000 | 3.6350 |
|
| 336 |
+
| 3.4518 | 47.04 | 2296000 | 3.6483 |
|
| 337 |
+
| 3.4592 | 47.2 | 2304000 | 3.6263 |
|
| 338 |
+
| 3.4592 | 47.37 | 2312000 | 3.6339 |
|
| 339 |
+
| 3.4569 | 47.53 | 2320000 | 3.6594 |
|
| 340 |
+
| 3.4569 | 47.7 | 2328000 | 3.6385 |
|
| 341 |
+
| 3.4524 | 47.86 | 2336000 | 3.6434 |
|
| 342 |
+
| 3.4524 | 48.02 | 2344000 | 3.6502 |
|
| 343 |
+
| 3.4644 | 48.19 | 2352000 | 3.6176 |
|
| 344 |
+
| 3.4644 | 48.35 | 2360000 | 3.6293 |
|
| 345 |
+
| 3.4586 | 48.52 | 2368000 | 3.6304 |
|
| 346 |
+
| 3.4586 | 48.68 | 2376000 | 3.6343 |
|
| 347 |
+
| 3.4439 | 48.84 | 2384000 | 3.6090 |
|
| 348 |
+
| 3.4439 | 49.01 | 2392000 | 3.6414 |
|
| 349 |
+
| 3.4474 | 49.17 | 2400000 | 3.6208 |
|
| 350 |
+
|
| 351 |
+
|
| 352 |
+
### Framework versions
|
| 353 |
+
|
| 354 |
+
- Transformers 4.35.0.dev0
|
| 355 |
+
- Pytorch 2.0.1+cu117
|
| 356 |
+
- Datasets 2.14.5
|
| 357 |
+
- 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": 49.17,
|
| 3 |
+
"eval_loss": 3.2573602199554443,
|
| 4 |
+
"eval_runtime": 41.9844,
|
| 5 |
+
"eval_samples": 41102,
|
| 6 |
+
"eval_samples_per_second": 978.983,
|
| 7 |
+
"eval_steps_per_second": 61.189,
|
| 8 |
+
"perplexity": 25.980862776689275,
|
| 9 |
+
"train_loss": 3.376089767252604,
|
| 10 |
+
"train_runtime": 158003.2062,
|
| 11 |
+
"train_samples": 780934,
|
| 12 |
+
"train_samples_per_second": 243.033,
|
| 13 |
+
"train_steps_per_second": 15.19
|
| 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": 49.17,
|
| 3 |
+
"eval_loss": 3.2573602199554443,
|
| 4 |
+
"eval_runtime": 41.9844,
|
| 5 |
+
"eval_samples": 41102,
|
| 6 |
+
"eval_samples_per_second": 978.983,
|
| 7 |
+
"eval_steps_per_second": 61.189,
|
| 8 |
+
"perplexity": 25.980862776689275
|
| 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:f0a5a9280e3c9e6a4e56f6d367b5e1ef6f66b6ce35c7a0262a73d3638110800f
|
| 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": 49.17,
|
| 3 |
+
"train_loss": 3.376089767252604,
|
| 4 |
+
"train_runtime": 158003.2062,
|
| 5 |
+
"train_samples": 780934,
|
| 6 |
+
"train_samples_per_second": 243.033,
|
| 7 |
+
"train_steps_per_second": 15.19
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3328 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 3.23770809173584,
|
| 3 |
+
"best_model_checkpoint": "./model_tweets_2020_Q1_90/checkpoint-128000",
|
| 4 |
+
"epoch": 49.171259398881354,
|
| 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": 0.16,
|
| 13 |
+
"eval_loss": 3.4494731426239014,
|
| 14 |
+
"eval_runtime": 46.3964,
|
| 15 |
+
"eval_samples_per_second": 885.888,
|
| 16 |
+
"eval_steps_per_second": 55.371,
|
| 17 |
+
"step": 8000
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.33,
|
| 21 |
+
"learning_rate": 9.939131159843243e-06,
|
| 22 |
+
"loss": 3.5684,
|
| 23 |
+
"step": 16000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.33,
|
| 27 |
+
"eval_loss": 3.416630744934082,
|
| 28 |
+
"eval_runtime": 46.3565,
|
| 29 |
+
"eval_samples_per_second": 886.65,
|
| 30 |
+
"eval_steps_per_second": 55.418,
|
| 31 |
+
"step": 16000
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.49,
|
| 35 |
+
"eval_loss": 3.3847219944000244,
|
| 36 |
+
"eval_runtime": 47.2297,
|
| 37 |
+
"eval_samples_per_second": 870.258,
|
| 38 |
+
"eval_steps_per_second": 54.394,
|
| 39 |
+
"step": 24000
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.66,
|
| 43 |
+
"learning_rate": 9.872425581589261e-06,
|
| 44 |
+
"loss": 3.3755,
|
| 45 |
+
"step": 32000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.66,
|
| 49 |
+
"eval_loss": 3.3664660453796387,
|
| 50 |
+
"eval_runtime": 47.0495,
|
| 51 |
+
"eval_samples_per_second": 873.591,
|
| 52 |
+
"eval_steps_per_second": 54.602,
|
| 53 |
+
"step": 32000
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.82,
|
| 57 |
+
"eval_loss": 3.3654134273529053,
|
| 58 |
+
"eval_runtime": 46.3932,
|
| 59 |
+
"eval_samples_per_second": 885.949,
|
| 60 |
+
"eval_steps_per_second": 55.374,
|
| 61 |
+
"step": 40000
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.98,
|
| 65 |
+
"learning_rate": 9.80572000333528e-06,
|
| 66 |
+
"loss": 3.3533,
|
| 67 |
+
"step": 48000
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 0.98,
|
| 71 |
+
"eval_loss": 3.3654167652130127,
|
| 72 |
+
"eval_runtime": 46.5322,
|
| 73 |
+
"eval_samples_per_second": 883.301,
|
| 74 |
+
"eval_steps_per_second": 55.209,
|
| 75 |
+
"step": 48000
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 1.15,
|
| 79 |
+
"eval_loss": 3.332759380340576,
|
| 80 |
+
"eval_runtime": 46.4492,
|
| 81 |
+
"eval_samples_per_second": 884.88,
|
| 82 |
+
"eval_steps_per_second": 55.308,
|
| 83 |
+
"step": 56000
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 1.31,
|
| 87 |
+
"learning_rate": 9.739014425081299e-06,
|
| 88 |
+
"loss": 3.3014,
|
| 89 |
+
"step": 64000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.31,
|
| 93 |
+
"eval_loss": 3.3209590911865234,
|
| 94 |
+
"eval_runtime": 45.8973,
|
| 95 |
+
"eval_samples_per_second": 895.521,
|
| 96 |
+
"eval_steps_per_second": 55.973,
|
| 97 |
+
"step": 64000
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 1.48,
|
| 101 |
+
"eval_loss": 3.3491690158843994,
|
| 102 |
+
"eval_runtime": 46.3252,
|
| 103 |
+
"eval_samples_per_second": 887.249,
|
| 104 |
+
"eval_steps_per_second": 55.456,
|
| 105 |
+
"step": 72000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.64,
|
| 109 |
+
"learning_rate": 9.672308846827316e-06,
|
| 110 |
+
"loss": 3.2888,
|
| 111 |
+
"step": 80000
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 1.64,
|
| 115 |
+
"eval_loss": 3.3213465213775635,
|
| 116 |
+
"eval_runtime": 45.915,
|
| 117 |
+
"eval_samples_per_second": 895.177,
|
| 118 |
+
"eval_steps_per_second": 55.951,
|
| 119 |
+
"step": 80000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.8,
|
| 123 |
+
"eval_loss": 3.2708065509796143,
|
| 124 |
+
"eval_runtime": 45.9723,
|
| 125 |
+
"eval_samples_per_second": 894.061,
|
| 126 |
+
"eval_steps_per_second": 55.882,
|
| 127 |
+
"step": 88000
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 1.97,
|
| 131 |
+
"learning_rate": 9.605603268573334e-06,
|
| 132 |
+
"loss": 3.2609,
|
| 133 |
+
"step": 96000
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 1.97,
|
| 137 |
+
"eval_loss": 3.290764808654785,
|
| 138 |
+
"eval_runtime": 46.6916,
|
| 139 |
+
"eval_samples_per_second": 880.287,
|
| 140 |
+
"eval_steps_per_second": 55.021,
|
| 141 |
+
"step": 96000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 2.13,
|
| 145 |
+
"eval_loss": 3.2766778469085693,
|
| 146 |
+
"eval_runtime": 45.6527,
|
| 147 |
+
"eval_samples_per_second": 900.318,
|
| 148 |
+
"eval_steps_per_second": 56.273,
|
| 149 |
+
"step": 104000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 2.29,
|
| 153 |
+
"learning_rate": 9.538897690319354e-06,
|
| 154 |
+
"loss": 3.2159,
|
| 155 |
+
"step": 112000
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 2.29,
|
| 159 |
+
"eval_loss": 3.259241819381714,
|
| 160 |
+
"eval_runtime": 45.9077,
|
| 161 |
+
"eval_samples_per_second": 895.319,
|
| 162 |
+
"eval_steps_per_second": 55.96,
|
| 163 |
+
"step": 112000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 2.46,
|
| 167 |
+
"eval_loss": 3.2411258220672607,
|
| 168 |
+
"eval_runtime": 46.8974,
|
| 169 |
+
"eval_samples_per_second": 876.424,
|
| 170 |
+
"eval_steps_per_second": 54.779,
|
| 171 |
+
"step": 120000
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 2.62,
|
| 175 |
+
"learning_rate": 9.472192112065373e-06,
|
| 176 |
+
"loss": 3.2167,
|
| 177 |
+
"step": 128000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 2.62,
|
| 181 |
+
"eval_loss": 3.23770809173584,
|
| 182 |
+
"eval_runtime": 46.0285,
|
| 183 |
+
"eval_samples_per_second": 892.969,
|
| 184 |
+
"eval_steps_per_second": 55.813,
|
| 185 |
+
"step": 128000
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.79,
|
| 189 |
+
"eval_loss": 3.2485291957855225,
|
| 190 |
+
"eval_runtime": 46.313,
|
| 191 |
+
"eval_samples_per_second": 887.483,
|
| 192 |
+
"eval_steps_per_second": 55.47,
|
| 193 |
+
"step": 136000
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 2.95,
|
| 197 |
+
"learning_rate": 9.405486533811392e-06,
|
| 198 |
+
"loss": 3.199,
|
| 199 |
+
"step": 144000
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 2.95,
|
| 203 |
+
"eval_loss": 3.2608513832092285,
|
| 204 |
+
"eval_runtime": 46.3737,
|
| 205 |
+
"eval_samples_per_second": 886.322,
|
| 206 |
+
"eval_steps_per_second": 55.398,
|
| 207 |
+
"step": 144000
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 3.11,
|
| 211 |
+
"eval_loss": 3.2552711963653564,
|
| 212 |
+
"eval_runtime": 45.7073,
|
| 213 |
+
"eval_samples_per_second": 899.243,
|
| 214 |
+
"eval_steps_per_second": 56.205,
|
| 215 |
+
"step": 152000
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 3.28,
|
| 219 |
+
"learning_rate": 9.338780955557409e-06,
|
| 220 |
+
"loss": 3.1905,
|
| 221 |
+
"step": 160000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 3.28,
|
| 225 |
+
"eval_loss": 3.2425193786621094,
|
| 226 |
+
"eval_runtime": 46.3189,
|
| 227 |
+
"eval_samples_per_second": 887.37,
|
| 228 |
+
"eval_steps_per_second": 55.463,
|
| 229 |
+
"step": 160000
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 3.44,
|
| 233 |
+
"eval_loss": 3.2421696186065674,
|
| 234 |
+
"eval_runtime": 46.3489,
|
| 235 |
+
"eval_samples_per_second": 886.796,
|
| 236 |
+
"eval_steps_per_second": 55.427,
|
| 237 |
+
"step": 168000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 3.61,
|
| 241 |
+
"learning_rate": 9.272075377303427e-06,
|
| 242 |
+
"loss": 3.1822,
|
| 243 |
+
"step": 176000
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 3.61,
|
| 247 |
+
"eval_loss": 3.262392997741699,
|
| 248 |
+
"eval_runtime": 46.6763,
|
| 249 |
+
"eval_samples_per_second": 880.575,
|
| 250 |
+
"eval_steps_per_second": 55.039,
|
| 251 |
+
"step": 176000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 3.77,
|
| 255 |
+
"eval_loss": 3.2507119178771973,
|
| 256 |
+
"eval_runtime": 46.8277,
|
| 257 |
+
"eval_samples_per_second": 877.728,
|
| 258 |
+
"eval_steps_per_second": 54.861,
|
| 259 |
+
"step": 184000
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 3.93,
|
| 263 |
+
"learning_rate": 9.205369799049446e-06,
|
| 264 |
+
"loss": 3.1852,
|
| 265 |
+
"step": 192000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 3.93,
|
| 269 |
+
"eval_loss": 3.2483315467834473,
|
| 270 |
+
"eval_runtime": 45.7607,
|
| 271 |
+
"eval_samples_per_second": 898.195,
|
| 272 |
+
"eval_steps_per_second": 56.14,
|
| 273 |
+
"step": 192000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 4.1,
|
| 277 |
+
"eval_loss": 3.251424789428711,
|
| 278 |
+
"eval_runtime": 46.3642,
|
| 279 |
+
"eval_samples_per_second": 886.503,
|
| 280 |
+
"eval_steps_per_second": 55.409,
|
| 281 |
+
"step": 200000
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 4.26,
|
| 285 |
+
"learning_rate": 9.138664220795464e-06,
|
| 286 |
+
"loss": 3.1767,
|
| 287 |
+
"step": 208000
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 4.26,
|
| 291 |
+
"eval_loss": 3.242562770843506,
|
| 292 |
+
"eval_runtime": 46.886,
|
| 293 |
+
"eval_samples_per_second": 876.637,
|
| 294 |
+
"eval_steps_per_second": 54.792,
|
| 295 |
+
"step": 208000
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 4.43,
|
| 299 |
+
"eval_loss": 3.234778642654419,
|
| 300 |
+
"eval_runtime": 46.4949,
|
| 301 |
+
"eval_samples_per_second": 884.01,
|
| 302 |
+
"eval_steps_per_second": 55.253,
|
| 303 |
+
"step": 216000
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 4.59,
|
| 307 |
+
"learning_rate": 9.071958642541483e-06,
|
| 308 |
+
"loss": 3.1767,
|
| 309 |
+
"step": 224000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 4.59,
|
| 313 |
+
"eval_loss": 3.2734625339508057,
|
| 314 |
+
"eval_runtime": 46.0486,
|
| 315 |
+
"eval_samples_per_second": 892.58,
|
| 316 |
+
"eval_steps_per_second": 55.789,
|
| 317 |
+
"step": 224000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 4.75,
|
| 321 |
+
"eval_loss": 3.2471694946289062,
|
| 322 |
+
"eval_runtime": 46.5054,
|
| 323 |
+
"eval_samples_per_second": 883.811,
|
| 324 |
+
"eval_steps_per_second": 55.241,
|
| 325 |
+
"step": 232000
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 4.92,
|
| 329 |
+
"learning_rate": 9.005253064287502e-06,
|
| 330 |
+
"loss": 3.1973,
|
| 331 |
+
"step": 240000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 4.92,
|
| 335 |
+
"eval_loss": 3.259644031524658,
|
| 336 |
+
"eval_runtime": 45.8405,
|
| 337 |
+
"eval_samples_per_second": 896.631,
|
| 338 |
+
"eval_steps_per_second": 56.042,
|
| 339 |
+
"step": 240000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 5.08,
|
| 343 |
+
"eval_loss": 3.2605602741241455,
|
| 344 |
+
"eval_runtime": 45.4485,
|
| 345 |
+
"eval_samples_per_second": 904.365,
|
| 346 |
+
"eval_steps_per_second": 56.526,
|
| 347 |
+
"step": 248000
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 5.24,
|
| 351 |
+
"learning_rate": 8.93854748603352e-06,
|
| 352 |
+
"loss": 3.1781,
|
| 353 |
+
"step": 256000
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 5.24,
|
| 357 |
+
"eval_loss": 3.281527519226074,
|
| 358 |
+
"eval_runtime": 46.3336,
|
| 359 |
+
"eval_samples_per_second": 887.089,
|
| 360 |
+
"eval_steps_per_second": 55.446,
|
| 361 |
+
"step": 256000
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 5.41,
|
| 365 |
+
"eval_loss": 3.273421049118042,
|
| 366 |
+
"eval_runtime": 45.4558,
|
| 367 |
+
"eval_samples_per_second": 904.218,
|
| 368 |
+
"eval_steps_per_second": 56.516,
|
| 369 |
+
"step": 264000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 5.57,
|
| 373 |
+
"learning_rate": 8.871841907779539e-06,
|
| 374 |
+
"loss": 3.1803,
|
| 375 |
+
"step": 272000
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 5.57,
|
| 379 |
+
"eval_loss": 3.2739477157592773,
|
| 380 |
+
"eval_runtime": 45.6455,
|
| 381 |
+
"eval_samples_per_second": 900.462,
|
| 382 |
+
"eval_steps_per_second": 56.282,
|
| 383 |
+
"step": 272000
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 5.74,
|
| 387 |
+
"eval_loss": 3.2712481021881104,
|
| 388 |
+
"eval_runtime": 46.5973,
|
| 389 |
+
"eval_samples_per_second": 882.068,
|
| 390 |
+
"eval_steps_per_second": 55.132,
|
| 391 |
+
"step": 280000
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 5.9,
|
| 395 |
+
"learning_rate": 8.805136329525557e-06,
|
| 396 |
+
"loss": 3.1989,
|
| 397 |
+
"step": 288000
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 5.9,
|
| 401 |
+
"eval_loss": 3.273439884185791,
|
| 402 |
+
"eval_runtime": 46.318,
|
| 403 |
+
"eval_samples_per_second": 887.387,
|
| 404 |
+
"eval_steps_per_second": 55.464,
|
| 405 |
+
"step": 288000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.06,
|
| 409 |
+
"eval_loss": 3.293893814086914,
|
| 410 |
+
"eval_runtime": 45.8003,
|
| 411 |
+
"eval_samples_per_second": 897.418,
|
| 412 |
+
"eval_steps_per_second": 56.091,
|
| 413 |
+
"step": 296000
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 6.23,
|
| 417 |
+
"learning_rate": 8.738430751271576e-06,
|
| 418 |
+
"loss": 3.1929,
|
| 419 |
+
"step": 304000
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.23,
|
| 423 |
+
"eval_loss": 3.288043737411499,
|
| 424 |
+
"eval_runtime": 46.6462,
|
| 425 |
+
"eval_samples_per_second": 881.144,
|
| 426 |
+
"eval_steps_per_second": 55.074,
|
| 427 |
+
"step": 304000
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 6.39,
|
| 431 |
+
"eval_loss": 3.289358139038086,
|
| 432 |
+
"eval_runtime": 45.8512,
|
| 433 |
+
"eval_samples_per_second": 896.422,
|
| 434 |
+
"eval_steps_per_second": 56.029,
|
| 435 |
+
"step": 312000
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 6.56,
|
| 439 |
+
"learning_rate": 8.671725173017595e-06,
|
| 440 |
+
"loss": 3.2083,
|
| 441 |
+
"step": 320000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 6.56,
|
| 445 |
+
"eval_loss": 3.308645725250244,
|
| 446 |
+
"eval_runtime": 46.2317,
|
| 447 |
+
"eval_samples_per_second": 889.043,
|
| 448 |
+
"eval_steps_per_second": 55.568,
|
| 449 |
+
"step": 320000
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 6.72,
|
| 453 |
+
"eval_loss": 3.3066623210906982,
|
| 454 |
+
"eval_runtime": 46.8669,
|
| 455 |
+
"eval_samples_per_second": 876.995,
|
| 456 |
+
"eval_steps_per_second": 54.815,
|
| 457 |
+
"step": 328000
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 6.88,
|
| 461 |
+
"learning_rate": 8.605019594763613e-06,
|
| 462 |
+
"loss": 3.2013,
|
| 463 |
+
"step": 336000
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 6.88,
|
| 467 |
+
"eval_loss": 3.278655529022217,
|
| 468 |
+
"eval_runtime": 45.904,
|
| 469 |
+
"eval_samples_per_second": 895.391,
|
| 470 |
+
"eval_steps_per_second": 55.965,
|
| 471 |
+
"step": 336000
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 7.05,
|
| 475 |
+
"eval_loss": 3.3152964115142822,
|
| 476 |
+
"eval_runtime": 46.5312,
|
| 477 |
+
"eval_samples_per_second": 883.322,
|
| 478 |
+
"eval_steps_per_second": 55.21,
|
| 479 |
+
"step": 344000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.21,
|
| 483 |
+
"learning_rate": 8.538314016509632e-06,
|
| 484 |
+
"loss": 3.2111,
|
| 485 |
+
"step": 352000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 7.21,
|
| 489 |
+
"eval_loss": 3.3246278762817383,
|
| 490 |
+
"eval_runtime": 46.7247,
|
| 491 |
+
"eval_samples_per_second": 879.664,
|
| 492 |
+
"eval_steps_per_second": 54.982,
|
| 493 |
+
"step": 352000
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.38,
|
| 497 |
+
"eval_loss": 3.3322579860687256,
|
| 498 |
+
"eval_runtime": 45.9989,
|
| 499 |
+
"eval_samples_per_second": 893.543,
|
| 500 |
+
"eval_steps_per_second": 55.849,
|
| 501 |
+
"step": 360000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 7.54,
|
| 505 |
+
"learning_rate": 8.471608438255649e-06,
|
| 506 |
+
"loss": 3.2186,
|
| 507 |
+
"step": 368000
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.54,
|
| 511 |
+
"eval_loss": 3.2938337326049805,
|
| 512 |
+
"eval_runtime": 46.144,
|
| 513 |
+
"eval_samples_per_second": 890.734,
|
| 514 |
+
"eval_steps_per_second": 55.674,
|
| 515 |
+
"step": 368000
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 7.7,
|
| 519 |
+
"eval_loss": 3.3499817848205566,
|
| 520 |
+
"eval_runtime": 45.582,
|
| 521 |
+
"eval_samples_per_second": 901.717,
|
| 522 |
+
"eval_steps_per_second": 56.36,
|
| 523 |
+
"step": 376000
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 7.87,
|
| 527 |
+
"learning_rate": 8.404902860001667e-06,
|
| 528 |
+
"loss": 3.2268,
|
| 529 |
+
"step": 384000
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 7.87,
|
| 533 |
+
"eval_loss": 3.3179759979248047,
|
| 534 |
+
"eval_runtime": 45.2091,
|
| 535 |
+
"eval_samples_per_second": 909.153,
|
| 536 |
+
"eval_steps_per_second": 56.825,
|
| 537 |
+
"step": 384000
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 8.03,
|
| 541 |
+
"eval_loss": 3.3171069622039795,
|
| 542 |
+
"eval_runtime": 46.0196,
|
| 543 |
+
"eval_samples_per_second": 893.141,
|
| 544 |
+
"eval_steps_per_second": 55.824,
|
| 545 |
+
"step": 392000
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 8.2,
|
| 549 |
+
"learning_rate": 8.338197281747686e-06,
|
| 550 |
+
"loss": 3.233,
|
| 551 |
+
"step": 400000
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 8.2,
|
| 555 |
+
"eval_loss": 3.3461642265319824,
|
| 556 |
+
"eval_runtime": 45.5487,
|
| 557 |
+
"eval_samples_per_second": 902.375,
|
| 558 |
+
"eval_steps_per_second": 56.401,
|
| 559 |
+
"step": 400000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 8.36,
|
| 563 |
+
"eval_loss": 3.341256618499756,
|
| 564 |
+
"eval_runtime": 45.4264,
|
| 565 |
+
"eval_samples_per_second": 904.804,
|
| 566 |
+
"eval_steps_per_second": 56.553,
|
| 567 |
+
"step": 408000
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.52,
|
| 571 |
+
"learning_rate": 8.271491703493705e-06,
|
| 572 |
+
"loss": 3.2432,
|
| 573 |
+
"step": 416000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 8.52,
|
| 577 |
+
"eval_loss": 3.328122615814209,
|
| 578 |
+
"eval_runtime": 45.9787,
|
| 579 |
+
"eval_samples_per_second": 893.936,
|
| 580 |
+
"eval_steps_per_second": 55.874,
|
| 581 |
+
"step": 416000
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 8.69,
|
| 585 |
+
"eval_loss": 3.342041492462158,
|
| 586 |
+
"eval_runtime": 45.4274,
|
| 587 |
+
"eval_samples_per_second": 904.784,
|
| 588 |
+
"eval_steps_per_second": 56.552,
|
| 589 |
+
"step": 424000
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 8.85,
|
| 593 |
+
"learning_rate": 8.204786125239725e-06,
|
| 594 |
+
"loss": 3.2586,
|
| 595 |
+
"step": 432000
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 8.85,
|
| 599 |
+
"eval_loss": 3.3609066009521484,
|
| 600 |
+
"eval_runtime": 45.3913,
|
| 601 |
+
"eval_samples_per_second": 905.504,
|
| 602 |
+
"eval_steps_per_second": 56.597,
|
| 603 |
+
"step": 432000
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 9.01,
|
| 607 |
+
"eval_loss": 3.352691173553467,
|
| 608 |
+
"eval_runtime": 46.0515,
|
| 609 |
+
"eval_samples_per_second": 892.522,
|
| 610 |
+
"eval_steps_per_second": 55.785,
|
| 611 |
+
"step": 440000
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 9.18,
|
| 615 |
+
"learning_rate": 8.138080546985743e-06,
|
| 616 |
+
"loss": 3.2567,
|
| 617 |
+
"step": 448000
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 9.18,
|
| 621 |
+
"eval_loss": 3.359393358230591,
|
| 622 |
+
"eval_runtime": 45.57,
|
| 623 |
+
"eval_samples_per_second": 901.953,
|
| 624 |
+
"eval_steps_per_second": 56.375,
|
| 625 |
+
"step": 448000
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 9.34,
|
| 629 |
+
"eval_loss": 3.3497443199157715,
|
| 630 |
+
"eval_runtime": 45.4208,
|
| 631 |
+
"eval_samples_per_second": 904.915,
|
| 632 |
+
"eval_steps_per_second": 56.56,
|
| 633 |
+
"step": 456000
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 9.51,
|
| 637 |
+
"learning_rate": 8.07137496873176e-06,
|
| 638 |
+
"loss": 3.2592,
|
| 639 |
+
"step": 464000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 9.51,
|
| 643 |
+
"eval_loss": 3.3606550693511963,
|
| 644 |
+
"eval_runtime": 46.15,
|
| 645 |
+
"eval_samples_per_second": 890.617,
|
| 646 |
+
"eval_steps_per_second": 55.666,
|
| 647 |
+
"step": 464000
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 9.67,
|
| 651 |
+
"eval_loss": 3.3839540481567383,
|
| 652 |
+
"eval_runtime": 45.5702,
|
| 653 |
+
"eval_samples_per_second": 901.95,
|
| 654 |
+
"eval_steps_per_second": 56.375,
|
| 655 |
+
"step": 472000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 9.83,
|
| 659 |
+
"learning_rate": 8.004669390477779e-06,
|
| 660 |
+
"loss": 3.2793,
|
| 661 |
+
"step": 480000
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 9.83,
|
| 665 |
+
"eval_loss": 3.366785764694214,
|
| 666 |
+
"eval_runtime": 45.749,
|
| 667 |
+
"eval_samples_per_second": 898.424,
|
| 668 |
+
"eval_steps_per_second": 56.154,
|
| 669 |
+
"step": 480000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 10.0,
|
| 673 |
+
"eval_loss": 3.3609416484832764,
|
| 674 |
+
"eval_runtime": 47.1383,
|
| 675 |
+
"eval_samples_per_second": 871.945,
|
| 676 |
+
"eval_steps_per_second": 54.499,
|
| 677 |
+
"step": 488000
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 10.16,
|
| 681 |
+
"learning_rate": 7.937963812223798e-06,
|
| 682 |
+
"loss": 3.257,
|
| 683 |
+
"step": 496000
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 10.16,
|
| 687 |
+
"eval_loss": 3.368229389190674,
|
| 688 |
+
"eval_runtime": 45.5778,
|
| 689 |
+
"eval_samples_per_second": 901.798,
|
| 690 |
+
"eval_steps_per_second": 56.365,
|
| 691 |
+
"step": 496000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 10.33,
|
| 695 |
+
"eval_loss": 3.4005918502807617,
|
| 696 |
+
"eval_runtime": 46.5843,
|
| 697 |
+
"eval_samples_per_second": 882.314,
|
| 698 |
+
"eval_steps_per_second": 55.147,
|
| 699 |
+
"step": 504000
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 10.49,
|
| 703 |
+
"learning_rate": 7.871258233969816e-06,
|
| 704 |
+
"loss": 3.2656,
|
| 705 |
+
"step": 512000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 10.49,
|
| 709 |
+
"eval_loss": 3.358835220336914,
|
| 710 |
+
"eval_runtime": 46.2545,
|
| 711 |
+
"eval_samples_per_second": 888.605,
|
| 712 |
+
"eval_steps_per_second": 55.541,
|
| 713 |
+
"step": 512000
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 10.65,
|
| 717 |
+
"eval_loss": 3.379861831665039,
|
| 718 |
+
"eval_runtime": 45.613,
|
| 719 |
+
"eval_samples_per_second": 901.103,
|
| 720 |
+
"eval_steps_per_second": 56.322,
|
| 721 |
+
"step": 520000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 10.82,
|
| 725 |
+
"learning_rate": 7.804552655715835e-06,
|
| 726 |
+
"loss": 3.2727,
|
| 727 |
+
"step": 528000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 10.82,
|
| 731 |
+
"eval_loss": 3.383315086364746,
|
| 732 |
+
"eval_runtime": 46.0041,
|
| 733 |
+
"eval_samples_per_second": 893.442,
|
| 734 |
+
"eval_steps_per_second": 55.843,
|
| 735 |
+
"step": 528000
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 10.98,
|
| 739 |
+
"eval_loss": 3.356590747833252,
|
| 740 |
+
"eval_runtime": 45.9202,
|
| 741 |
+
"eval_samples_per_second": 895.074,
|
| 742 |
+
"eval_steps_per_second": 55.945,
|
| 743 |
+
"step": 536000
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 11.15,
|
| 747 |
+
"learning_rate": 7.737847077461853e-06,
|
| 748 |
+
"loss": 3.2705,
|
| 749 |
+
"step": 544000
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 11.15,
|
| 753 |
+
"eval_loss": 3.3793959617614746,
|
| 754 |
+
"eval_runtime": 45.6075,
|
| 755 |
+
"eval_samples_per_second": 901.211,
|
| 756 |
+
"eval_steps_per_second": 56.328,
|
| 757 |
+
"step": 544000
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 11.31,
|
| 761 |
+
"eval_loss": 3.3838233947753906,
|
| 762 |
+
"eval_runtime": 46.1859,
|
| 763 |
+
"eval_samples_per_second": 889.925,
|
| 764 |
+
"eval_steps_per_second": 55.623,
|
| 765 |
+
"step": 552000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 11.47,
|
| 769 |
+
"learning_rate": 7.671141499207872e-06,
|
| 770 |
+
"loss": 3.2676,
|
| 771 |
+
"step": 560000
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 11.47,
|
| 775 |
+
"eval_loss": 3.3659656047821045,
|
| 776 |
+
"eval_runtime": 45.7183,
|
| 777 |
+
"eval_samples_per_second": 899.027,
|
| 778 |
+
"eval_steps_per_second": 56.192,
|
| 779 |
+
"step": 560000
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 11.64,
|
| 783 |
+
"eval_loss": 3.3937699794769287,
|
| 784 |
+
"eval_runtime": 45.9326,
|
| 785 |
+
"eval_samples_per_second": 894.832,
|
| 786 |
+
"eval_steps_per_second": 55.93,
|
| 787 |
+
"step": 568000
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 11.8,
|
| 791 |
+
"learning_rate": 7.604435920953891e-06,
|
| 792 |
+
"loss": 3.258,
|
| 793 |
+
"step": 576000
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 11.8,
|
| 797 |
+
"eval_loss": 3.3661420345306396,
|
| 798 |
+
"eval_runtime": 46.4625,
|
| 799 |
+
"eval_samples_per_second": 884.627,
|
| 800 |
+
"eval_steps_per_second": 55.292,
|
| 801 |
+
"step": 576000
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 11.97,
|
| 805 |
+
"eval_loss": 3.3490447998046875,
|
| 806 |
+
"eval_runtime": 45.8318,
|
| 807 |
+
"eval_samples_per_second": 896.801,
|
| 808 |
+
"eval_steps_per_second": 56.053,
|
| 809 |
+
"step": 584000
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 12.13,
|
| 813 |
+
"learning_rate": 7.537730342699909e-06,
|
| 814 |
+
"loss": 3.2646,
|
| 815 |
+
"step": 592000
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 12.13,
|
| 819 |
+
"eval_loss": 3.3716230392456055,
|
| 820 |
+
"eval_runtime": 45.6734,
|
| 821 |
+
"eval_samples_per_second": 899.91,
|
| 822 |
+
"eval_steps_per_second": 56.247,
|
| 823 |
+
"step": 592000
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 12.29,
|
| 827 |
+
"eval_loss": 3.3877346515655518,
|
| 828 |
+
"eval_runtime": 46.2161,
|
| 829 |
+
"eval_samples_per_second": 889.344,
|
| 830 |
+
"eval_steps_per_second": 55.587,
|
| 831 |
+
"step": 600000
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 12.46,
|
| 835 |
+
"learning_rate": 7.471024764445928e-06,
|
| 836 |
+
"loss": 3.2578,
|
| 837 |
+
"step": 608000
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 12.46,
|
| 841 |
+
"eval_loss": 3.3930206298828125,
|
| 842 |
+
"eval_runtime": 45.3985,
|
| 843 |
+
"eval_samples_per_second": 905.361,
|
| 844 |
+
"eval_steps_per_second": 56.588,
|
| 845 |
+
"step": 608000
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 12.62,
|
| 849 |
+
"eval_loss": 3.392077922821045,
|
| 850 |
+
"eval_runtime": 45.1724,
|
| 851 |
+
"eval_samples_per_second": 909.893,
|
| 852 |
+
"eval_steps_per_second": 56.871,
|
| 853 |
+
"step": 616000
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 12.78,
|
| 857 |
+
"learning_rate": 7.4043191861919465e-06,
|
| 858 |
+
"loss": 3.2719,
|
| 859 |
+
"step": 624000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 12.78,
|
| 863 |
+
"eval_loss": 3.395730495452881,
|
| 864 |
+
"eval_runtime": 45.8195,
|
| 865 |
+
"eval_samples_per_second": 897.042,
|
| 866 |
+
"eval_steps_per_second": 56.068,
|
| 867 |
+
"step": 624000
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 12.95,
|
| 871 |
+
"eval_loss": 3.4196434020996094,
|
| 872 |
+
"eval_runtime": 45.2614,
|
| 873 |
+
"eval_samples_per_second": 908.103,
|
| 874 |
+
"eval_steps_per_second": 56.759,
|
| 875 |
+
"step": 632000
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 13.11,
|
| 879 |
+
"learning_rate": 7.337613607937964e-06,
|
| 880 |
+
"loss": 3.2828,
|
| 881 |
+
"step": 640000
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 13.11,
|
| 885 |
+
"eval_loss": 3.4077515602111816,
|
| 886 |
+
"eval_runtime": 45.5674,
|
| 887 |
+
"eval_samples_per_second": 902.004,
|
| 888 |
+
"eval_steps_per_second": 56.378,
|
| 889 |
+
"step": 640000
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 13.28,
|
| 893 |
+
"eval_loss": 3.4202864170074463,
|
| 894 |
+
"eval_runtime": 46.3249,
|
| 895 |
+
"eval_samples_per_second": 887.255,
|
| 896 |
+
"eval_steps_per_second": 55.456,
|
| 897 |
+
"step": 648000
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 13.44,
|
| 901 |
+
"learning_rate": 7.270908029683983e-06,
|
| 902 |
+
"loss": 3.2805,
|
| 903 |
+
"step": 656000
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 13.44,
|
| 907 |
+
"eval_loss": 3.3899548053741455,
|
| 908 |
+
"eval_runtime": 46.1588,
|
| 909 |
+
"eval_samples_per_second": 890.448,
|
| 910 |
+
"eval_steps_per_second": 55.656,
|
| 911 |
+
"step": 656000
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 13.6,
|
| 915 |
+
"eval_loss": 3.4037835597991943,
|
| 916 |
+
"eval_runtime": 46.9454,
|
| 917 |
+
"eval_samples_per_second": 875.527,
|
| 918 |
+
"eval_steps_per_second": 54.723,
|
| 919 |
+
"step": 664000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 13.77,
|
| 923 |
+
"learning_rate": 7.2042024514300015e-06,
|
| 924 |
+
"loss": 3.2975,
|
| 925 |
+
"step": 672000
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 13.77,
|
| 929 |
+
"eval_loss": 3.405585765838623,
|
| 930 |
+
"eval_runtime": 46.2706,
|
| 931 |
+
"eval_samples_per_second": 888.297,
|
| 932 |
+
"eval_steps_per_second": 55.521,
|
| 933 |
+
"step": 672000
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 13.93,
|
| 937 |
+
"eval_loss": 3.428373336791992,
|
| 938 |
+
"eval_runtime": 45.9889,
|
| 939 |
+
"eval_samples_per_second": 893.738,
|
| 940 |
+
"eval_steps_per_second": 55.861,
|
| 941 |
+
"step": 680000
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 14.1,
|
| 945 |
+
"learning_rate": 7.13749687317602e-06,
|
| 946 |
+
"loss": 3.2965,
|
| 947 |
+
"step": 688000
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 14.1,
|
| 951 |
+
"eval_loss": 3.41803240776062,
|
| 952 |
+
"eval_runtime": 46.9126,
|
| 953 |
+
"eval_samples_per_second": 876.14,
|
| 954 |
+
"eval_steps_per_second": 54.761,
|
| 955 |
+
"step": 688000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 14.26,
|
| 959 |
+
"eval_loss": 3.419599771499634,
|
| 960 |
+
"eval_runtime": 46.1796,
|
| 961 |
+
"eval_samples_per_second": 890.047,
|
| 962 |
+
"eval_steps_per_second": 55.631,
|
| 963 |
+
"step": 696000
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 14.42,
|
| 967 |
+
"learning_rate": 7.070791294922038e-06,
|
| 968 |
+
"loss": 3.3069,
|
| 969 |
+
"step": 704000
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 14.42,
|
| 973 |
+
"eval_loss": 3.425711154937744,
|
| 974 |
+
"eval_runtime": 46.2298,
|
| 975 |
+
"eval_samples_per_second": 889.08,
|
| 976 |
+
"eval_steps_per_second": 55.57,
|
| 977 |
+
"step": 704000
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 14.59,
|
| 981 |
+
"eval_loss": 3.4299447536468506,
|
| 982 |
+
"eval_runtime": 46.768,
|
| 983 |
+
"eval_samples_per_second": 878.85,
|
| 984 |
+
"eval_steps_per_second": 54.931,
|
| 985 |
+
"step": 712000
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 14.75,
|
| 989 |
+
"learning_rate": 7.0040857166680564e-06,
|
| 990 |
+
"loss": 3.3152,
|
| 991 |
+
"step": 720000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 14.75,
|
| 995 |
+
"eval_loss": 3.4787514209747314,
|
| 996 |
+
"eval_runtime": 46.0913,
|
| 997 |
+
"eval_samples_per_second": 891.752,
|
| 998 |
+
"eval_steps_per_second": 55.737,
|
| 999 |
+
"step": 720000
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 14.92,
|
| 1003 |
+
"eval_loss": 3.4424662590026855,
|
| 1004 |
+
"eval_runtime": 46.3411,
|
| 1005 |
+
"eval_samples_per_second": 886.945,
|
| 1006 |
+
"eval_steps_per_second": 55.437,
|
| 1007 |
+
"step": 728000
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 15.08,
|
| 1011 |
+
"learning_rate": 6.937380138414076e-06,
|
| 1012 |
+
"loss": 3.3125,
|
| 1013 |
+
"step": 736000
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 15.08,
|
| 1017 |
+
"eval_loss": 3.430126667022705,
|
| 1018 |
+
"eval_runtime": 46.9882,
|
| 1019 |
+
"eval_samples_per_second": 874.73,
|
| 1020 |
+
"eval_steps_per_second": 54.673,
|
| 1021 |
+
"step": 736000
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 15.24,
|
| 1025 |
+
"eval_loss": 3.4440979957580566,
|
| 1026 |
+
"eval_runtime": 46.1825,
|
| 1027 |
+
"eval_samples_per_second": 889.99,
|
| 1028 |
+
"eval_steps_per_second": 55.627,
|
| 1029 |
+
"step": 744000
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 15.41,
|
| 1033 |
+
"learning_rate": 6.8706745601600945e-06,
|
| 1034 |
+
"loss": 3.3174,
|
| 1035 |
+
"step": 752000
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 15.41,
|
| 1039 |
+
"eval_loss": 3.4396116733551025,
|
| 1040 |
+
"eval_runtime": 46.2686,
|
| 1041 |
+
"eval_samples_per_second": 888.334,
|
| 1042 |
+
"eval_steps_per_second": 55.524,
|
| 1043 |
+
"step": 752000
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 15.57,
|
| 1047 |
+
"eval_loss": 3.463931083679199,
|
| 1048 |
+
"eval_runtime": 46.7798,
|
| 1049 |
+
"eval_samples_per_second": 878.627,
|
| 1050 |
+
"eval_steps_per_second": 54.917,
|
| 1051 |
+
"step": 760000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 15.73,
|
| 1055 |
+
"learning_rate": 6.803968981906113e-06,
|
| 1056 |
+
"loss": 3.3242,
|
| 1057 |
+
"step": 768000
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 15.73,
|
| 1061 |
+
"eval_loss": 3.4523837566375732,
|
| 1062 |
+
"eval_runtime": 45.7867,
|
| 1063 |
+
"eval_samples_per_second": 897.685,
|
| 1064 |
+
"eval_steps_per_second": 56.108,
|
| 1065 |
+
"step": 768000
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 15.9,
|
| 1069 |
+
"eval_loss": 3.455958366394043,
|
| 1070 |
+
"eval_runtime": 45.3124,
|
| 1071 |
+
"eval_samples_per_second": 907.08,
|
| 1072 |
+
"eval_steps_per_second": 56.695,
|
| 1073 |
+
"step": 776000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 16.06,
|
| 1077 |
+
"learning_rate": 6.737263403652131e-06,
|
| 1078 |
+
"loss": 3.3385,
|
| 1079 |
+
"step": 784000
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 16.06,
|
| 1083 |
+
"eval_loss": 3.4779999256134033,
|
| 1084 |
+
"eval_runtime": 46.0072,
|
| 1085 |
+
"eval_samples_per_second": 893.383,
|
| 1086 |
+
"eval_steps_per_second": 55.839,
|
| 1087 |
+
"step": 784000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 16.23,
|
| 1091 |
+
"eval_loss": 3.4773714542388916,
|
| 1092 |
+
"eval_runtime": 45.131,
|
| 1093 |
+
"eval_samples_per_second": 910.727,
|
| 1094 |
+
"eval_steps_per_second": 56.923,
|
| 1095 |
+
"step": 792000
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 16.39,
|
| 1099 |
+
"learning_rate": 6.6705578253981495e-06,
|
| 1100 |
+
"loss": 3.3371,
|
| 1101 |
+
"step": 800000
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 16.39,
|
| 1105 |
+
"eval_loss": 3.47719669342041,
|
| 1106 |
+
"eval_runtime": 45.6308,
|
| 1107 |
+
"eval_samples_per_second": 900.751,
|
| 1108 |
+
"eval_steps_per_second": 56.3,
|
| 1109 |
+
"step": 800000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 16.55,
|
| 1113 |
+
"eval_loss": 3.4955241680145264,
|
| 1114 |
+
"eval_runtime": 46.0477,
|
| 1115 |
+
"eval_samples_per_second": 892.597,
|
| 1116 |
+
"eval_steps_per_second": 55.79,
|
| 1117 |
+
"step": 808000
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 16.72,
|
| 1121 |
+
"learning_rate": 6.603852247144168e-06,
|
| 1122 |
+
"loss": 3.3633,
|
| 1123 |
+
"step": 816000
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 16.72,
|
| 1127 |
+
"eval_loss": 3.486057996749878,
|
| 1128 |
+
"eval_runtime": 44.9231,
|
| 1129 |
+
"eval_samples_per_second": 914.941,
|
| 1130 |
+
"eval_steps_per_second": 57.187,
|
| 1131 |
+
"step": 816000
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 16.88,
|
| 1135 |
+
"eval_loss": 3.506316661834717,
|
| 1136 |
+
"eval_runtime": 45.7078,
|
| 1137 |
+
"eval_samples_per_second": 899.234,
|
| 1138 |
+
"eval_steps_per_second": 56.205,
|
| 1139 |
+
"step": 824000
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 17.05,
|
| 1143 |
+
"learning_rate": 6.537146668890187e-06,
|
| 1144 |
+
"loss": 3.3678,
|
| 1145 |
+
"step": 832000
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 17.05,
|
| 1149 |
+
"eval_loss": 3.50439190864563,
|
| 1150 |
+
"eval_runtime": 45.0245,
|
| 1151 |
+
"eval_samples_per_second": 912.882,
|
| 1152 |
+
"eval_steps_per_second": 57.058,
|
| 1153 |
+
"step": 832000
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 17.21,
|
| 1157 |
+
"eval_loss": 3.520247220993042,
|
| 1158 |
+
"eval_runtime": 45.2071,
|
| 1159 |
+
"eval_samples_per_second": 909.193,
|
| 1160 |
+
"eval_steps_per_second": 56.827,
|
| 1161 |
+
"step": 840000
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 17.37,
|
| 1165 |
+
"learning_rate": 6.4704410906362044e-06,
|
| 1166 |
+
"loss": 3.3634,
|
| 1167 |
+
"step": 848000
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 17.37,
|
| 1171 |
+
"eval_loss": 3.4941418170928955,
|
| 1172 |
+
"eval_runtime": 46.4208,
|
| 1173 |
+
"eval_samples_per_second": 885.423,
|
| 1174 |
+
"eval_steps_per_second": 55.342,
|
| 1175 |
+
"step": 848000
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 17.54,
|
| 1179 |
+
"eval_loss": 3.522303819656372,
|
| 1180 |
+
"eval_runtime": 46.164,
|
| 1181 |
+
"eval_samples_per_second": 890.347,
|
| 1182 |
+
"eval_steps_per_second": 55.649,
|
| 1183 |
+
"step": 856000
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 17.7,
|
| 1187 |
+
"learning_rate": 6.403735512382223e-06,
|
| 1188 |
+
"loss": 3.3797,
|
| 1189 |
+
"step": 864000
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 17.7,
|
| 1193 |
+
"eval_loss": 3.502774715423584,
|
| 1194 |
+
"eval_runtime": 45.8285,
|
| 1195 |
+
"eval_samples_per_second": 896.865,
|
| 1196 |
+
"eval_steps_per_second": 56.057,
|
| 1197 |
+
"step": 864000
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 17.87,
|
| 1201 |
+
"eval_loss": 3.526393175125122,
|
| 1202 |
+
"eval_runtime": 46.6422,
|
| 1203 |
+
"eval_samples_per_second": 881.219,
|
| 1204 |
+
"eval_steps_per_second": 55.079,
|
| 1205 |
+
"step": 872000
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 18.03,
|
| 1209 |
+
"learning_rate": 6.337029934128242e-06,
|
| 1210 |
+
"loss": 3.3802,
|
| 1211 |
+
"step": 880000
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 18.03,
|
| 1215 |
+
"eval_loss": 3.531257152557373,
|
| 1216 |
+
"eval_runtime": 46.217,
|
| 1217 |
+
"eval_samples_per_second": 889.327,
|
| 1218 |
+
"eval_steps_per_second": 55.586,
|
| 1219 |
+
"step": 880000
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 18.19,
|
| 1223 |
+
"eval_loss": 3.496319055557251,
|
| 1224 |
+
"eval_runtime": 45.9803,
|
| 1225 |
+
"eval_samples_per_second": 893.904,
|
| 1226 |
+
"eval_steps_per_second": 55.872,
|
| 1227 |
+
"step": 888000
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 18.36,
|
| 1231 |
+
"learning_rate": 6.270324355874261e-06,
|
| 1232 |
+
"loss": 3.357,
|
| 1233 |
+
"step": 896000
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 18.36,
|
| 1237 |
+
"eval_loss": 3.5171141624450684,
|
| 1238 |
+
"eval_runtime": 47.1622,
|
| 1239 |
+
"eval_samples_per_second": 871.504,
|
| 1240 |
+
"eval_steps_per_second": 54.472,
|
| 1241 |
+
"step": 896000
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 18.52,
|
| 1245 |
+
"eval_loss": 3.530701160430908,
|
| 1246 |
+
"eval_runtime": 46.113,
|
| 1247 |
+
"eval_samples_per_second": 891.332,
|
| 1248 |
+
"eval_steps_per_second": 55.711,
|
| 1249 |
+
"step": 904000
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 18.69,
|
| 1253 |
+
"learning_rate": 6.20361877762028e-06,
|
| 1254 |
+
"loss": 3.3866,
|
| 1255 |
+
"step": 912000
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 18.69,
|
| 1259 |
+
"eval_loss": 3.5221967697143555,
|
| 1260 |
+
"eval_runtime": 46.035,
|
| 1261 |
+
"eval_samples_per_second": 892.843,
|
| 1262 |
+
"eval_steps_per_second": 55.805,
|
| 1263 |
+
"step": 912000
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 18.85,
|
| 1267 |
+
"eval_loss": 3.5319056510925293,
|
| 1268 |
+
"eval_runtime": 46.8446,
|
| 1269 |
+
"eval_samples_per_second": 877.412,
|
| 1270 |
+
"eval_steps_per_second": 54.841,
|
| 1271 |
+
"step": 920000
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 19.01,
|
| 1275 |
+
"learning_rate": 6.1369131993662975e-06,
|
| 1276 |
+
"loss": 3.3818,
|
| 1277 |
+
"step": 928000
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 19.01,
|
| 1281 |
+
"eval_loss": 3.532552480697632,
|
| 1282 |
+
"eval_runtime": 46.3901,
|
| 1283 |
+
"eval_samples_per_second": 886.007,
|
| 1284 |
+
"eval_steps_per_second": 55.378,
|
| 1285 |
+
"step": 928000
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 19.18,
|
| 1289 |
+
"eval_loss": 3.5116307735443115,
|
| 1290 |
+
"eval_runtime": 45.2931,
|
| 1291 |
+
"eval_samples_per_second": 907.466,
|
| 1292 |
+
"eval_steps_per_second": 56.719,
|
| 1293 |
+
"step": 936000
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 19.34,
|
| 1297 |
+
"learning_rate": 6.070207621112316e-06,
|
| 1298 |
+
"loss": 3.3754,
|
| 1299 |
+
"step": 944000
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 19.34,
|
| 1303 |
+
"eval_loss": 3.5228991508483887,
|
| 1304 |
+
"eval_runtime": 47.0715,
|
| 1305 |
+
"eval_samples_per_second": 873.183,
|
| 1306 |
+
"eval_steps_per_second": 54.577,
|
| 1307 |
+
"step": 944000
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 19.5,
|
| 1311 |
+
"eval_loss": 3.538318634033203,
|
| 1312 |
+
"eval_runtime": 45.9256,
|
| 1313 |
+
"eval_samples_per_second": 894.97,
|
| 1314 |
+
"eval_steps_per_second": 55.938,
|
| 1315 |
+
"step": 952000
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 19.67,
|
| 1319 |
+
"learning_rate": 6.003502042858335e-06,
|
| 1320 |
+
"loss": 3.3893,
|
| 1321 |
+
"step": 960000
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 19.67,
|
| 1325 |
+
"eval_loss": 3.544513463973999,
|
| 1326 |
+
"eval_runtime": 46.8245,
|
| 1327 |
+
"eval_samples_per_second": 877.788,
|
| 1328 |
+
"eval_steps_per_second": 54.864,
|
| 1329 |
+
"step": 960000
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 19.83,
|
| 1333 |
+
"eval_loss": 3.5230634212493896,
|
| 1334 |
+
"eval_runtime": 47.3348,
|
| 1335 |
+
"eval_samples_per_second": 868.325,
|
| 1336 |
+
"eval_steps_per_second": 54.273,
|
| 1337 |
+
"step": 968000
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 20.0,
|
| 1341 |
+
"learning_rate": 5.936796464604353e-06,
|
| 1342 |
+
"loss": 3.3899,
|
| 1343 |
+
"step": 976000
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 20.0,
|
| 1347 |
+
"eval_loss": 3.531026840209961,
|
| 1348 |
+
"eval_runtime": 45.7886,
|
| 1349 |
+
"eval_samples_per_second": 897.647,
|
| 1350 |
+
"eval_steps_per_second": 56.106,
|
| 1351 |
+
"step": 976000
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 20.16,
|
| 1355 |
+
"eval_loss": 3.53287935256958,
|
| 1356 |
+
"eval_runtime": 46.7771,
|
| 1357 |
+
"eval_samples_per_second": 878.677,
|
| 1358 |
+
"eval_steps_per_second": 54.92,
|
| 1359 |
+
"step": 984000
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 20.32,
|
| 1363 |
+
"learning_rate": 5.870090886350371e-06,
|
| 1364 |
+
"loss": 3.3918,
|
| 1365 |
+
"step": 992000
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 20.32,
|
| 1369 |
+
"eval_loss": 3.5158653259277344,
|
| 1370 |
+
"eval_runtime": 46.2173,
|
| 1371 |
+
"eval_samples_per_second": 889.32,
|
| 1372 |
+
"eval_steps_per_second": 55.585,
|
| 1373 |
+
"step": 992000
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 20.49,
|
| 1377 |
+
"eval_loss": 3.562788486480713,
|
| 1378 |
+
"eval_runtime": 45.7474,
|
| 1379 |
+
"eval_samples_per_second": 898.456,
|
| 1380 |
+
"eval_steps_per_second": 56.156,
|
| 1381 |
+
"step": 1000000
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 20.65,
|
| 1385 |
+
"learning_rate": 5.80338530809639e-06,
|
| 1386 |
+
"loss": 3.3786,
|
| 1387 |
+
"step": 1008000
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 20.65,
|
| 1391 |
+
"eval_loss": 3.5290534496307373,
|
| 1392 |
+
"eval_runtime": 46.4581,
|
| 1393 |
+
"eval_samples_per_second": 884.711,
|
| 1394 |
+
"eval_steps_per_second": 55.297,
|
| 1395 |
+
"step": 1008000
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 20.82,
|
| 1399 |
+
"eval_loss": 3.5163111686706543,
|
| 1400 |
+
"eval_runtime": 45.899,
|
| 1401 |
+
"eval_samples_per_second": 895.487,
|
| 1402 |
+
"eval_steps_per_second": 55.971,
|
| 1403 |
+
"step": 1016000
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 20.98,
|
| 1407 |
+
"learning_rate": 5.736679729842408e-06,
|
| 1408 |
+
"loss": 3.3862,
|
| 1409 |
+
"step": 1024000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 20.98,
|
| 1413 |
+
"eval_loss": 3.531219959259033,
|
| 1414 |
+
"eval_runtime": 45.4959,
|
| 1415 |
+
"eval_samples_per_second": 903.423,
|
| 1416 |
+
"eval_steps_per_second": 56.467,
|
| 1417 |
+
"step": 1024000
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 21.14,
|
| 1421 |
+
"eval_loss": 3.514033317565918,
|
| 1422 |
+
"eval_runtime": 46.6408,
|
| 1423 |
+
"eval_samples_per_second": 881.245,
|
| 1424 |
+
"eval_steps_per_second": 55.08,
|
| 1425 |
+
"step": 1032000
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 21.31,
|
| 1429 |
+
"learning_rate": 5.669974151588427e-06,
|
| 1430 |
+
"loss": 3.3855,
|
| 1431 |
+
"step": 1040000
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 21.31,
|
| 1435 |
+
"eval_loss": 3.5617153644561768,
|
| 1436 |
+
"eval_runtime": 45.7071,
|
| 1437 |
+
"eval_samples_per_second": 899.248,
|
| 1438 |
+
"eval_steps_per_second": 56.206,
|
| 1439 |
+
"step": 1040000
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 21.47,
|
| 1443 |
+
"eval_loss": 3.5374927520751953,
|
| 1444 |
+
"eval_runtime": 45.668,
|
| 1445 |
+
"eval_samples_per_second": 900.018,
|
| 1446 |
+
"eval_steps_per_second": 56.254,
|
| 1447 |
+
"step": 1048000
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 21.64,
|
| 1451 |
+
"learning_rate": 5.603268573334446e-06,
|
| 1452 |
+
"loss": 3.3872,
|
| 1453 |
+
"step": 1056000
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 21.64,
|
| 1457 |
+
"eval_loss": 3.532823085784912,
|
| 1458 |
+
"eval_runtime": 46.5514,
|
| 1459 |
+
"eval_samples_per_second": 882.938,
|
| 1460 |
+
"eval_steps_per_second": 55.186,
|
| 1461 |
+
"step": 1056000
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 21.8,
|
| 1465 |
+
"eval_loss": 3.561626434326172,
|
| 1466 |
+
"eval_runtime": 45.9586,
|
| 1467 |
+
"eval_samples_per_second": 894.327,
|
| 1468 |
+
"eval_steps_per_second": 55.898,
|
| 1469 |
+
"step": 1064000
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 21.96,
|
| 1473 |
+
"learning_rate": 5.536562995080464e-06,
|
| 1474 |
+
"loss": 3.3931,
|
| 1475 |
+
"step": 1072000
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 21.96,
|
| 1479 |
+
"eval_loss": 3.5647873878479004,
|
| 1480 |
+
"eval_runtime": 46.8936,
|
| 1481 |
+
"eval_samples_per_second": 876.495,
|
| 1482 |
+
"eval_steps_per_second": 54.784,
|
| 1483 |
+
"step": 1072000
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 22.13,
|
| 1487 |
+
"eval_loss": 3.544335126876831,
|
| 1488 |
+
"eval_runtime": 46.3686,
|
| 1489 |
+
"eval_samples_per_second": 886.419,
|
| 1490 |
+
"eval_steps_per_second": 55.404,
|
| 1491 |
+
"step": 1080000
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 22.29,
|
| 1495 |
+
"learning_rate": 5.469857416826483e-06,
|
| 1496 |
+
"loss": 3.3708,
|
| 1497 |
+
"step": 1088000
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 22.29,
|
| 1501 |
+
"eval_loss": 3.5400941371917725,
|
| 1502 |
+
"eval_runtime": 45.8359,
|
| 1503 |
+
"eval_samples_per_second": 896.72,
|
| 1504 |
+
"eval_steps_per_second": 56.048,
|
| 1505 |
+
"step": 1088000
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 22.45,
|
| 1509 |
+
"eval_loss": 3.55292010307312,
|
| 1510 |
+
"eval_runtime": 46.8082,
|
| 1511 |
+
"eval_samples_per_second": 878.095,
|
| 1512 |
+
"eval_steps_per_second": 54.884,
|
| 1513 |
+
"step": 1096000
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 22.62,
|
| 1517 |
+
"learning_rate": 5.403151838572501e-06,
|
| 1518 |
+
"loss": 3.4099,
|
| 1519 |
+
"step": 1104000
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 22.62,
|
| 1523 |
+
"eval_loss": 3.533414602279663,
|
| 1524 |
+
"eval_runtime": 46.1107,
|
| 1525 |
+
"eval_samples_per_second": 891.377,
|
| 1526 |
+
"eval_steps_per_second": 55.714,
|
| 1527 |
+
"step": 1104000
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 22.78,
|
| 1531 |
+
"eval_loss": 3.5325212478637695,
|
| 1532 |
+
"eval_runtime": 46.1351,
|
| 1533 |
+
"eval_samples_per_second": 890.905,
|
| 1534 |
+
"eval_steps_per_second": 55.684,
|
| 1535 |
+
"step": 1112000
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 22.95,
|
| 1539 |
+
"learning_rate": 5.33644626031852e-06,
|
| 1540 |
+
"loss": 3.4027,
|
| 1541 |
+
"step": 1120000
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 22.95,
|
| 1545 |
+
"eval_loss": 3.5818660259246826,
|
| 1546 |
+
"eval_runtime": 46.7428,
|
| 1547 |
+
"eval_samples_per_second": 879.323,
|
| 1548 |
+
"eval_steps_per_second": 54.96,
|
| 1549 |
+
"step": 1120000
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 23.11,
|
| 1553 |
+
"eval_loss": 3.5470829010009766,
|
| 1554 |
+
"eval_runtime": 46.1344,
|
| 1555 |
+
"eval_samples_per_second": 890.92,
|
| 1556 |
+
"eval_steps_per_second": 55.685,
|
| 1557 |
+
"step": 1128000
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 23.27,
|
| 1561 |
+
"learning_rate": 5.269740682064538e-06,
|
| 1562 |
+
"loss": 3.4035,
|
| 1563 |
+
"step": 1136000
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 23.27,
|
| 1567 |
+
"eval_loss": 3.548552989959717,
|
| 1568 |
+
"eval_runtime": 46.1071,
|
| 1569 |
+
"eval_samples_per_second": 891.446,
|
| 1570 |
+
"eval_steps_per_second": 55.718,
|
| 1571 |
+
"step": 1136000
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 23.44,
|
| 1575 |
+
"eval_loss": 3.5470151901245117,
|
| 1576 |
+
"eval_runtime": 46.849,
|
| 1577 |
+
"eval_samples_per_second": 877.33,
|
| 1578 |
+
"eval_steps_per_second": 54.836,
|
| 1579 |
+
"step": 1144000
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 23.6,
|
| 1583 |
+
"learning_rate": 5.203035103810556e-06,
|
| 1584 |
+
"loss": 3.3964,
|
| 1585 |
+
"step": 1152000
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 23.6,
|
| 1589 |
+
"eval_loss": 3.572176694869995,
|
| 1590 |
+
"eval_runtime": 46.3661,
|
| 1591 |
+
"eval_samples_per_second": 886.467,
|
| 1592 |
+
"eval_steps_per_second": 55.407,
|
| 1593 |
+
"step": 1152000
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 23.77,
|
| 1597 |
+
"eval_loss": 3.55098295211792,
|
| 1598 |
+
"eval_runtime": 46.1812,
|
| 1599 |
+
"eval_samples_per_second": 890.015,
|
| 1600 |
+
"eval_steps_per_second": 55.629,
|
| 1601 |
+
"step": 1160000
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 23.93,
|
| 1605 |
+
"learning_rate": 5.136329525556575e-06,
|
| 1606 |
+
"loss": 3.4115,
|
| 1607 |
+
"step": 1168000
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 23.93,
|
| 1611 |
+
"eval_loss": 3.561007499694824,
|
| 1612 |
+
"eval_runtime": 47.5429,
|
| 1613 |
+
"eval_samples_per_second": 864.525,
|
| 1614 |
+
"eval_steps_per_second": 54.035,
|
| 1615 |
+
"step": 1168000
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 24.09,
|
| 1619 |
+
"eval_loss": 3.5757482051849365,
|
| 1620 |
+
"eval_runtime": 46.3962,
|
| 1621 |
+
"eval_samples_per_second": 885.891,
|
| 1622 |
+
"eval_steps_per_second": 55.371,
|
| 1623 |
+
"step": 1176000
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 24.26,
|
| 1627 |
+
"learning_rate": 5.0696239473025935e-06,
|
| 1628 |
+
"loss": 3.4173,
|
| 1629 |
+
"step": 1184000
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 24.26,
|
| 1633 |
+
"eval_loss": 3.554094076156616,
|
| 1634 |
+
"eval_runtime": 45.5708,
|
| 1635 |
+
"eval_samples_per_second": 901.936,
|
| 1636 |
+
"eval_steps_per_second": 56.374,
|
| 1637 |
+
"step": 1184000
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 24.42,
|
| 1641 |
+
"eval_loss": 3.577660083770752,
|
| 1642 |
+
"eval_runtime": 47.0565,
|
| 1643 |
+
"eval_samples_per_second": 873.461,
|
| 1644 |
+
"eval_steps_per_second": 54.594,
|
| 1645 |
+
"step": 1192000
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 24.59,
|
| 1649 |
+
"learning_rate": 5.002918369048611e-06,
|
| 1650 |
+
"loss": 3.4169,
|
| 1651 |
+
"step": 1200000
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 24.59,
|
| 1655 |
+
"eval_loss": 3.5637948513031006,
|
| 1656 |
+
"eval_runtime": 47.0711,
|
| 1657 |
+
"eval_samples_per_second": 873.19,
|
| 1658 |
+
"eval_steps_per_second": 54.577,
|
| 1659 |
+
"step": 1200000
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 24.75,
|
| 1663 |
+
"eval_loss": 3.5462896823883057,
|
| 1664 |
+
"eval_runtime": 46.8215,
|
| 1665 |
+
"eval_samples_per_second": 877.845,
|
| 1666 |
+
"eval_steps_per_second": 54.868,
|
| 1667 |
+
"step": 1208000
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 24.91,
|
| 1671 |
+
"learning_rate": 4.936212790794631e-06,
|
| 1672 |
+
"loss": 3.4031,
|
| 1673 |
+
"step": 1216000
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 24.91,
|
| 1677 |
+
"eval_loss": 3.5299670696258545,
|
| 1678 |
+
"eval_runtime": 46.9742,
|
| 1679 |
+
"eval_samples_per_second": 874.99,
|
| 1680 |
+
"eval_steps_per_second": 54.69,
|
| 1681 |
+
"step": 1216000
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 25.08,
|
| 1685 |
+
"eval_loss": 3.558427333831787,
|
| 1686 |
+
"eval_runtime": 46.1322,
|
| 1687 |
+
"eval_samples_per_second": 890.961,
|
| 1688 |
+
"eval_steps_per_second": 55.688,
|
| 1689 |
+
"step": 1224000
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 25.24,
|
| 1693 |
+
"learning_rate": 4.869507212540649e-06,
|
| 1694 |
+
"loss": 3.4094,
|
| 1695 |
+
"step": 1232000
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 25.24,
|
| 1699 |
+
"eval_loss": 3.568174123764038,
|
| 1700 |
+
"eval_runtime": 46.3049,
|
| 1701 |
+
"eval_samples_per_second": 887.638,
|
| 1702 |
+
"eval_steps_per_second": 55.48,
|
| 1703 |
+
"step": 1232000
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 25.41,
|
| 1707 |
+
"eval_loss": 3.555844783782959,
|
| 1708 |
+
"eval_runtime": 46.0676,
|
| 1709 |
+
"eval_samples_per_second": 892.211,
|
| 1710 |
+
"eval_steps_per_second": 55.766,
|
| 1711 |
+
"step": 1240000
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 25.57,
|
| 1715 |
+
"learning_rate": 4.802801634286667e-06,
|
| 1716 |
+
"loss": 3.4116,
|
| 1717 |
+
"step": 1248000
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 25.57,
|
| 1721 |
+
"eval_loss": 3.5629091262817383,
|
| 1722 |
+
"eval_runtime": 45.5765,
|
| 1723 |
+
"eval_samples_per_second": 901.825,
|
| 1724 |
+
"eval_steps_per_second": 56.367,
|
| 1725 |
+
"step": 1248000
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 25.73,
|
| 1729 |
+
"eval_loss": 3.5490224361419678,
|
| 1730 |
+
"eval_runtime": 46.4409,
|
| 1731 |
+
"eval_samples_per_second": 885.039,
|
| 1732 |
+
"eval_steps_per_second": 55.318,
|
| 1733 |
+
"step": 1256000
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 25.9,
|
| 1737 |
+
"learning_rate": 4.7360960560326865e-06,
|
| 1738 |
+
"loss": 3.4199,
|
| 1739 |
+
"step": 1264000
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 25.9,
|
| 1743 |
+
"eval_loss": 3.567878484725952,
|
| 1744 |
+
"eval_runtime": 46.1595,
|
| 1745 |
+
"eval_samples_per_second": 890.434,
|
| 1746 |
+
"eval_steps_per_second": 55.655,
|
| 1747 |
+
"step": 1264000
|
| 1748 |
+
},
|
| 1749 |
+
{
|
| 1750 |
+
"epoch": 26.06,
|
| 1751 |
+
"eval_loss": 3.5885465145111084,
|
| 1752 |
+
"eval_runtime": 45.9316,
|
| 1753 |
+
"eval_samples_per_second": 894.853,
|
| 1754 |
+
"eval_steps_per_second": 55.931,
|
| 1755 |
+
"step": 1272000
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 26.22,
|
| 1759 |
+
"learning_rate": 4.669390477778704e-06,
|
| 1760 |
+
"loss": 3.412,
|
| 1761 |
+
"step": 1280000
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 26.22,
|
| 1765 |
+
"eval_loss": 3.5578629970550537,
|
| 1766 |
+
"eval_runtime": 46.4337,
|
| 1767 |
+
"eval_samples_per_second": 885.176,
|
| 1768 |
+
"eval_steps_per_second": 55.326,
|
| 1769 |
+
"step": 1280000
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 26.39,
|
| 1773 |
+
"eval_loss": 3.5465352535247803,
|
| 1774 |
+
"eval_runtime": 45.7517,
|
| 1775 |
+
"eval_samples_per_second": 898.371,
|
| 1776 |
+
"eval_steps_per_second": 56.151,
|
| 1777 |
+
"step": 1288000
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 26.55,
|
| 1781 |
+
"learning_rate": 4.602684899524723e-06,
|
| 1782 |
+
"loss": 3.4123,
|
| 1783 |
+
"step": 1296000
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 26.55,
|
| 1787 |
+
"eval_loss": 3.572610855102539,
|
| 1788 |
+
"eval_runtime": 45.5426,
|
| 1789 |
+
"eval_samples_per_second": 902.496,
|
| 1790 |
+
"eval_steps_per_second": 56.409,
|
| 1791 |
+
"step": 1296000
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 26.72,
|
| 1795 |
+
"eval_loss": 3.577484130859375,
|
| 1796 |
+
"eval_runtime": 46.4204,
|
| 1797 |
+
"eval_samples_per_second": 885.431,
|
| 1798 |
+
"eval_steps_per_second": 55.342,
|
| 1799 |
+
"step": 1304000
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 26.88,
|
| 1803 |
+
"learning_rate": 4.5359793212707415e-06,
|
| 1804 |
+
"loss": 3.4132,
|
| 1805 |
+
"step": 1312000
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 26.88,
|
| 1809 |
+
"eval_loss": 3.5477850437164307,
|
| 1810 |
+
"eval_runtime": 45.6512,
|
| 1811 |
+
"eval_samples_per_second": 900.348,
|
| 1812 |
+
"eval_steps_per_second": 56.275,
|
| 1813 |
+
"step": 1312000
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 27.04,
|
| 1817 |
+
"eval_loss": 3.5588574409484863,
|
| 1818 |
+
"eval_runtime": 46.0446,
|
| 1819 |
+
"eval_samples_per_second": 892.657,
|
| 1820 |
+
"eval_steps_per_second": 55.794,
|
| 1821 |
+
"step": 1320000
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 27.21,
|
| 1825 |
+
"learning_rate": 4.46927374301676e-06,
|
| 1826 |
+
"loss": 3.4161,
|
| 1827 |
+
"step": 1328000
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 27.21,
|
| 1831 |
+
"eval_loss": 3.56620717048645,
|
| 1832 |
+
"eval_runtime": 46.4839,
|
| 1833 |
+
"eval_samples_per_second": 884.22,
|
| 1834 |
+
"eval_steps_per_second": 55.266,
|
| 1835 |
+
"step": 1328000
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 27.37,
|
| 1839 |
+
"eval_loss": 3.589487075805664,
|
| 1840 |
+
"eval_runtime": 46.3966,
|
| 1841 |
+
"eval_samples_per_second": 885.884,
|
| 1842 |
+
"eval_steps_per_second": 55.37,
|
| 1843 |
+
"step": 1336000
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 27.54,
|
| 1847 |
+
"learning_rate": 4.402568164762779e-06,
|
| 1848 |
+
"loss": 3.4097,
|
| 1849 |
+
"step": 1344000
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 27.54,
|
| 1853 |
+
"eval_loss": 3.5940632820129395,
|
| 1854 |
+
"eval_runtime": 46.4364,
|
| 1855 |
+
"eval_samples_per_second": 885.125,
|
| 1856 |
+
"eval_steps_per_second": 55.323,
|
| 1857 |
+
"step": 1344000
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 27.7,
|
| 1861 |
+
"eval_loss": 3.5912110805511475,
|
| 1862 |
+
"eval_runtime": 45.9687,
|
| 1863 |
+
"eval_samples_per_second": 894.131,
|
| 1864 |
+
"eval_steps_per_second": 55.886,
|
| 1865 |
+
"step": 1352000
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 27.86,
|
| 1869 |
+
"learning_rate": 4.335862586508797e-06,
|
| 1870 |
+
"loss": 3.415,
|
| 1871 |
+
"step": 1360000
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 27.86,
|
| 1875 |
+
"eval_loss": 3.565756320953369,
|
| 1876 |
+
"eval_runtime": 45.7621,
|
| 1877 |
+
"eval_samples_per_second": 898.168,
|
| 1878 |
+
"eval_steps_per_second": 56.138,
|
| 1879 |
+
"step": 1360000
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 28.03,
|
| 1883 |
+
"eval_loss": 3.5553781986236572,
|
| 1884 |
+
"eval_runtime": 46.2903,
|
| 1885 |
+
"eval_samples_per_second": 887.919,
|
| 1886 |
+
"eval_steps_per_second": 55.498,
|
| 1887 |
+
"step": 1368000
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 28.19,
|
| 1891 |
+
"learning_rate": 4.269157008254816e-06,
|
| 1892 |
+
"loss": 3.4193,
|
| 1893 |
+
"step": 1376000
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 28.19,
|
| 1897 |
+
"eval_loss": 3.589851140975952,
|
| 1898 |
+
"eval_runtime": 45.8411,
|
| 1899 |
+
"eval_samples_per_second": 896.618,
|
| 1900 |
+
"eval_steps_per_second": 56.041,
|
| 1901 |
+
"step": 1376000
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 28.36,
|
| 1905 |
+
"eval_loss": 3.5652260780334473,
|
| 1906 |
+
"eval_runtime": 45.5538,
|
| 1907 |
+
"eval_samples_per_second": 902.275,
|
| 1908 |
+
"eval_steps_per_second": 56.395,
|
| 1909 |
+
"step": 1384000
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 28.52,
|
| 1913 |
+
"learning_rate": 4.202451430000834e-06,
|
| 1914 |
+
"loss": 3.4136,
|
| 1915 |
+
"step": 1392000
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 28.52,
|
| 1919 |
+
"eval_loss": 3.5832390785217285,
|
| 1920 |
+
"eval_runtime": 46.575,
|
| 1921 |
+
"eval_samples_per_second": 882.491,
|
| 1922 |
+
"eval_steps_per_second": 55.158,
|
| 1923 |
+
"step": 1392000
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 28.68,
|
| 1927 |
+
"eval_loss": 3.5885210037231445,
|
| 1928 |
+
"eval_runtime": 45.9659,
|
| 1929 |
+
"eval_samples_per_second": 894.184,
|
| 1930 |
+
"eval_steps_per_second": 55.889,
|
| 1931 |
+
"step": 1400000
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 28.85,
|
| 1935 |
+
"learning_rate": 4.135745851746852e-06,
|
| 1936 |
+
"loss": 3.4294,
|
| 1937 |
+
"step": 1408000
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 28.85,
|
| 1941 |
+
"eval_loss": 3.583249807357788,
|
| 1942 |
+
"eval_runtime": 45.7927,
|
| 1943 |
+
"eval_samples_per_second": 897.568,
|
| 1944 |
+
"eval_steps_per_second": 56.101,
|
| 1945 |
+
"step": 1408000
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 29.01,
|
| 1949 |
+
"eval_loss": 3.6025209426879883,
|
| 1950 |
+
"eval_runtime": 46.362,
|
| 1951 |
+
"eval_samples_per_second": 886.546,
|
| 1952 |
+
"eval_steps_per_second": 55.412,
|
| 1953 |
+
"step": 1416000
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 29.17,
|
| 1957 |
+
"learning_rate": 4.069040273492872e-06,
|
| 1958 |
+
"loss": 3.4243,
|
| 1959 |
+
"step": 1424000
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 29.17,
|
| 1963 |
+
"eval_loss": 3.6040360927581787,
|
| 1964 |
+
"eval_runtime": 45.7855,
|
| 1965 |
+
"eval_samples_per_second": 897.708,
|
| 1966 |
+
"eval_steps_per_second": 56.11,
|
| 1967 |
+
"step": 1424000
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 29.34,
|
| 1971 |
+
"eval_loss": 3.5890395641326904,
|
| 1972 |
+
"eval_runtime": 46.5109,
|
| 1973 |
+
"eval_samples_per_second": 883.707,
|
| 1974 |
+
"eval_steps_per_second": 55.234,
|
| 1975 |
+
"step": 1432000
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 29.5,
|
| 1979 |
+
"learning_rate": 4.0023346952388895e-06,
|
| 1980 |
+
"loss": 3.4427,
|
| 1981 |
+
"step": 1440000
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 29.5,
|
| 1985 |
+
"eval_loss": 3.58347749710083,
|
| 1986 |
+
"eval_runtime": 46.2896,
|
| 1987 |
+
"eval_samples_per_second": 887.931,
|
| 1988 |
+
"eval_steps_per_second": 55.498,
|
| 1989 |
+
"step": 1440000
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 29.67,
|
| 1993 |
+
"eval_loss": 3.6185286045074463,
|
| 1994 |
+
"eval_runtime": 46.4189,
|
| 1995 |
+
"eval_samples_per_second": 885.459,
|
| 1996 |
+
"eval_steps_per_second": 55.344,
|
| 1997 |
+
"step": 1448000
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 29.83,
|
| 2001 |
+
"learning_rate": 3.935629116984908e-06,
|
| 2002 |
+
"loss": 3.4293,
|
| 2003 |
+
"step": 1456000
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 29.83,
|
| 2007 |
+
"eval_loss": 3.6028919219970703,
|
| 2008 |
+
"eval_runtime": 46.7251,
|
| 2009 |
+
"eval_samples_per_second": 879.656,
|
| 2010 |
+
"eval_steps_per_second": 54.981,
|
| 2011 |
+
"step": 1456000
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 29.99,
|
| 2015 |
+
"eval_loss": 3.616161823272705,
|
| 2016 |
+
"eval_runtime": 45.7265,
|
| 2017 |
+
"eval_samples_per_second": 898.865,
|
| 2018 |
+
"eval_steps_per_second": 56.182,
|
| 2019 |
+
"step": 1464000
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 30.16,
|
| 2023 |
+
"learning_rate": 3.868923538730927e-06,
|
| 2024 |
+
"loss": 3.4363,
|
| 2025 |
+
"step": 1472000
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 30.16,
|
| 2029 |
+
"eval_loss": 3.6257941722869873,
|
| 2030 |
+
"eval_runtime": 45.6532,
|
| 2031 |
+
"eval_samples_per_second": 900.308,
|
| 2032 |
+
"eval_steps_per_second": 56.272,
|
| 2033 |
+
"step": 1472000
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 30.32,
|
| 2037 |
+
"eval_loss": 3.6038014888763428,
|
| 2038 |
+
"eval_runtime": 46.717,
|
| 2039 |
+
"eval_samples_per_second": 879.808,
|
| 2040 |
+
"eval_steps_per_second": 54.991,
|
| 2041 |
+
"step": 1480000
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 30.49,
|
| 2045 |
+
"learning_rate": 3.8022179604769453e-06,
|
| 2046 |
+
"loss": 3.4532,
|
| 2047 |
+
"step": 1488000
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 30.49,
|
| 2051 |
+
"eval_loss": 3.6039483547210693,
|
| 2052 |
+
"eval_runtime": 45.742,
|
| 2053 |
+
"eval_samples_per_second": 898.562,
|
| 2054 |
+
"eval_steps_per_second": 56.163,
|
| 2055 |
+
"step": 1488000
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 30.65,
|
| 2059 |
+
"eval_loss": 3.605367422103882,
|
| 2060 |
+
"eval_runtime": 45.7078,
|
| 2061 |
+
"eval_samples_per_second": 899.234,
|
| 2062 |
+
"eval_steps_per_second": 56.205,
|
| 2063 |
+
"step": 1496000
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 30.81,
|
| 2067 |
+
"learning_rate": 3.735512382222964e-06,
|
| 2068 |
+
"loss": 3.4401,
|
| 2069 |
+
"step": 1504000
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 30.81,
|
| 2073 |
+
"eval_loss": 3.6269376277923584,
|
| 2074 |
+
"eval_runtime": 46.6124,
|
| 2075 |
+
"eval_samples_per_second": 881.783,
|
| 2076 |
+
"eval_steps_per_second": 55.114,
|
| 2077 |
+
"step": 1504000
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 30.98,
|
| 2081 |
+
"eval_loss": 3.600417137145996,
|
| 2082 |
+
"eval_runtime": 47.0146,
|
| 2083 |
+
"eval_samples_per_second": 874.239,
|
| 2084 |
+
"eval_steps_per_second": 54.643,
|
| 2085 |
+
"step": 1512000
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 31.14,
|
| 2089 |
+
"learning_rate": 3.668806803968982e-06,
|
| 2090 |
+
"loss": 3.4491,
|
| 2091 |
+
"step": 1520000
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 31.14,
|
| 2095 |
+
"eval_loss": 3.6095597743988037,
|
| 2096 |
+
"eval_runtime": 47.1653,
|
| 2097 |
+
"eval_samples_per_second": 871.446,
|
| 2098 |
+
"eval_steps_per_second": 54.468,
|
| 2099 |
+
"step": 1520000
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 31.31,
|
| 2103 |
+
"eval_loss": 3.6216766834259033,
|
| 2104 |
+
"eval_runtime": 48.343,
|
| 2105 |
+
"eval_samples_per_second": 850.216,
|
| 2106 |
+
"eval_steps_per_second": 53.141,
|
| 2107 |
+
"step": 1528000
|
| 2108 |
+
},
|
| 2109 |
+
{
|
| 2110 |
+
"epoch": 31.47,
|
| 2111 |
+
"learning_rate": 3.6021012257150007e-06,
|
| 2112 |
+
"loss": 3.4438,
|
| 2113 |
+
"step": 1536000
|
| 2114 |
+
},
|
| 2115 |
+
{
|
| 2116 |
+
"epoch": 31.47,
|
| 2117 |
+
"eval_loss": 3.6081080436706543,
|
| 2118 |
+
"eval_runtime": 47.4804,
|
| 2119 |
+
"eval_samples_per_second": 865.663,
|
| 2120 |
+
"eval_steps_per_second": 54.107,
|
| 2121 |
+
"step": 1536000
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 31.63,
|
| 2125 |
+
"eval_loss": 3.6190168857574463,
|
| 2126 |
+
"eval_runtime": 48.3587,
|
| 2127 |
+
"eval_samples_per_second": 849.941,
|
| 2128 |
+
"eval_steps_per_second": 53.124,
|
| 2129 |
+
"step": 1544000
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 31.8,
|
| 2133 |
+
"learning_rate": 3.535395647461019e-06,
|
| 2134 |
+
"loss": 3.4337,
|
| 2135 |
+
"step": 1552000
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 31.8,
|
| 2139 |
+
"eval_loss": 3.611992835998535,
|
| 2140 |
+
"eval_runtime": 47.5342,
|
| 2141 |
+
"eval_samples_per_second": 864.683,
|
| 2142 |
+
"eval_steps_per_second": 54.045,
|
| 2143 |
+
"step": 1552000
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 31.96,
|
| 2147 |
+
"eval_loss": 3.586127996444702,
|
| 2148 |
+
"eval_runtime": 46.8726,
|
| 2149 |
+
"eval_samples_per_second": 876.888,
|
| 2150 |
+
"eval_steps_per_second": 54.808,
|
| 2151 |
+
"step": 1560000
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 32.13,
|
| 2155 |
+
"learning_rate": 3.468690069207038e-06,
|
| 2156 |
+
"loss": 3.4475,
|
| 2157 |
+
"step": 1568000
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 32.13,
|
| 2161 |
+
"eval_loss": 3.620932102203369,
|
| 2162 |
+
"eval_runtime": 48.2654,
|
| 2163 |
+
"eval_samples_per_second": 851.582,
|
| 2164 |
+
"eval_steps_per_second": 53.226,
|
| 2165 |
+
"step": 1568000
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 32.29,
|
| 2169 |
+
"eval_loss": 3.6301937103271484,
|
| 2170 |
+
"eval_runtime": 47.2416,
|
| 2171 |
+
"eval_samples_per_second": 870.039,
|
| 2172 |
+
"eval_steps_per_second": 54.38,
|
| 2173 |
+
"step": 1576000
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 32.45,
|
| 2177 |
+
"learning_rate": 3.4019844909530565e-06,
|
| 2178 |
+
"loss": 3.4406,
|
| 2179 |
+
"step": 1584000
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 32.45,
|
| 2183 |
+
"eval_loss": 3.6052932739257812,
|
| 2184 |
+
"eval_runtime": 46.0861,
|
| 2185 |
+
"eval_samples_per_second": 891.852,
|
| 2186 |
+
"eval_steps_per_second": 55.743,
|
| 2187 |
+
"step": 1584000
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 32.62,
|
| 2191 |
+
"eval_loss": 3.593369960784912,
|
| 2192 |
+
"eval_runtime": 49.6475,
|
| 2193 |
+
"eval_samples_per_second": 827.876,
|
| 2194 |
+
"eval_steps_per_second": 51.745,
|
| 2195 |
+
"step": 1592000
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 32.78,
|
| 2199 |
+
"learning_rate": 3.3352789126990747e-06,
|
| 2200 |
+
"loss": 3.4392,
|
| 2201 |
+
"step": 1600000
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 32.78,
|
| 2205 |
+
"eval_loss": 3.594203472137451,
|
| 2206 |
+
"eval_runtime": 47.8907,
|
| 2207 |
+
"eval_samples_per_second": 858.246,
|
| 2208 |
+
"eval_steps_per_second": 53.643,
|
| 2209 |
+
"step": 1600000
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 32.94,
|
| 2213 |
+
"eval_loss": 3.601329803466797,
|
| 2214 |
+
"eval_runtime": 46.6549,
|
| 2215 |
+
"eval_samples_per_second": 880.98,
|
| 2216 |
+
"eval_steps_per_second": 55.064,
|
| 2217 |
+
"step": 1608000
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 33.11,
|
| 2221 |
+
"learning_rate": 3.2685733344450933e-06,
|
| 2222 |
+
"loss": 3.4514,
|
| 2223 |
+
"step": 1616000
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 33.11,
|
| 2227 |
+
"eval_loss": 3.6505630016326904,
|
| 2228 |
+
"eval_runtime": 47.3453,
|
| 2229 |
+
"eval_samples_per_second": 868.132,
|
| 2230 |
+
"eval_steps_per_second": 54.261,
|
| 2231 |
+
"step": 1616000
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 33.27,
|
| 2235 |
+
"eval_loss": 3.604905128479004,
|
| 2236 |
+
"eval_runtime": 47.3478,
|
| 2237 |
+
"eval_samples_per_second": 868.087,
|
| 2238 |
+
"eval_steps_per_second": 54.258,
|
| 2239 |
+
"step": 1624000
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 33.44,
|
| 2243 |
+
"learning_rate": 3.2018677561911115e-06,
|
| 2244 |
+
"loss": 3.4406,
|
| 2245 |
+
"step": 1632000
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 33.44,
|
| 2249 |
+
"eval_loss": 3.6285159587860107,
|
| 2250 |
+
"eval_runtime": 45.2665,
|
| 2251 |
+
"eval_samples_per_second": 908.001,
|
| 2252 |
+
"eval_steps_per_second": 56.753,
|
| 2253 |
+
"step": 1632000
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 33.6,
|
| 2257 |
+
"eval_loss": 3.6107122898101807,
|
| 2258 |
+
"eval_runtime": 47.0075,
|
| 2259 |
+
"eval_samples_per_second": 874.372,
|
| 2260 |
+
"eval_steps_per_second": 54.651,
|
| 2261 |
+
"step": 1640000
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 33.76,
|
| 2265 |
+
"learning_rate": 3.1351621779371306e-06,
|
| 2266 |
+
"loss": 3.4522,
|
| 2267 |
+
"step": 1648000
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 33.76,
|
| 2271 |
+
"eval_loss": 3.6080775260925293,
|
| 2272 |
+
"eval_runtime": 46.384,
|
| 2273 |
+
"eval_samples_per_second": 886.124,
|
| 2274 |
+
"eval_steps_per_second": 55.385,
|
| 2275 |
+
"step": 1648000
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 33.93,
|
| 2279 |
+
"eval_loss": 3.6121394634246826,
|
| 2280 |
+
"eval_runtime": 47.5808,
|
| 2281 |
+
"eval_samples_per_second": 863.836,
|
| 2282 |
+
"eval_steps_per_second": 53.992,
|
| 2283 |
+
"step": 1656000
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 34.09,
|
| 2287 |
+
"learning_rate": 3.0684565996831487e-06,
|
| 2288 |
+
"loss": 3.4592,
|
| 2289 |
+
"step": 1664000
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 34.09,
|
| 2293 |
+
"eval_loss": 3.639568567276001,
|
| 2294 |
+
"eval_runtime": 47.4907,
|
| 2295 |
+
"eval_samples_per_second": 865.474,
|
| 2296 |
+
"eval_steps_per_second": 54.095,
|
| 2297 |
+
"step": 1664000
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 34.26,
|
| 2301 |
+
"eval_loss": 3.628408432006836,
|
| 2302 |
+
"eval_runtime": 45.8805,
|
| 2303 |
+
"eval_samples_per_second": 895.849,
|
| 2304 |
+
"eval_steps_per_second": 55.993,
|
| 2305 |
+
"step": 1672000
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 34.42,
|
| 2309 |
+
"learning_rate": 3.0017510214291673e-06,
|
| 2310 |
+
"loss": 3.4587,
|
| 2311 |
+
"step": 1680000
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 34.42,
|
| 2315 |
+
"eval_loss": 3.619464635848999,
|
| 2316 |
+
"eval_runtime": 46.7813,
|
| 2317 |
+
"eval_samples_per_second": 878.599,
|
| 2318 |
+
"eval_steps_per_second": 54.915,
|
| 2319 |
+
"step": 1680000
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 34.58,
|
| 2323 |
+
"eval_loss": 3.6168148517608643,
|
| 2324 |
+
"eval_runtime": 46.0408,
|
| 2325 |
+
"eval_samples_per_second": 892.731,
|
| 2326 |
+
"eval_steps_per_second": 55.798,
|
| 2327 |
+
"step": 1688000
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 34.75,
|
| 2331 |
+
"learning_rate": 2.9350454431751855e-06,
|
| 2332 |
+
"loss": 3.4589,
|
| 2333 |
+
"step": 1696000
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 34.75,
|
| 2337 |
+
"eval_loss": 3.631527900695801,
|
| 2338 |
+
"eval_runtime": 45.9831,
|
| 2339 |
+
"eval_samples_per_second": 893.85,
|
| 2340 |
+
"eval_steps_per_second": 55.868,
|
| 2341 |
+
"step": 1696000
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 34.91,
|
| 2345 |
+
"eval_loss": 3.6044745445251465,
|
| 2346 |
+
"eval_runtime": 46.5293,
|
| 2347 |
+
"eval_samples_per_second": 883.356,
|
| 2348 |
+
"eval_steps_per_second": 55.212,
|
| 2349 |
+
"step": 1704000
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 35.08,
|
| 2353 |
+
"learning_rate": 2.868339864921204e-06,
|
| 2354 |
+
"loss": 3.4703,
|
| 2355 |
+
"step": 1712000
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 35.08,
|
| 2359 |
+
"eval_loss": 3.6251227855682373,
|
| 2360 |
+
"eval_runtime": 45.5912,
|
| 2361 |
+
"eval_samples_per_second": 901.533,
|
| 2362 |
+
"eval_steps_per_second": 56.349,
|
| 2363 |
+
"step": 1712000
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 35.24,
|
| 2367 |
+
"eval_loss": 3.6251931190490723,
|
| 2368 |
+
"eval_runtime": 45.7404,
|
| 2369 |
+
"eval_samples_per_second": 898.593,
|
| 2370 |
+
"eval_steps_per_second": 56.165,
|
| 2371 |
+
"step": 1720000
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 35.4,
|
| 2375 |
+
"learning_rate": 2.801634286667223e-06,
|
| 2376 |
+
"loss": 3.4565,
|
| 2377 |
+
"step": 1728000
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 35.4,
|
| 2381 |
+
"eval_loss": 3.62538743019104,
|
| 2382 |
+
"eval_runtime": 46.4207,
|
| 2383 |
+
"eval_samples_per_second": 885.423,
|
| 2384 |
+
"eval_steps_per_second": 55.342,
|
| 2385 |
+
"step": 1728000
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 35.57,
|
| 2389 |
+
"eval_loss": 3.6544113159179688,
|
| 2390 |
+
"eval_runtime": 45.7864,
|
| 2391 |
+
"eval_samples_per_second": 897.691,
|
| 2392 |
+
"eval_steps_per_second": 56.108,
|
| 2393 |
+
"step": 1736000
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 35.73,
|
| 2397 |
+
"learning_rate": 2.7349287084132413e-06,
|
| 2398 |
+
"loss": 3.4634,
|
| 2399 |
+
"step": 1744000
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 35.73,
|
| 2403 |
+
"eval_loss": 3.629049062728882,
|
| 2404 |
+
"eval_runtime": 46.556,
|
| 2405 |
+
"eval_samples_per_second": 882.85,
|
| 2406 |
+
"eval_steps_per_second": 55.181,
|
| 2407 |
+
"step": 1744000
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 35.9,
|
| 2411 |
+
"eval_loss": 3.612429618835449,
|
| 2412 |
+
"eval_runtime": 46.5059,
|
| 2413 |
+
"eval_samples_per_second": 883.802,
|
| 2414 |
+
"eval_steps_per_second": 55.24,
|
| 2415 |
+
"step": 1752000
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 36.06,
|
| 2419 |
+
"learning_rate": 2.66822313015926e-06,
|
| 2420 |
+
"loss": 3.4625,
|
| 2421 |
+
"step": 1760000
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 36.06,
|
| 2425 |
+
"eval_loss": 3.6262378692626953,
|
| 2426 |
+
"eval_runtime": 45.8554,
|
| 2427 |
+
"eval_samples_per_second": 896.34,
|
| 2428 |
+
"eval_steps_per_second": 56.024,
|
| 2429 |
+
"step": 1760000
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 36.22,
|
| 2433 |
+
"eval_loss": 3.6317975521087646,
|
| 2434 |
+
"eval_runtime": 46.7318,
|
| 2435 |
+
"eval_samples_per_second": 879.529,
|
| 2436 |
+
"eval_steps_per_second": 54.973,
|
| 2437 |
+
"step": 1768000
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 36.39,
|
| 2441 |
+
"learning_rate": 2.601517551905278e-06,
|
| 2442 |
+
"loss": 3.457,
|
| 2443 |
+
"step": 1776000
|
| 2444 |
+
},
|
| 2445 |
+
{
|
| 2446 |
+
"epoch": 36.39,
|
| 2447 |
+
"eval_loss": 3.640812397003174,
|
| 2448 |
+
"eval_runtime": 45.9688,
|
| 2449 |
+
"eval_samples_per_second": 894.129,
|
| 2450 |
+
"eval_steps_per_second": 55.886,
|
| 2451 |
+
"step": 1776000
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 36.55,
|
| 2455 |
+
"eval_loss": 3.6433026790618896,
|
| 2456 |
+
"eval_runtime": 45.8154,
|
| 2457 |
+
"eval_samples_per_second": 897.122,
|
| 2458 |
+
"eval_steps_per_second": 56.073,
|
| 2459 |
+
"step": 1784000
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 36.71,
|
| 2463 |
+
"learning_rate": 2.5348119736512967e-06,
|
| 2464 |
+
"loss": 3.4618,
|
| 2465 |
+
"step": 1792000
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 36.71,
|
| 2469 |
+
"eval_loss": 3.627612352371216,
|
| 2470 |
+
"eval_runtime": 46.6149,
|
| 2471 |
+
"eval_samples_per_second": 881.735,
|
| 2472 |
+
"eval_steps_per_second": 55.111,
|
| 2473 |
+
"step": 1792000
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 36.88,
|
| 2477 |
+
"eval_loss": 3.631366014480591,
|
| 2478 |
+
"eval_runtime": 46.0925,
|
| 2479 |
+
"eval_samples_per_second": 891.729,
|
| 2480 |
+
"eval_steps_per_second": 55.736,
|
| 2481 |
+
"step": 1800000
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 37.04,
|
| 2485 |
+
"learning_rate": 2.4681063953973154e-06,
|
| 2486 |
+
"loss": 3.4611,
|
| 2487 |
+
"step": 1808000
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 37.04,
|
| 2491 |
+
"eval_loss": 3.6415860652923584,
|
| 2492 |
+
"eval_runtime": 46.287,
|
| 2493 |
+
"eval_samples_per_second": 887.982,
|
| 2494 |
+
"eval_steps_per_second": 55.502,
|
| 2495 |
+
"step": 1808000
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 37.21,
|
| 2499 |
+
"eval_loss": 3.665800094604492,
|
| 2500 |
+
"eval_runtime": 46.839,
|
| 2501 |
+
"eval_samples_per_second": 877.517,
|
| 2502 |
+
"eval_steps_per_second": 54.847,
|
| 2503 |
+
"step": 1816000
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 37.37,
|
| 2507 |
+
"learning_rate": 2.4014008171433335e-06,
|
| 2508 |
+
"loss": 3.4651,
|
| 2509 |
+
"step": 1824000
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 37.37,
|
| 2513 |
+
"eval_loss": 3.638195037841797,
|
| 2514 |
+
"eval_runtime": 46.0815,
|
| 2515 |
+
"eval_samples_per_second": 891.942,
|
| 2516 |
+
"eval_steps_per_second": 55.749,
|
| 2517 |
+
"step": 1824000
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 37.53,
|
| 2521 |
+
"eval_loss": 3.656243085861206,
|
| 2522 |
+
"eval_runtime": 45.3257,
|
| 2523 |
+
"eval_samples_per_second": 906.815,
|
| 2524 |
+
"eval_steps_per_second": 56.679,
|
| 2525 |
+
"step": 1832000
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 37.7,
|
| 2529 |
+
"learning_rate": 2.334695238889352e-06,
|
| 2530 |
+
"loss": 3.4625,
|
| 2531 |
+
"step": 1840000
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 37.7,
|
| 2535 |
+
"eval_loss": 3.6376214027404785,
|
| 2536 |
+
"eval_runtime": 47.1734,
|
| 2537 |
+
"eval_samples_per_second": 871.296,
|
| 2538 |
+
"eval_steps_per_second": 54.459,
|
| 2539 |
+
"step": 1840000
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 37.86,
|
| 2543 |
+
"eval_loss": 3.651963710784912,
|
| 2544 |
+
"eval_runtime": 46.059,
|
| 2545 |
+
"eval_samples_per_second": 892.377,
|
| 2546 |
+
"eval_steps_per_second": 55.776,
|
| 2547 |
+
"step": 1848000
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 38.03,
|
| 2551 |
+
"learning_rate": 2.2679896606353707e-06,
|
| 2552 |
+
"loss": 3.4561,
|
| 2553 |
+
"step": 1856000
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 38.03,
|
| 2557 |
+
"eval_loss": 3.6300716400146484,
|
| 2558 |
+
"eval_runtime": 46.8158,
|
| 2559 |
+
"eval_samples_per_second": 877.951,
|
| 2560 |
+
"eval_steps_per_second": 54.875,
|
| 2561 |
+
"step": 1856000
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 38.19,
|
| 2565 |
+
"eval_loss": 3.619462728500366,
|
| 2566 |
+
"eval_runtime": 45.8596,
|
| 2567 |
+
"eval_samples_per_second": 896.258,
|
| 2568 |
+
"eval_steps_per_second": 56.019,
|
| 2569 |
+
"step": 1864000
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 38.35,
|
| 2573 |
+
"learning_rate": 2.2012840823813894e-06,
|
| 2574 |
+
"loss": 3.4655,
|
| 2575 |
+
"step": 1872000
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 38.35,
|
| 2579 |
+
"eval_loss": 3.6279447078704834,
|
| 2580 |
+
"eval_runtime": 46.2215,
|
| 2581 |
+
"eval_samples_per_second": 889.241,
|
| 2582 |
+
"eval_steps_per_second": 55.58,
|
| 2583 |
+
"step": 1872000
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 38.52,
|
| 2587 |
+
"eval_loss": 3.636460542678833,
|
| 2588 |
+
"eval_runtime": 46.7533,
|
| 2589 |
+
"eval_samples_per_second": 879.125,
|
| 2590 |
+
"eval_steps_per_second": 54.948,
|
| 2591 |
+
"step": 1880000
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 38.68,
|
| 2595 |
+
"learning_rate": 2.134578504127408e-06,
|
| 2596 |
+
"loss": 3.4637,
|
| 2597 |
+
"step": 1888000
|
| 2598 |
+
},
|
| 2599 |
+
{
|
| 2600 |
+
"epoch": 38.68,
|
| 2601 |
+
"eval_loss": 3.638620138168335,
|
| 2602 |
+
"eval_runtime": 46.2177,
|
| 2603 |
+
"eval_samples_per_second": 889.313,
|
| 2604 |
+
"eval_steps_per_second": 55.585,
|
| 2605 |
+
"step": 1888000
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 38.85,
|
| 2609 |
+
"eval_loss": 3.643373489379883,
|
| 2610 |
+
"eval_runtime": 45.9947,
|
| 2611 |
+
"eval_samples_per_second": 893.624,
|
| 2612 |
+
"eval_steps_per_second": 55.854,
|
| 2613 |
+
"step": 1896000
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 39.01,
|
| 2617 |
+
"learning_rate": 2.067872925873426e-06,
|
| 2618 |
+
"loss": 3.458,
|
| 2619 |
+
"step": 1904000
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 39.01,
|
| 2623 |
+
"eval_loss": 3.65189266204834,
|
| 2624 |
+
"eval_runtime": 46.7003,
|
| 2625 |
+
"eval_samples_per_second": 880.122,
|
| 2626 |
+
"eval_steps_per_second": 55.01,
|
| 2627 |
+
"step": 1904000
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 39.17,
|
| 2631 |
+
"eval_loss": 3.6438076496124268,
|
| 2632 |
+
"eval_runtime": 46.3785,
|
| 2633 |
+
"eval_samples_per_second": 886.229,
|
| 2634 |
+
"eval_steps_per_second": 55.392,
|
| 2635 |
+
"step": 1912000
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 39.34,
|
| 2639 |
+
"learning_rate": 2.0011673476194448e-06,
|
| 2640 |
+
"loss": 3.4523,
|
| 2641 |
+
"step": 1920000
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 39.34,
|
| 2645 |
+
"eval_loss": 3.640777349472046,
|
| 2646 |
+
"eval_runtime": 46.701,
|
| 2647 |
+
"eval_samples_per_second": 880.109,
|
| 2648 |
+
"eval_steps_per_second": 55.01,
|
| 2649 |
+
"step": 1920000
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 39.5,
|
| 2653 |
+
"eval_loss": 3.6513171195983887,
|
| 2654 |
+
"eval_runtime": 46.884,
|
| 2655 |
+
"eval_samples_per_second": 876.675,
|
| 2656 |
+
"eval_steps_per_second": 54.795,
|
| 2657 |
+
"step": 1928000
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 39.66,
|
| 2661 |
+
"learning_rate": 1.9344617693654634e-06,
|
| 2662 |
+
"loss": 3.4743,
|
| 2663 |
+
"step": 1936000
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 39.66,
|
| 2667 |
+
"eval_loss": 3.6177797317504883,
|
| 2668 |
+
"eval_runtime": 46.0686,
|
| 2669 |
+
"eval_samples_per_second": 892.192,
|
| 2670 |
+
"eval_steps_per_second": 55.765,
|
| 2671 |
+
"step": 1936000
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 39.83,
|
| 2675 |
+
"eval_loss": 3.6398518085479736,
|
| 2676 |
+
"eval_runtime": 46.8575,
|
| 2677 |
+
"eval_samples_per_second": 877.171,
|
| 2678 |
+
"eval_steps_per_second": 54.826,
|
| 2679 |
+
"step": 1944000
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 39.99,
|
| 2683 |
+
"learning_rate": 1.867756191111482e-06,
|
| 2684 |
+
"loss": 3.4626,
|
| 2685 |
+
"step": 1952000
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 39.99,
|
| 2689 |
+
"eval_loss": 3.624283790588379,
|
| 2690 |
+
"eval_runtime": 46.1682,
|
| 2691 |
+
"eval_samples_per_second": 890.266,
|
| 2692 |
+
"eval_steps_per_second": 55.644,
|
| 2693 |
+
"step": 1952000
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 40.16,
|
| 2697 |
+
"eval_loss": 3.6325714588165283,
|
| 2698 |
+
"eval_runtime": 45.9837,
|
| 2699 |
+
"eval_samples_per_second": 893.838,
|
| 2700 |
+
"eval_steps_per_second": 55.868,
|
| 2701 |
+
"step": 1960000
|
| 2702 |
+
},
|
| 2703 |
+
{
|
| 2704 |
+
"epoch": 40.32,
|
| 2705 |
+
"learning_rate": 1.8010506128575004e-06,
|
| 2706 |
+
"loss": 3.4692,
|
| 2707 |
+
"step": 1968000
|
| 2708 |
+
},
|
| 2709 |
+
{
|
| 2710 |
+
"epoch": 40.32,
|
| 2711 |
+
"eval_loss": 3.6723103523254395,
|
| 2712 |
+
"eval_runtime": 46.8787,
|
| 2713 |
+
"eval_samples_per_second": 876.773,
|
| 2714 |
+
"eval_steps_per_second": 54.801,
|
| 2715 |
+
"step": 1968000
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 40.48,
|
| 2719 |
+
"eval_loss": 3.6456410884857178,
|
| 2720 |
+
"eval_runtime": 46.0442,
|
| 2721 |
+
"eval_samples_per_second": 892.664,
|
| 2722 |
+
"eval_steps_per_second": 55.794,
|
| 2723 |
+
"step": 1976000
|
| 2724 |
+
},
|
| 2725 |
+
{
|
| 2726 |
+
"epoch": 40.65,
|
| 2727 |
+
"learning_rate": 1.734345034603519e-06,
|
| 2728 |
+
"loss": 3.4765,
|
| 2729 |
+
"step": 1984000
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 40.65,
|
| 2733 |
+
"eval_loss": 3.6437156200408936,
|
| 2734 |
+
"eval_runtime": 45.2826,
|
| 2735 |
+
"eval_samples_per_second": 907.678,
|
| 2736 |
+
"eval_steps_per_second": 56.733,
|
| 2737 |
+
"step": 1984000
|
| 2738 |
+
},
|
| 2739 |
+
{
|
| 2740 |
+
"epoch": 40.81,
|
| 2741 |
+
"eval_loss": 3.647704839706421,
|
| 2742 |
+
"eval_runtime": 46.8981,
|
| 2743 |
+
"eval_samples_per_second": 876.41,
|
| 2744 |
+
"eval_steps_per_second": 54.778,
|
| 2745 |
+
"step": 1992000
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 40.98,
|
| 2749 |
+
"learning_rate": 1.6676394563495374e-06,
|
| 2750 |
+
"loss": 3.4747,
|
| 2751 |
+
"step": 2000000
|
| 2752 |
+
},
|
| 2753 |
+
{
|
| 2754 |
+
"epoch": 40.98,
|
| 2755 |
+
"eval_loss": 3.638388156890869,
|
| 2756 |
+
"eval_runtime": 46.0328,
|
| 2757 |
+
"eval_samples_per_second": 892.886,
|
| 2758 |
+
"eval_steps_per_second": 55.808,
|
| 2759 |
+
"step": 2000000
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 41.14,
|
| 2763 |
+
"eval_loss": 3.6370368003845215,
|
| 2764 |
+
"eval_runtime": 46.7372,
|
| 2765 |
+
"eval_samples_per_second": 879.427,
|
| 2766 |
+
"eval_steps_per_second": 54.967,
|
| 2767 |
+
"step": 2008000
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 41.3,
|
| 2771 |
+
"learning_rate": 1.6009338780955558e-06,
|
| 2772 |
+
"loss": 3.4683,
|
| 2773 |
+
"step": 2016000
|
| 2774 |
+
},
|
| 2775 |
+
{
|
| 2776 |
+
"epoch": 41.3,
|
| 2777 |
+
"eval_loss": 3.662468433380127,
|
| 2778 |
+
"eval_runtime": 46.61,
|
| 2779 |
+
"eval_samples_per_second": 881.828,
|
| 2780 |
+
"eval_steps_per_second": 55.117,
|
| 2781 |
+
"step": 2016000
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 41.47,
|
| 2785 |
+
"eval_loss": 3.6453213691711426,
|
| 2786 |
+
"eval_runtime": 45.8611,
|
| 2787 |
+
"eval_samples_per_second": 896.229,
|
| 2788 |
+
"eval_steps_per_second": 56.017,
|
| 2789 |
+
"step": 2024000
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 41.63,
|
| 2793 |
+
"learning_rate": 1.5342282998415744e-06,
|
| 2794 |
+
"loss": 3.4599,
|
| 2795 |
+
"step": 2032000
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 41.63,
|
| 2799 |
+
"eval_loss": 3.64886212348938,
|
| 2800 |
+
"eval_runtime": 46.762,
|
| 2801 |
+
"eval_samples_per_second": 878.962,
|
| 2802 |
+
"eval_steps_per_second": 54.938,
|
| 2803 |
+
"step": 2032000
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 41.8,
|
| 2807 |
+
"eval_loss": 3.6310884952545166,
|
| 2808 |
+
"eval_runtime": 46.4576,
|
| 2809 |
+
"eval_samples_per_second": 884.72,
|
| 2810 |
+
"eval_steps_per_second": 55.298,
|
| 2811 |
+
"step": 2040000
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 41.96,
|
| 2815 |
+
"learning_rate": 1.4675227215875928e-06,
|
| 2816 |
+
"loss": 3.4713,
|
| 2817 |
+
"step": 2048000
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 41.96,
|
| 2821 |
+
"eval_loss": 3.619154691696167,
|
| 2822 |
+
"eval_runtime": 45.9184,
|
| 2823 |
+
"eval_samples_per_second": 895.109,
|
| 2824 |
+
"eval_steps_per_second": 55.947,
|
| 2825 |
+
"step": 2048000
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 42.12,
|
| 2829 |
+
"eval_loss": 3.651060104370117,
|
| 2830 |
+
"eval_runtime": 47.0032,
|
| 2831 |
+
"eval_samples_per_second": 874.451,
|
| 2832 |
+
"eval_steps_per_second": 54.656,
|
| 2833 |
+
"step": 2056000
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 42.29,
|
| 2837 |
+
"learning_rate": 1.4008171433336116e-06,
|
| 2838 |
+
"loss": 3.4677,
|
| 2839 |
+
"step": 2064000
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 42.29,
|
| 2843 |
+
"eval_loss": 3.6425869464874268,
|
| 2844 |
+
"eval_runtime": 46.3503,
|
| 2845 |
+
"eval_samples_per_second": 886.769,
|
| 2846 |
+
"eval_steps_per_second": 55.426,
|
| 2847 |
+
"step": 2064000
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 42.45,
|
| 2851 |
+
"eval_loss": 3.6362836360931396,
|
| 2852 |
+
"eval_runtime": 46.2845,
|
| 2853 |
+
"eval_samples_per_second": 888.029,
|
| 2854 |
+
"eval_steps_per_second": 55.505,
|
| 2855 |
+
"step": 2072000
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 42.62,
|
| 2859 |
+
"learning_rate": 1.33411156507963e-06,
|
| 2860 |
+
"loss": 3.4689,
|
| 2861 |
+
"step": 2080000
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 42.62,
|
| 2865 |
+
"eval_loss": 3.6378438472747803,
|
| 2866 |
+
"eval_runtime": 47.0132,
|
| 2867 |
+
"eval_samples_per_second": 874.265,
|
| 2868 |
+
"eval_steps_per_second": 54.644,
|
| 2869 |
+
"step": 2080000
|
| 2870 |
+
},
|
| 2871 |
+
{
|
| 2872 |
+
"epoch": 42.78,
|
| 2873 |
+
"eval_loss": 3.6450445652008057,
|
| 2874 |
+
"eval_runtime": 46.1055,
|
| 2875 |
+
"eval_samples_per_second": 891.478,
|
| 2876 |
+
"eval_steps_per_second": 55.72,
|
| 2877 |
+
"step": 2088000
|
| 2878 |
+
},
|
| 2879 |
+
{
|
| 2880 |
+
"epoch": 42.94,
|
| 2881 |
+
"learning_rate": 1.2674059868256484e-06,
|
| 2882 |
+
"loss": 3.4598,
|
| 2883 |
+
"step": 2096000
|
| 2884 |
+
},
|
| 2885 |
+
{
|
| 2886 |
+
"epoch": 42.94,
|
| 2887 |
+
"eval_loss": 3.64805006980896,
|
| 2888 |
+
"eval_runtime": 46.8684,
|
| 2889 |
+
"eval_samples_per_second": 876.967,
|
| 2890 |
+
"eval_steps_per_second": 54.813,
|
| 2891 |
+
"step": 2096000
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 43.11,
|
| 2895 |
+
"eval_loss": 3.6675028800964355,
|
| 2896 |
+
"eval_runtime": 46.4765,
|
| 2897 |
+
"eval_samples_per_second": 884.36,
|
| 2898 |
+
"eval_steps_per_second": 55.275,
|
| 2899 |
+
"step": 2104000
|
| 2900 |
+
},
|
| 2901 |
+
{
|
| 2902 |
+
"epoch": 43.27,
|
| 2903 |
+
"learning_rate": 1.2007004085716668e-06,
|
| 2904 |
+
"loss": 3.4487,
|
| 2905 |
+
"step": 2112000
|
| 2906 |
+
},
|
| 2907 |
+
{
|
| 2908 |
+
"epoch": 43.27,
|
| 2909 |
+
"eval_loss": 3.6557657718658447,
|
| 2910 |
+
"eval_runtime": 46.0356,
|
| 2911 |
+
"eval_samples_per_second": 892.83,
|
| 2912 |
+
"eval_steps_per_second": 55.805,
|
| 2913 |
+
"step": 2112000
|
| 2914 |
+
},
|
| 2915 |
+
{
|
| 2916 |
+
"epoch": 43.43,
|
| 2917 |
+
"eval_loss": 3.6451427936553955,
|
| 2918 |
+
"eval_runtime": 47.3121,
|
| 2919 |
+
"eval_samples_per_second": 868.741,
|
| 2920 |
+
"eval_steps_per_second": 54.299,
|
| 2921 |
+
"step": 2120000
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 43.6,
|
| 2925 |
+
"learning_rate": 1.1339948303176854e-06,
|
| 2926 |
+
"loss": 3.4555,
|
| 2927 |
+
"step": 2128000
|
| 2928 |
+
},
|
| 2929 |
+
{
|
| 2930 |
+
"epoch": 43.6,
|
| 2931 |
+
"eval_loss": 3.643132448196411,
|
| 2932 |
+
"eval_runtime": 46.2499,
|
| 2933 |
+
"eval_samples_per_second": 888.694,
|
| 2934 |
+
"eval_steps_per_second": 55.546,
|
| 2935 |
+
"step": 2128000
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 43.76,
|
| 2939 |
+
"eval_loss": 3.6470389366149902,
|
| 2940 |
+
"eval_runtime": 45.8331,
|
| 2941 |
+
"eval_samples_per_second": 896.776,
|
| 2942 |
+
"eval_steps_per_second": 56.051,
|
| 2943 |
+
"step": 2136000
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 43.93,
|
| 2947 |
+
"learning_rate": 1.067289252063704e-06,
|
| 2948 |
+
"loss": 3.4727,
|
| 2949 |
+
"step": 2144000
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 43.93,
|
| 2953 |
+
"eval_loss": 3.6265406608581543,
|
| 2954 |
+
"eval_runtime": 47.1162,
|
| 2955 |
+
"eval_samples_per_second": 872.353,
|
| 2956 |
+
"eval_steps_per_second": 54.525,
|
| 2957 |
+
"step": 2144000
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 44.09,
|
| 2961 |
+
"eval_loss": 3.6335132122039795,
|
| 2962 |
+
"eval_runtime": 45.9499,
|
| 2963 |
+
"eval_samples_per_second": 894.497,
|
| 2964 |
+
"eval_steps_per_second": 55.909,
|
| 2965 |
+
"step": 2152000
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 44.25,
|
| 2969 |
+
"learning_rate": 1.0005836738097224e-06,
|
| 2970 |
+
"loss": 3.4626,
|
| 2971 |
+
"step": 2160000
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 44.25,
|
| 2975 |
+
"eval_loss": 3.639557123184204,
|
| 2976 |
+
"eval_runtime": 46.75,
|
| 2977 |
+
"eval_samples_per_second": 879.187,
|
| 2978 |
+
"eval_steps_per_second": 54.952,
|
| 2979 |
+
"step": 2160000
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 44.42,
|
| 2983 |
+
"eval_loss": 3.653687000274658,
|
| 2984 |
+
"eval_runtime": 47.165,
|
| 2985 |
+
"eval_samples_per_second": 871.452,
|
| 2986 |
+
"eval_steps_per_second": 54.468,
|
| 2987 |
+
"step": 2168000
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 44.58,
|
| 2991 |
+
"learning_rate": 9.33878095555741e-07,
|
| 2992 |
+
"loss": 3.4724,
|
| 2993 |
+
"step": 2176000
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 44.58,
|
| 2997 |
+
"eval_loss": 3.61678409576416,
|
| 2998 |
+
"eval_runtime": 46.2585,
|
| 2999 |
+
"eval_samples_per_second": 888.528,
|
| 3000 |
+
"eval_steps_per_second": 55.536,
|
| 3001 |
+
"step": 2176000
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 44.75,
|
| 3005 |
+
"eval_loss": 3.644352674484253,
|
| 3006 |
+
"eval_runtime": 47.0469,
|
| 3007 |
+
"eval_samples_per_second": 873.64,
|
| 3008 |
+
"eval_steps_per_second": 54.605,
|
| 3009 |
+
"step": 2184000
|
| 3010 |
+
},
|
| 3011 |
+
{
|
| 3012 |
+
"epoch": 44.91,
|
| 3013 |
+
"learning_rate": 8.671725173017595e-07,
|
| 3014 |
+
"loss": 3.4545,
|
| 3015 |
+
"step": 2192000
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 44.91,
|
| 3019 |
+
"eval_loss": 3.6440114974975586,
|
| 3020 |
+
"eval_runtime": 46.2426,
|
| 3021 |
+
"eval_samples_per_second": 888.835,
|
| 3022 |
+
"eval_steps_per_second": 55.555,
|
| 3023 |
+
"step": 2192000
|
| 3024 |
+
},
|
| 3025 |
+
{
|
| 3026 |
+
"epoch": 45.07,
|
| 3027 |
+
"eval_loss": 3.6327061653137207,
|
| 3028 |
+
"eval_runtime": 46.09,
|
| 3029 |
+
"eval_samples_per_second": 891.776,
|
| 3030 |
+
"eval_steps_per_second": 55.739,
|
| 3031 |
+
"step": 2200000
|
| 3032 |
+
},
|
| 3033 |
+
{
|
| 3034 |
+
"epoch": 45.24,
|
| 3035 |
+
"learning_rate": 8.004669390477779e-07,
|
| 3036 |
+
"loss": 3.461,
|
| 3037 |
+
"step": 2208000
|
| 3038 |
+
},
|
| 3039 |
+
{
|
| 3040 |
+
"epoch": 45.24,
|
| 3041 |
+
"eval_loss": 3.6362533569335938,
|
| 3042 |
+
"eval_runtime": 47.1445,
|
| 3043 |
+
"eval_samples_per_second": 871.831,
|
| 3044 |
+
"eval_steps_per_second": 54.492,
|
| 3045 |
+
"step": 2208000
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 45.4,
|
| 3049 |
+
"eval_loss": 3.653747081756592,
|
| 3050 |
+
"eval_runtime": 46.2235,
|
| 3051 |
+
"eval_samples_per_second": 889.202,
|
| 3052 |
+
"eval_steps_per_second": 55.578,
|
| 3053 |
+
"step": 2216000
|
| 3054 |
+
},
|
| 3055 |
+
{
|
| 3056 |
+
"epoch": 45.57,
|
| 3057 |
+
"learning_rate": 7.337613607937964e-07,
|
| 3058 |
+
"loss": 3.4702,
|
| 3059 |
+
"step": 2224000
|
| 3060 |
+
},
|
| 3061 |
+
{
|
| 3062 |
+
"epoch": 45.57,
|
| 3063 |
+
"eval_loss": 3.6123247146606445,
|
| 3064 |
+
"eval_runtime": 46.168,
|
| 3065 |
+
"eval_samples_per_second": 890.27,
|
| 3066 |
+
"eval_steps_per_second": 55.645,
|
| 3067 |
+
"step": 2224000
|
| 3068 |
+
},
|
| 3069 |
+
{
|
| 3070 |
+
"epoch": 45.73,
|
| 3071 |
+
"eval_loss": 3.6554455757141113,
|
| 3072 |
+
"eval_runtime": 47.1193,
|
| 3073 |
+
"eval_samples_per_second": 872.296,
|
| 3074 |
+
"eval_steps_per_second": 54.521,
|
| 3075 |
+
"step": 2232000
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 45.89,
|
| 3079 |
+
"learning_rate": 6.67055782539815e-07,
|
| 3080 |
+
"loss": 3.4565,
|
| 3081 |
+
"step": 2240000
|
| 3082 |
+
},
|
| 3083 |
+
{
|
| 3084 |
+
"epoch": 45.89,
|
| 3085 |
+
"eval_loss": 3.6522979736328125,
|
| 3086 |
+
"eval_runtime": 46.1449,
|
| 3087 |
+
"eval_samples_per_second": 890.716,
|
| 3088 |
+
"eval_steps_per_second": 55.672,
|
| 3089 |
+
"step": 2240000
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 46.06,
|
| 3093 |
+
"eval_loss": 3.6339659690856934,
|
| 3094 |
+
"eval_runtime": 47.2579,
|
| 3095 |
+
"eval_samples_per_second": 869.739,
|
| 3096 |
+
"eval_steps_per_second": 54.361,
|
| 3097 |
+
"step": 2248000
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 46.22,
|
| 3101 |
+
"learning_rate": 6.003502042858334e-07,
|
| 3102 |
+
"loss": 3.4517,
|
| 3103 |
+
"step": 2256000
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 46.22,
|
| 3107 |
+
"eval_loss": 3.6459498405456543,
|
| 3108 |
+
"eval_runtime": 46.9038,
|
| 3109 |
+
"eval_samples_per_second": 876.305,
|
| 3110 |
+
"eval_steps_per_second": 54.772,
|
| 3111 |
+
"step": 2256000
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 46.38,
|
| 3115 |
+
"eval_loss": 3.656141996383667,
|
| 3116 |
+
"eval_runtime": 46.3654,
|
| 3117 |
+
"eval_samples_per_second": 886.48,
|
| 3118 |
+
"eval_steps_per_second": 55.408,
|
| 3119 |
+
"step": 2264000
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 46.55,
|
| 3123 |
+
"learning_rate": 5.33644626031852e-07,
|
| 3124 |
+
"loss": 3.4631,
|
| 3125 |
+
"step": 2272000
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 46.55,
|
| 3129 |
+
"eval_loss": 3.6547927856445312,
|
| 3130 |
+
"eval_runtime": 47.1154,
|
| 3131 |
+
"eval_samples_per_second": 872.368,
|
| 3132 |
+
"eval_steps_per_second": 54.526,
|
| 3133 |
+
"step": 2272000
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 46.71,
|
| 3137 |
+
"eval_loss": 3.6228716373443604,
|
| 3138 |
+
"eval_runtime": 46.2908,
|
| 3139 |
+
"eval_samples_per_second": 887.908,
|
| 3140 |
+
"eval_steps_per_second": 55.497,
|
| 3141 |
+
"step": 2280000
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 46.88,
|
| 3145 |
+
"learning_rate": 4.669390477778705e-07,
|
| 3146 |
+
"loss": 3.4518,
|
| 3147 |
+
"step": 2288000
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 46.88,
|
| 3151 |
+
"eval_loss": 3.6350128650665283,
|
| 3152 |
+
"eval_runtime": 46.3584,
|
| 3153 |
+
"eval_samples_per_second": 886.613,
|
| 3154 |
+
"eval_steps_per_second": 55.416,
|
| 3155 |
+
"step": 2288000
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 47.04,
|
| 3159 |
+
"eval_loss": 3.6483192443847656,
|
| 3160 |
+
"eval_runtime": 47.24,
|
| 3161 |
+
"eval_samples_per_second": 870.067,
|
| 3162 |
+
"eval_steps_per_second": 54.382,
|
| 3163 |
+
"step": 2296000
|
| 3164 |
+
},
|
| 3165 |
+
{
|
| 3166 |
+
"epoch": 47.2,
|
| 3167 |
+
"learning_rate": 4.0023346952388894e-07,
|
| 3168 |
+
"loss": 3.4592,
|
| 3169 |
+
"step": 2304000
|
| 3170 |
+
},
|
| 3171 |
+
{
|
| 3172 |
+
"epoch": 47.2,
|
| 3173 |
+
"eval_loss": 3.6263089179992676,
|
| 3174 |
+
"eval_runtime": 47.0185,
|
| 3175 |
+
"eval_samples_per_second": 874.166,
|
| 3176 |
+
"eval_steps_per_second": 54.638,
|
| 3177 |
+
"step": 2304000
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 47.37,
|
| 3181 |
+
"eval_loss": 3.6339097023010254,
|
| 3182 |
+
"eval_runtime": 46.0199,
|
| 3183 |
+
"eval_samples_per_second": 893.135,
|
| 3184 |
+
"eval_steps_per_second": 55.824,
|
| 3185 |
+
"step": 2312000
|
| 3186 |
+
},
|
| 3187 |
+
{
|
| 3188 |
+
"epoch": 47.53,
|
| 3189 |
+
"learning_rate": 3.335278912699075e-07,
|
| 3190 |
+
"loss": 3.4569,
|
| 3191 |
+
"step": 2320000
|
| 3192 |
+
},
|
| 3193 |
+
{
|
| 3194 |
+
"epoch": 47.53,
|
| 3195 |
+
"eval_loss": 3.659444808959961,
|
| 3196 |
+
"eval_runtime": 47.1636,
|
| 3197 |
+
"eval_samples_per_second": 871.477,
|
| 3198 |
+
"eval_steps_per_second": 54.47,
|
| 3199 |
+
"step": 2320000
|
| 3200 |
+
},
|
| 3201 |
+
{
|
| 3202 |
+
"epoch": 47.7,
|
| 3203 |
+
"eval_loss": 3.638535737991333,
|
| 3204 |
+
"eval_runtime": 46.1693,
|
| 3205 |
+
"eval_samples_per_second": 890.246,
|
| 3206 |
+
"eval_steps_per_second": 55.643,
|
| 3207 |
+
"step": 2328000
|
| 3208 |
+
},
|
| 3209 |
+
{
|
| 3210 |
+
"epoch": 47.86,
|
| 3211 |
+
"learning_rate": 2.66822313015926e-07,
|
| 3212 |
+
"loss": 3.4524,
|
| 3213 |
+
"step": 2336000
|
| 3214 |
+
},
|
| 3215 |
+
{
|
| 3216 |
+
"epoch": 47.86,
|
| 3217 |
+
"eval_loss": 3.6434078216552734,
|
| 3218 |
+
"eval_runtime": 47.0318,
|
| 3219 |
+
"eval_samples_per_second": 873.919,
|
| 3220 |
+
"eval_steps_per_second": 54.623,
|
| 3221 |
+
"step": 2336000
|
| 3222 |
+
},
|
| 3223 |
+
{
|
| 3224 |
+
"epoch": 48.02,
|
| 3225 |
+
"eval_loss": 3.650230646133423,
|
| 3226 |
+
"eval_runtime": 46.5514,
|
| 3227 |
+
"eval_samples_per_second": 882.938,
|
| 3228 |
+
"eval_steps_per_second": 55.186,
|
| 3229 |
+
"step": 2344000
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 48.19,
|
| 3233 |
+
"learning_rate": 2.0011673476194447e-07,
|
| 3234 |
+
"loss": 3.4644,
|
| 3235 |
+
"step": 2352000
|
| 3236 |
+
},
|
| 3237 |
+
{
|
| 3238 |
+
"epoch": 48.19,
|
| 3239 |
+
"eval_loss": 3.617619276046753,
|
| 3240 |
+
"eval_runtime": 46.2116,
|
| 3241 |
+
"eval_samples_per_second": 889.43,
|
| 3242 |
+
"eval_steps_per_second": 55.592,
|
| 3243 |
+
"step": 2352000
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 48.35,
|
| 3247 |
+
"eval_loss": 3.6293184757232666,
|
| 3248 |
+
"eval_runtime": 47.399,
|
| 3249 |
+
"eval_samples_per_second": 867.15,
|
| 3250 |
+
"eval_steps_per_second": 54.199,
|
| 3251 |
+
"step": 2360000
|
| 3252 |
+
},
|
| 3253 |
+
{
|
| 3254 |
+
"epoch": 48.52,
|
| 3255 |
+
"learning_rate": 1.33411156507963e-07,
|
| 3256 |
+
"loss": 3.4586,
|
| 3257 |
+
"step": 2368000
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 48.52,
|
| 3261 |
+
"eval_loss": 3.630380392074585,
|
| 3262 |
+
"eval_runtime": 46.3912,
|
| 3263 |
+
"eval_samples_per_second": 885.987,
|
| 3264 |
+
"eval_steps_per_second": 55.377,
|
| 3265 |
+
"step": 2368000
|
| 3266 |
+
},
|
| 3267 |
+
{
|
| 3268 |
+
"epoch": 48.68,
|
| 3269 |
+
"eval_loss": 3.6343326568603516,
|
| 3270 |
+
"eval_runtime": 46.2144,
|
| 3271 |
+
"eval_samples_per_second": 889.376,
|
| 3272 |
+
"eval_steps_per_second": 55.589,
|
| 3273 |
+
"step": 2376000
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 48.84,
|
| 3277 |
+
"learning_rate": 6.67055782539815e-08,
|
| 3278 |
+
"loss": 3.4439,
|
| 3279 |
+
"step": 2384000
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 48.84,
|
| 3283 |
+
"eval_loss": 3.6090333461761475,
|
| 3284 |
+
"eval_runtime": 47.3482,
|
| 3285 |
+
"eval_samples_per_second": 868.08,
|
| 3286 |
+
"eval_steps_per_second": 54.258,
|
| 3287 |
+
"step": 2384000
|
| 3288 |
+
},
|
| 3289 |
+
{
|
| 3290 |
+
"epoch": 49.01,
|
| 3291 |
+
"eval_loss": 3.6414153575897217,
|
| 3292 |
+
"eval_runtime": 46.5994,
|
| 3293 |
+
"eval_samples_per_second": 882.029,
|
| 3294 |
+
"eval_steps_per_second": 55.13,
|
| 3295 |
+
"step": 2392000
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 49.17,
|
| 3299 |
+
"learning_rate": 0.0,
|
| 3300 |
+
"loss": 3.4474,
|
| 3301 |
+
"step": 2400000
|
| 3302 |
+
},
|
| 3303 |
+
{
|
| 3304 |
+
"epoch": 49.17,
|
| 3305 |
+
"eval_loss": 3.620838165283203,
|
| 3306 |
+
"eval_runtime": 46.9825,
|
| 3307 |
+
"eval_samples_per_second": 874.835,
|
| 3308 |
+
"eval_steps_per_second": 54.68,
|
| 3309 |
+
"step": 2400000
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 49.17,
|
| 3313 |
+
"step": 2400000,
|
| 3314 |
+
"total_flos": 6.906141294629226e+17,
|
| 3315 |
+
"train_loss": 3.376089767252604,
|
| 3316 |
+
"train_runtime": 158003.2062,
|
| 3317 |
+
"train_samples_per_second": 243.033,
|
| 3318 |
+
"train_steps_per_second": 15.19
|
| 3319 |
+
}
|
| 3320 |
+
],
|
| 3321 |
+
"logging_steps": 16000,
|
| 3322 |
+
"max_steps": 2400000,
|
| 3323 |
+
"num_train_epochs": 50,
|
| 3324 |
+
"save_steps": 32000,
|
| 3325 |
+
"total_flos": 6.906141294629226e+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 sha256:015483c8d7a8f28e3cfc7685d3435adc6087ae5ef117bbe1f324d578dc554544
|
| 3 |
+
size 4219
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|