Uploaded PeptideGPT non-fouling model
Browse files- README.md +259 -0
- all_results.json +15 -0
- config.json +39 -0
- eval_results.json +10 -0
- events.out.tfevents.1708315853.lambda-a6000.2440433.0 +3 -0
- generation_config.json +6 -0
- merges.txt +0 -0
- model.safetensors +3 -0
- special_tokens_map.json +23 -0
- tokenizer.json +0 -0
- tokenizer_config.json +19 -0
- train_results.json +8 -0
- trainer_state.json +3030 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
ADDED
|
@@ -0,0 +1,259 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: nferruz/ProtGPT2
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
metrics:
|
| 7 |
+
- accuracy
|
| 8 |
+
model-index:
|
| 9 |
+
- name: output_nf_3
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# output_nf_3
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [nferruz/ProtGPT2](https://huggingface.co/nferruz/ProtGPT2) on an unknown dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 5.6642
|
| 21 |
+
- Accuracy: 0.3099
|
| 22 |
+
|
| 23 |
+
## Model description
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Intended uses & limitations
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training and evaluation data
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Training procedure
|
| 36 |
+
|
| 37 |
+
### Training hyperparameters
|
| 38 |
+
|
| 39 |
+
The following hyperparameters were used during training:
|
| 40 |
+
- learning_rate: 1e-06
|
| 41 |
+
- train_batch_size: 1
|
| 42 |
+
- eval_batch_size: 8
|
| 43 |
+
- seed: 42
|
| 44 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
+
- lr_scheduler_type: linear
|
| 46 |
+
- num_epochs: 200.0
|
| 47 |
+
|
| 48 |
+
### Training results
|
| 49 |
+
|
| 50 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 51 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 52 |
+
| 12.1762 | 1.0 | 9 | 11.8181 | 0.0029 |
|
| 53 |
+
| 11.6538 | 2.0 | 18 | 11.3468 | 0.0010 |
|
| 54 |
+
| 11.2876 | 3.0 | 27 | 10.9127 | 0.0010 |
|
| 55 |
+
| 10.8664 | 4.0 | 36 | 10.5162 | 0.0010 |
|
| 56 |
+
| 10.6184 | 5.0 | 45 | 10.1817 | 0.0020 |
|
| 57 |
+
| 10.3132 | 6.0 | 54 | 9.8765 | 0.0010 |
|
| 58 |
+
| 9.9087 | 7.0 | 63 | 9.6073 | 0.0010 |
|
| 59 |
+
| 9.6953 | 8.0 | 72 | 9.3244 | 0.0 |
|
| 60 |
+
| 9.3741 | 9.0 | 81 | 9.0318 | 0.0 |
|
| 61 |
+
| 9.2045 | 10.0 | 90 | 8.7461 | 0.0 |
|
| 62 |
+
| 8.9079 | 11.0 | 99 | 8.4755 | 0.0020 |
|
| 63 |
+
| 8.7047 | 12.0 | 108 | 8.2222 | 0.0117 |
|
| 64 |
+
| 8.4622 | 13.0 | 117 | 7.9951 | 0.0313 |
|
| 65 |
+
| 8.2649 | 14.0 | 126 | 7.8052 | 0.0655 |
|
| 66 |
+
| 8.043 | 15.0 | 135 | 7.6495 | 0.1065 |
|
| 67 |
+
| 7.9092 | 16.0 | 144 | 7.5160 | 0.1378 |
|
| 68 |
+
| 7.7103 | 17.0 | 153 | 7.3881 | 0.1691 |
|
| 69 |
+
| 7.5701 | 18.0 | 162 | 7.2693 | 0.1818 |
|
| 70 |
+
| 7.4483 | 19.0 | 171 | 7.1544 | 0.1994 |
|
| 71 |
+
| 7.314 | 20.0 | 180 | 7.0469 | 0.2160 |
|
| 72 |
+
| 7.2101 | 21.0 | 189 | 6.9440 | 0.2326 |
|
| 73 |
+
| 7.1026 | 22.0 | 198 | 6.8542 | 0.2366 |
|
| 74 |
+
| 6.9954 | 23.0 | 207 | 6.7684 | 0.2483 |
|
| 75 |
+
| 6.9206 | 24.0 | 216 | 6.6851 | 0.2512 |
|
| 76 |
+
| 6.8588 | 25.0 | 225 | 6.6217 | 0.2561 |
|
| 77 |
+
| 6.7975 | 26.0 | 234 | 6.5621 | 0.2590 |
|
| 78 |
+
| 6.7355 | 27.0 | 243 | 6.5111 | 0.2610 |
|
| 79 |
+
| 6.6928 | 28.0 | 252 | 6.4630 | 0.2639 |
|
| 80 |
+
| 6.6483 | 29.0 | 261 | 6.4137 | 0.2698 |
|
| 81 |
+
| 6.6169 | 30.0 | 270 | 6.3675 | 0.2717 |
|
| 82 |
+
| 6.5498 | 31.0 | 279 | 6.3292 | 0.2747 |
|
| 83 |
+
| 6.5385 | 32.0 | 288 | 6.2967 | 0.2766 |
|
| 84 |
+
| 6.5111 | 33.0 | 297 | 6.2592 | 0.2786 |
|
| 85 |
+
| 6.4748 | 34.0 | 306 | 6.2316 | 0.2815 |
|
| 86 |
+
| 6.4575 | 35.0 | 315 | 6.2096 | 0.2835 |
|
| 87 |
+
| 6.4251 | 36.0 | 324 | 6.1847 | 0.2845 |
|
| 88 |
+
| 6.4096 | 37.0 | 333 | 6.1613 | 0.2854 |
|
| 89 |
+
| 6.3741 | 38.0 | 342 | 6.1402 | 0.2854 |
|
| 90 |
+
| 6.3645 | 39.0 | 351 | 6.1285 | 0.2874 |
|
| 91 |
+
| 6.3511 | 40.0 | 360 | 6.1134 | 0.2874 |
|
| 92 |
+
| 6.3254 | 41.0 | 369 | 6.0957 | 0.2874 |
|
| 93 |
+
| 6.3077 | 42.0 | 378 | 6.0818 | 0.2874 |
|
| 94 |
+
| 6.301 | 43.0 | 387 | 6.0688 | 0.2874 |
|
| 95 |
+
| 6.2846 | 44.0 | 396 | 6.0511 | 0.2884 |
|
| 96 |
+
| 6.2739 | 45.0 | 405 | 6.0398 | 0.2893 |
|
| 97 |
+
| 6.2569 | 46.0 | 414 | 6.0301 | 0.2913 |
|
| 98 |
+
| 6.258 | 47.0 | 423 | 6.0176 | 0.2913 |
|
| 99 |
+
| 6.2273 | 48.0 | 432 | 6.0037 | 0.2933 |
|
| 100 |
+
| 6.2256 | 49.0 | 441 | 5.9962 | 0.2933 |
|
| 101 |
+
| 6.2065 | 50.0 | 450 | 5.9869 | 0.2933 |
|
| 102 |
+
| 6.1991 | 51.0 | 459 | 5.9756 | 0.2933 |
|
| 103 |
+
| 6.1895 | 52.0 | 468 | 5.9682 | 0.2942 |
|
| 104 |
+
| 6.1763 | 53.0 | 477 | 5.9611 | 0.2942 |
|
| 105 |
+
| 6.1734 | 54.0 | 486 | 5.9535 | 0.2942 |
|
| 106 |
+
| 6.1702 | 55.0 | 495 | 5.9457 | 0.2942 |
|
| 107 |
+
| 6.1556 | 56.0 | 504 | 5.9376 | 0.2952 |
|
| 108 |
+
| 6.1481 | 57.0 | 513 | 5.9306 | 0.2952 |
|
| 109 |
+
| 6.1425 | 58.0 | 522 | 5.9222 | 0.2942 |
|
| 110 |
+
| 6.1416 | 59.0 | 531 | 5.9170 | 0.2942 |
|
| 111 |
+
| 6.1328 | 60.0 | 540 | 5.9142 | 0.2942 |
|
| 112 |
+
| 6.1176 | 61.0 | 549 | 5.9064 | 0.2942 |
|
| 113 |
+
| 6.1091 | 62.0 | 558 | 5.9010 | 0.2942 |
|
| 114 |
+
| 6.104 | 63.0 | 567 | 5.8961 | 0.2942 |
|
| 115 |
+
| 6.0986 | 64.0 | 576 | 5.8915 | 0.2942 |
|
| 116 |
+
| 6.089 | 65.0 | 585 | 5.8854 | 0.2952 |
|
| 117 |
+
| 6.0734 | 66.0 | 594 | 5.8810 | 0.2942 |
|
| 118 |
+
| 6.0905 | 67.0 | 603 | 5.8762 | 0.2942 |
|
| 119 |
+
| 6.0701 | 68.0 | 612 | 5.8737 | 0.2962 |
|
| 120 |
+
| 6.0595 | 69.0 | 621 | 5.8694 | 0.2962 |
|
| 121 |
+
| 6.0616 | 70.0 | 630 | 5.8661 | 0.2962 |
|
| 122 |
+
| 6.0512 | 71.0 | 639 | 5.8635 | 0.2962 |
|
| 123 |
+
| 6.0415 | 72.0 | 648 | 5.8613 | 0.2962 |
|
| 124 |
+
| 6.0391 | 73.0 | 657 | 5.8583 | 0.2962 |
|
| 125 |
+
| 6.032 | 74.0 | 666 | 5.8548 | 0.2962 |
|
| 126 |
+
| 6.0317 | 75.0 | 675 | 5.8511 | 0.2962 |
|
| 127 |
+
| 6.0343 | 76.0 | 684 | 5.8479 | 0.2972 |
|
| 128 |
+
| 6.0156 | 77.0 | 693 | 5.8435 | 0.2981 |
|
| 129 |
+
| 6.0167 | 78.0 | 702 | 5.8403 | 0.2981 |
|
| 130 |
+
| 6.0052 | 79.0 | 711 | 5.8364 | 0.2981 |
|
| 131 |
+
| 6.0057 | 80.0 | 720 | 5.8336 | 0.2981 |
|
| 132 |
+
| 6.0001 | 81.0 | 729 | 5.8319 | 0.2981 |
|
| 133 |
+
| 6.001 | 82.0 | 738 | 5.8288 | 0.2991 |
|
| 134 |
+
| 5.9905 | 83.0 | 747 | 5.8266 | 0.3011 |
|
| 135 |
+
| 5.9906 | 84.0 | 756 | 5.8242 | 0.3021 |
|
| 136 |
+
| 5.9862 | 85.0 | 765 | 5.8216 | 0.3021 |
|
| 137 |
+
| 5.9829 | 86.0 | 774 | 5.8191 | 0.3050 |
|
| 138 |
+
| 5.9725 | 87.0 | 783 | 5.8173 | 0.3069 |
|
| 139 |
+
| 5.9795 | 88.0 | 792 | 5.8142 | 0.3089 |
|
| 140 |
+
| 5.9695 | 89.0 | 801 | 5.8115 | 0.3069 |
|
| 141 |
+
| 5.9607 | 90.0 | 810 | 5.8081 | 0.3089 |
|
| 142 |
+
| 5.9605 | 91.0 | 819 | 5.8057 | 0.3089 |
|
| 143 |
+
| 5.9591 | 92.0 | 828 | 5.8037 | 0.3079 |
|
| 144 |
+
| 5.9481 | 93.0 | 837 | 5.8011 | 0.3069 |
|
| 145 |
+
| 5.9501 | 94.0 | 846 | 5.7983 | 0.3079 |
|
| 146 |
+
| 5.948 | 95.0 | 855 | 5.7943 | 0.3089 |
|
| 147 |
+
| 5.9488 | 96.0 | 864 | 5.7907 | 0.3109 |
|
| 148 |
+
| 5.9449 | 97.0 | 873 | 5.7888 | 0.3118 |
|
| 149 |
+
| 5.9357 | 98.0 | 882 | 5.7872 | 0.3099 |
|
| 150 |
+
| 5.9363 | 99.0 | 891 | 5.7834 | 0.3079 |
|
| 151 |
+
| 5.9368 | 100.0 | 900 | 5.7817 | 0.3099 |
|
| 152 |
+
| 5.9215 | 101.0 | 909 | 5.7791 | 0.3099 |
|
| 153 |
+
| 5.9264 | 102.0 | 918 | 5.7749 | 0.3099 |
|
| 154 |
+
| 5.9126 | 103.0 | 927 | 5.7725 | 0.3089 |
|
| 155 |
+
| 5.9163 | 104.0 | 936 | 5.7698 | 0.3069 |
|
| 156 |
+
| 5.9183 | 105.0 | 945 | 5.7673 | 0.3069 |
|
| 157 |
+
| 5.9154 | 106.0 | 954 | 5.7639 | 0.3089 |
|
| 158 |
+
| 5.9149 | 107.0 | 963 | 5.7612 | 0.3079 |
|
| 159 |
+
| 5.9082 | 108.0 | 972 | 5.7579 | 0.3089 |
|
| 160 |
+
| 5.9051 | 109.0 | 981 | 5.7559 | 0.3079 |
|
| 161 |
+
| 5.908 | 110.0 | 990 | 5.7531 | 0.3079 |
|
| 162 |
+
| 5.901 | 111.0 | 999 | 5.7502 | 0.3079 |
|
| 163 |
+
| 5.9064 | 112.0 | 1008 | 5.7484 | 0.3089 |
|
| 164 |
+
| 5.895 | 113.0 | 1017 | 5.7460 | 0.3099 |
|
| 165 |
+
| 5.894 | 114.0 | 1026 | 5.7438 | 0.3089 |
|
| 166 |
+
| 5.8809 | 115.0 | 1035 | 5.7419 | 0.3079 |
|
| 167 |
+
| 5.8893 | 116.0 | 1044 | 5.7419 | 0.3079 |
|
| 168 |
+
| 5.8874 | 117.0 | 1053 | 5.7392 | 0.3069 |
|
| 169 |
+
| 5.8798 | 118.0 | 1062 | 5.7349 | 0.3089 |
|
| 170 |
+
| 5.8826 | 119.0 | 1071 | 5.7324 | 0.3099 |
|
| 171 |
+
| 5.8736 | 120.0 | 1080 | 5.7299 | 0.3099 |
|
| 172 |
+
| 5.8751 | 121.0 | 1089 | 5.7270 | 0.3099 |
|
| 173 |
+
| 5.8699 | 122.0 | 1098 | 5.7253 | 0.3099 |
|
| 174 |
+
| 5.8802 | 123.0 | 1107 | 5.7231 | 0.3099 |
|
| 175 |
+
| 5.8707 | 124.0 | 1116 | 5.7240 | 0.3099 |
|
| 176 |
+
| 5.8653 | 125.0 | 1125 | 5.7232 | 0.3109 |
|
| 177 |
+
| 5.8693 | 126.0 | 1134 | 5.7180 | 0.3109 |
|
| 178 |
+
| 5.8662 | 127.0 | 1143 | 5.7147 | 0.3109 |
|
| 179 |
+
| 5.8539 | 128.0 | 1152 | 5.7132 | 0.3099 |
|
| 180 |
+
| 5.8611 | 129.0 | 1161 | 5.7127 | 0.3099 |
|
| 181 |
+
| 5.8495 | 130.0 | 1170 | 5.7135 | 0.3079 |
|
| 182 |
+
| 5.8602 | 131.0 | 1179 | 5.7111 | 0.3089 |
|
| 183 |
+
| 5.8512 | 132.0 | 1188 | 5.7077 | 0.3109 |
|
| 184 |
+
| 5.8493 | 133.0 | 1197 | 5.7050 | 0.3109 |
|
| 185 |
+
| 5.8477 | 134.0 | 1206 | 5.7041 | 0.3099 |
|
| 186 |
+
| 5.8464 | 135.0 | 1215 | 5.7042 | 0.3089 |
|
| 187 |
+
| 5.8459 | 136.0 | 1224 | 5.7023 | 0.3099 |
|
| 188 |
+
| 5.8475 | 137.0 | 1233 | 5.7000 | 0.3099 |
|
| 189 |
+
| 5.8384 | 138.0 | 1242 | 5.6982 | 0.3099 |
|
| 190 |
+
| 5.8453 | 139.0 | 1251 | 5.6976 | 0.3089 |
|
| 191 |
+
| 5.8441 | 140.0 | 1260 | 5.6973 | 0.3089 |
|
| 192 |
+
| 5.838 | 141.0 | 1269 | 5.6971 | 0.3079 |
|
| 193 |
+
| 5.8463 | 142.0 | 1278 | 5.6950 | 0.3099 |
|
| 194 |
+
| 5.8385 | 143.0 | 1287 | 5.6921 | 0.3099 |
|
| 195 |
+
| 5.8354 | 144.0 | 1296 | 5.6909 | 0.3099 |
|
| 196 |
+
| 5.8283 | 145.0 | 1305 | 5.6908 | 0.3079 |
|
| 197 |
+
| 5.8363 | 146.0 | 1314 | 5.6902 | 0.3079 |
|
| 198 |
+
| 5.8433 | 147.0 | 1323 | 5.6890 | 0.3079 |
|
| 199 |
+
| 5.8302 | 148.0 | 1332 | 5.6890 | 0.3079 |
|
| 200 |
+
| 5.8276 | 149.0 | 1341 | 5.6881 | 0.3089 |
|
| 201 |
+
| 5.8366 | 150.0 | 1350 | 5.6864 | 0.3089 |
|
| 202 |
+
| 5.826 | 151.0 | 1359 | 5.6852 | 0.3099 |
|
| 203 |
+
| 5.8293 | 152.0 | 1368 | 5.6844 | 0.3099 |
|
| 204 |
+
| 5.8278 | 153.0 | 1377 | 5.6834 | 0.3099 |
|
| 205 |
+
| 5.8239 | 154.0 | 1386 | 5.6830 | 0.3089 |
|
| 206 |
+
| 5.8262 | 155.0 | 1395 | 5.6818 | 0.3089 |
|
| 207 |
+
| 5.8253 | 156.0 | 1404 | 5.6808 | 0.3099 |
|
| 208 |
+
| 5.8169 | 157.0 | 1413 | 5.6793 | 0.3109 |
|
| 209 |
+
| 5.8201 | 158.0 | 1422 | 5.6795 | 0.3109 |
|
| 210 |
+
| 5.8077 | 159.0 | 1431 | 5.6799 | 0.3099 |
|
| 211 |
+
| 5.8222 | 160.0 | 1440 | 5.6789 | 0.3099 |
|
| 212 |
+
| 5.8191 | 161.0 | 1449 | 5.6778 | 0.3099 |
|
| 213 |
+
| 5.83 | 162.0 | 1458 | 5.6768 | 0.3099 |
|
| 214 |
+
| 5.8183 | 163.0 | 1467 | 5.6757 | 0.3109 |
|
| 215 |
+
| 5.8124 | 164.0 | 1476 | 5.6747 | 0.3109 |
|
| 216 |
+
| 5.8119 | 165.0 | 1485 | 5.6745 | 0.3109 |
|
| 217 |
+
| 5.821 | 166.0 | 1494 | 5.6744 | 0.3099 |
|
| 218 |
+
| 5.807 | 167.0 | 1503 | 5.6737 | 0.3099 |
|
| 219 |
+
| 5.8177 | 168.0 | 1512 | 5.6725 | 0.3109 |
|
| 220 |
+
| 5.8046 | 169.0 | 1521 | 5.6710 | 0.3109 |
|
| 221 |
+
| 5.8093 | 170.0 | 1530 | 5.6708 | 0.3109 |
|
| 222 |
+
| 5.8145 | 171.0 | 1539 | 5.6710 | 0.3109 |
|
| 223 |
+
| 5.803 | 172.0 | 1548 | 5.6704 | 0.3109 |
|
| 224 |
+
| 5.8038 | 173.0 | 1557 | 5.6697 | 0.3109 |
|
| 225 |
+
| 5.807 | 174.0 | 1566 | 5.6689 | 0.3109 |
|
| 226 |
+
| 5.7974 | 175.0 | 1575 | 5.6683 | 0.3109 |
|
| 227 |
+
| 5.8089 | 176.0 | 1584 | 5.6688 | 0.3109 |
|
| 228 |
+
| 5.8067 | 177.0 | 1593 | 5.6693 | 0.3109 |
|
| 229 |
+
| 5.8092 | 178.0 | 1602 | 5.6695 | 0.3109 |
|
| 230 |
+
| 5.8047 | 179.0 | 1611 | 5.6687 | 0.3099 |
|
| 231 |
+
| 5.8007 | 180.0 | 1620 | 5.6679 | 0.3099 |
|
| 232 |
+
| 5.8041 | 181.0 | 1629 | 5.6672 | 0.3099 |
|
| 233 |
+
| 5.8072 | 182.0 | 1638 | 5.6667 | 0.3099 |
|
| 234 |
+
| 5.8093 | 183.0 | 1647 | 5.6662 | 0.3099 |
|
| 235 |
+
| 5.7948 | 184.0 | 1656 | 5.6658 | 0.3099 |
|
| 236 |
+
| 5.7968 | 185.0 | 1665 | 5.6656 | 0.3099 |
|
| 237 |
+
| 5.8033 | 186.0 | 1674 | 5.6653 | 0.3099 |
|
| 238 |
+
| 5.8031 | 187.0 | 1683 | 5.6651 | 0.3099 |
|
| 239 |
+
| 5.7953 | 188.0 | 1692 | 5.6650 | 0.3099 |
|
| 240 |
+
| 5.8085 | 189.0 | 1701 | 5.6647 | 0.3099 |
|
| 241 |
+
| 5.8021 | 190.0 | 1710 | 5.6646 | 0.3099 |
|
| 242 |
+
| 5.7995 | 191.0 | 1719 | 5.6643 | 0.3099 |
|
| 243 |
+
| 5.8057 | 192.0 | 1728 | 5.6642 | 0.3099 |
|
| 244 |
+
| 5.7989 | 193.0 | 1737 | 5.6642 | 0.3099 |
|
| 245 |
+
| 5.7977 | 194.0 | 1746 | 5.6642 | 0.3099 |
|
| 246 |
+
| 5.8009 | 195.0 | 1755 | 5.6644 | 0.3099 |
|
| 247 |
+
| 5.7988 | 196.0 | 1764 | 5.6645 | 0.3099 |
|
| 248 |
+
| 5.8016 | 197.0 | 1773 | 5.6645 | 0.3099 |
|
| 249 |
+
| 5.7929 | 198.0 | 1782 | 5.6645 | 0.3099 |
|
| 250 |
+
| 5.7973 | 199.0 | 1791 | 5.6645 | 0.3099 |
|
| 251 |
+
| 5.8022 | 200.0 | 1800 | 5.6645 | 0.3099 |
|
| 252 |
+
|
| 253 |
+
|
| 254 |
+
### Framework versions
|
| 255 |
+
|
| 256 |
+
- Transformers 4.38.0.dev0
|
| 257 |
+
- Pytorch 2.2.0
|
| 258 |
+
- Datasets 2.16.1
|
| 259 |
+
- Tokenizers 0.15.1
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 200.0,
|
| 3 |
+
"eval_accuracy": 0.3098729227761486,
|
| 4 |
+
"eval_loss": 5.664193153381348,
|
| 5 |
+
"eval_runtime": 3.4037,
|
| 6 |
+
"eval_samples": 1,
|
| 7 |
+
"eval_samples_per_second": 0.294,
|
| 8 |
+
"eval_steps_per_second": 0.294,
|
| 9 |
+
"perplexity": 288.3552287902082,
|
| 10 |
+
"train_loss": 6.335995424058702,
|
| 11 |
+
"train_runtime": 30730.8013,
|
| 12 |
+
"train_samples": 9,
|
| 13 |
+
"train_samples_per_second": 0.059,
|
| 14 |
+
"train_steps_per_second": 0.059
|
| 15 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "nferruz/ProtGPT2",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"embd_pdrop": 0.1,
|
| 10 |
+
"eos_token_id": 0,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"layer_norm_epsilon": 1e-05,
|
| 13 |
+
"model_type": "gpt2",
|
| 14 |
+
"n_ctx": 1024,
|
| 15 |
+
"n_embd": 1280,
|
| 16 |
+
"n_head": 20,
|
| 17 |
+
"n_inner": null,
|
| 18 |
+
"n_layer": 36,
|
| 19 |
+
"n_positions": 1024,
|
| 20 |
+
"reorder_and_upcast_attn": false,
|
| 21 |
+
"resid_pdrop": 0.1,
|
| 22 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 23 |
+
"scale_attn_weights": true,
|
| 24 |
+
"summary_activation": null,
|
| 25 |
+
"summary_first_dropout": 0.1,
|
| 26 |
+
"summary_proj_to_labels": true,
|
| 27 |
+
"summary_type": "cls_index",
|
| 28 |
+
"summary_use_proj": true,
|
| 29 |
+
"task_specific_params": {
|
| 30 |
+
"text-generation": {
|
| 31 |
+
"do_sample": true,
|
| 32 |
+
"max_length": 50
|
| 33 |
+
}
|
| 34 |
+
},
|
| 35 |
+
"torch_dtype": "float32",
|
| 36 |
+
"transformers_version": "4.38.0.dev0",
|
| 37 |
+
"use_cache": true,
|
| 38 |
+
"vocab_size": 50257
|
| 39 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 200.0,
|
| 3 |
+
"eval_accuracy": 0.3098729227761486,
|
| 4 |
+
"eval_loss": 5.664193153381348,
|
| 5 |
+
"eval_runtime": 3.4037,
|
| 6 |
+
"eval_samples": 1,
|
| 7 |
+
"eval_samples_per_second": 0.294,
|
| 8 |
+
"eval_steps_per_second": 0.294,
|
| 9 |
+
"perplexity": 288.3552287902082
|
| 10 |
+
}
|
events.out.tfevents.1708315853.lambda-a6000.2440433.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8cd879dade5383b85360197eacb44e3d5c8b9ed8c032f9e16f866e8df272aee
|
| 3 |
+
size 100848
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": 0,
|
| 5 |
+
"transformers_version": "4.38.0.dev0"
|
| 6 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ecec3742bf6d98cd536a99a4024057ade3080823d36a222764ea97e338f5ffab
|
| 3 |
+
size 3096165928
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<|endoftext|>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "<|endoftext|>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"unk_token": {
|
| 17 |
+
"content": "<|endoftext|>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": true,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
}
|
| 23 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "<|endoftext|>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
}
|
| 12 |
+
},
|
| 13 |
+
"bos_token": "<|endoftext|>",
|
| 14 |
+
"clean_up_tokenization_spaces": true,
|
| 15 |
+
"eos_token": "<|endoftext|>",
|
| 16 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 17 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 18 |
+
"unk_token": "<|endoftext|>"
|
| 19 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 200.0,
|
| 3 |
+
"train_loss": 6.335995424058702,
|
| 4 |
+
"train_runtime": 30730.8013,
|
| 5 |
+
"train_samples": 9,
|
| 6 |
+
"train_samples_per_second": 0.059,
|
| 7 |
+
"train_steps_per_second": 0.059
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3030 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 5.664193153381348,
|
| 3 |
+
"best_model_checkpoint": "output_nf_3/checkpoint-1737",
|
| 4 |
+
"epoch": 200.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1800,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"learning_rate": 9.95e-07,
|
| 14 |
+
"loss": 12.1762,
|
| 15 |
+
"step": 9
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 1.0,
|
| 19 |
+
"eval_accuracy": 0.002932551319648094,
|
| 20 |
+
"eval_loss": 11.818068504333496,
|
| 21 |
+
"eval_runtime": 3.6297,
|
| 22 |
+
"eval_samples_per_second": 0.276,
|
| 23 |
+
"eval_steps_per_second": 0.276,
|
| 24 |
+
"step": 9
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 2.0,
|
| 28 |
+
"learning_rate": 9.9e-07,
|
| 29 |
+
"loss": 11.6538,
|
| 30 |
+
"step": 18
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 2.0,
|
| 34 |
+
"eval_accuracy": 0.0009775171065493646,
|
| 35 |
+
"eval_loss": 11.346792221069336,
|
| 36 |
+
"eval_runtime": 3.7372,
|
| 37 |
+
"eval_samples_per_second": 0.268,
|
| 38 |
+
"eval_steps_per_second": 0.268,
|
| 39 |
+
"step": 18
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 3.0,
|
| 43 |
+
"learning_rate": 9.849999999999999e-07,
|
| 44 |
+
"loss": 11.2876,
|
| 45 |
+
"step": 27
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 3.0,
|
| 49 |
+
"eval_accuracy": 0.0009775171065493646,
|
| 50 |
+
"eval_loss": 10.912660598754883,
|
| 51 |
+
"eval_runtime": 4.2343,
|
| 52 |
+
"eval_samples_per_second": 0.236,
|
| 53 |
+
"eval_steps_per_second": 0.236,
|
| 54 |
+
"step": 27
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 4.0,
|
| 58 |
+
"learning_rate": 9.8e-07,
|
| 59 |
+
"loss": 10.8664,
|
| 60 |
+
"step": 36
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 4.0,
|
| 64 |
+
"eval_accuracy": 0.0009775171065493646,
|
| 65 |
+
"eval_loss": 10.516217231750488,
|
| 66 |
+
"eval_runtime": 3.4285,
|
| 67 |
+
"eval_samples_per_second": 0.292,
|
| 68 |
+
"eval_steps_per_second": 0.292,
|
| 69 |
+
"step": 36
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 5.0,
|
| 73 |
+
"learning_rate": 9.75e-07,
|
| 74 |
+
"loss": 10.6184,
|
| 75 |
+
"step": 45
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 5.0,
|
| 79 |
+
"eval_accuracy": 0.0019550342130987292,
|
| 80 |
+
"eval_loss": 10.181687355041504,
|
| 81 |
+
"eval_runtime": 3.9045,
|
| 82 |
+
"eval_samples_per_second": 0.256,
|
| 83 |
+
"eval_steps_per_second": 0.256,
|
| 84 |
+
"step": 45
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 6.0,
|
| 88 |
+
"learning_rate": 9.7e-07,
|
| 89 |
+
"loss": 10.3132,
|
| 90 |
+
"step": 54
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 6.0,
|
| 94 |
+
"eval_accuracy": 0.0009775171065493646,
|
| 95 |
+
"eval_loss": 9.876521110534668,
|
| 96 |
+
"eval_runtime": 3.4438,
|
| 97 |
+
"eval_samples_per_second": 0.29,
|
| 98 |
+
"eval_steps_per_second": 0.29,
|
| 99 |
+
"step": 54
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 7.0,
|
| 103 |
+
"learning_rate": 9.649999999999999e-07,
|
| 104 |
+
"loss": 9.9087,
|
| 105 |
+
"step": 63
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 7.0,
|
| 109 |
+
"eval_accuracy": 0.0009775171065493646,
|
| 110 |
+
"eval_loss": 9.607266426086426,
|
| 111 |
+
"eval_runtime": 3.7078,
|
| 112 |
+
"eval_samples_per_second": 0.27,
|
| 113 |
+
"eval_steps_per_second": 0.27,
|
| 114 |
+
"step": 63
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 8.0,
|
| 118 |
+
"learning_rate": 9.6e-07,
|
| 119 |
+
"loss": 9.6953,
|
| 120 |
+
"step": 72
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 8.0,
|
| 124 |
+
"eval_accuracy": 0.0,
|
| 125 |
+
"eval_loss": 9.324377059936523,
|
| 126 |
+
"eval_runtime": 3.5648,
|
| 127 |
+
"eval_samples_per_second": 0.281,
|
| 128 |
+
"eval_steps_per_second": 0.281,
|
| 129 |
+
"step": 72
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 9.0,
|
| 133 |
+
"learning_rate": 9.55e-07,
|
| 134 |
+
"loss": 9.3741,
|
| 135 |
+
"step": 81
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 9.0,
|
| 139 |
+
"eval_accuracy": 0.0,
|
| 140 |
+
"eval_loss": 9.031828880310059,
|
| 141 |
+
"eval_runtime": 3.42,
|
| 142 |
+
"eval_samples_per_second": 0.292,
|
| 143 |
+
"eval_steps_per_second": 0.292,
|
| 144 |
+
"step": 81
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 10.0,
|
| 148 |
+
"learning_rate": 9.499999999999999e-07,
|
| 149 |
+
"loss": 9.2045,
|
| 150 |
+
"step": 90
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 10.0,
|
| 154 |
+
"eval_accuracy": 0.0,
|
| 155 |
+
"eval_loss": 8.746149063110352,
|
| 156 |
+
"eval_runtime": 3.746,
|
| 157 |
+
"eval_samples_per_second": 0.267,
|
| 158 |
+
"eval_steps_per_second": 0.267,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 11.0,
|
| 163 |
+
"learning_rate": 9.45e-07,
|
| 164 |
+
"loss": 8.9079,
|
| 165 |
+
"step": 99
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 11.0,
|
| 169 |
+
"eval_accuracy": 0.0019550342130987292,
|
| 170 |
+
"eval_loss": 8.47550106048584,
|
| 171 |
+
"eval_runtime": 3.3518,
|
| 172 |
+
"eval_samples_per_second": 0.298,
|
| 173 |
+
"eval_steps_per_second": 0.298,
|
| 174 |
+
"step": 99
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 12.0,
|
| 178 |
+
"learning_rate": 9.399999999999999e-07,
|
| 179 |
+
"loss": 8.7047,
|
| 180 |
+
"step": 108
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 12.0,
|
| 184 |
+
"eval_accuracy": 0.011730205278592375,
|
| 185 |
+
"eval_loss": 8.222208976745605,
|
| 186 |
+
"eval_runtime": 3.4648,
|
| 187 |
+
"eval_samples_per_second": 0.289,
|
| 188 |
+
"eval_steps_per_second": 0.289,
|
| 189 |
+
"step": 108
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 13.0,
|
| 193 |
+
"learning_rate": 9.35e-07,
|
| 194 |
+
"loss": 8.4622,
|
| 195 |
+
"step": 117
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 13.0,
|
| 199 |
+
"eval_accuracy": 0.03128054740957967,
|
| 200 |
+
"eval_loss": 7.995096206665039,
|
| 201 |
+
"eval_runtime": 3.5365,
|
| 202 |
+
"eval_samples_per_second": 0.283,
|
| 203 |
+
"eval_steps_per_second": 0.283,
|
| 204 |
+
"step": 117
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 14.0,
|
| 208 |
+
"learning_rate": 9.3e-07,
|
| 209 |
+
"loss": 8.2649,
|
| 210 |
+
"step": 126
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 14.0,
|
| 214 |
+
"eval_accuracy": 0.06549364613880743,
|
| 215 |
+
"eval_loss": 7.805217266082764,
|
| 216 |
+
"eval_runtime": 3.456,
|
| 217 |
+
"eval_samples_per_second": 0.289,
|
| 218 |
+
"eval_steps_per_second": 0.289,
|
| 219 |
+
"step": 126
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 15.0,
|
| 223 |
+
"learning_rate": 9.25e-07,
|
| 224 |
+
"loss": 8.043,
|
| 225 |
+
"step": 135
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 15.0,
|
| 229 |
+
"eval_accuracy": 0.10654936461388075,
|
| 230 |
+
"eval_loss": 7.649549961090088,
|
| 231 |
+
"eval_runtime": 3.7554,
|
| 232 |
+
"eval_samples_per_second": 0.266,
|
| 233 |
+
"eval_steps_per_second": 0.266,
|
| 234 |
+
"step": 135
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 16.0,
|
| 238 |
+
"learning_rate": 9.2e-07,
|
| 239 |
+
"loss": 7.9092,
|
| 240 |
+
"step": 144
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 16.0,
|
| 244 |
+
"eval_accuracy": 0.1378299120234604,
|
| 245 |
+
"eval_loss": 7.516005516052246,
|
| 246 |
+
"eval_runtime": 4.0508,
|
| 247 |
+
"eval_samples_per_second": 0.247,
|
| 248 |
+
"eval_steps_per_second": 0.247,
|
| 249 |
+
"step": 144
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 17.0,
|
| 253 |
+
"learning_rate": 9.15e-07,
|
| 254 |
+
"loss": 7.7103,
|
| 255 |
+
"step": 153
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 17.0,
|
| 259 |
+
"eval_accuracy": 0.16911045943304007,
|
| 260 |
+
"eval_loss": 7.388144016265869,
|
| 261 |
+
"eval_runtime": 3.4529,
|
| 262 |
+
"eval_samples_per_second": 0.29,
|
| 263 |
+
"eval_steps_per_second": 0.29,
|
| 264 |
+
"step": 153
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 18.0,
|
| 268 |
+
"learning_rate": 9.1e-07,
|
| 269 |
+
"loss": 7.5701,
|
| 270 |
+
"step": 162
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 18.0,
|
| 274 |
+
"eval_accuracy": 0.18181818181818182,
|
| 275 |
+
"eval_loss": 7.269326686859131,
|
| 276 |
+
"eval_runtime": 3.467,
|
| 277 |
+
"eval_samples_per_second": 0.288,
|
| 278 |
+
"eval_steps_per_second": 0.288,
|
| 279 |
+
"step": 162
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 19.0,
|
| 283 |
+
"learning_rate": 9.05e-07,
|
| 284 |
+
"loss": 7.4483,
|
| 285 |
+
"step": 171
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 19.0,
|
| 289 |
+
"eval_accuracy": 0.19941348973607037,
|
| 290 |
+
"eval_loss": 7.154385089874268,
|
| 291 |
+
"eval_runtime": 4.1166,
|
| 292 |
+
"eval_samples_per_second": 0.243,
|
| 293 |
+
"eval_steps_per_second": 0.243,
|
| 294 |
+
"step": 171
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 20.0,
|
| 298 |
+
"learning_rate": 9e-07,
|
| 299 |
+
"loss": 7.314,
|
| 300 |
+
"step": 180
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 20.0,
|
| 304 |
+
"eval_accuracy": 0.21603128054740958,
|
| 305 |
+
"eval_loss": 7.046892166137695,
|
| 306 |
+
"eval_runtime": 5.2069,
|
| 307 |
+
"eval_samples_per_second": 0.192,
|
| 308 |
+
"eval_steps_per_second": 0.192,
|
| 309 |
+
"step": 180
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 21.0,
|
| 313 |
+
"learning_rate": 8.95e-07,
|
| 314 |
+
"loss": 7.2101,
|
| 315 |
+
"step": 189
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 21.0,
|
| 319 |
+
"eval_accuracy": 0.23264907135874877,
|
| 320 |
+
"eval_loss": 6.9440107345581055,
|
| 321 |
+
"eval_runtime": 3.9239,
|
| 322 |
+
"eval_samples_per_second": 0.255,
|
| 323 |
+
"eval_steps_per_second": 0.255,
|
| 324 |
+
"step": 189
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 22.0,
|
| 328 |
+
"learning_rate": 8.9e-07,
|
| 329 |
+
"loss": 7.1026,
|
| 330 |
+
"step": 198
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 22.0,
|
| 334 |
+
"eval_accuracy": 0.23655913978494625,
|
| 335 |
+
"eval_loss": 6.854238033294678,
|
| 336 |
+
"eval_runtime": 3.4311,
|
| 337 |
+
"eval_samples_per_second": 0.291,
|
| 338 |
+
"eval_steps_per_second": 0.291,
|
| 339 |
+
"step": 198
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 23.0,
|
| 343 |
+
"learning_rate": 8.85e-07,
|
| 344 |
+
"loss": 6.9954,
|
| 345 |
+
"step": 207
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 23.0,
|
| 349 |
+
"eval_accuracy": 0.2482893450635386,
|
| 350 |
+
"eval_loss": 6.768421173095703,
|
| 351 |
+
"eval_runtime": 4.5026,
|
| 352 |
+
"eval_samples_per_second": 0.222,
|
| 353 |
+
"eval_steps_per_second": 0.222,
|
| 354 |
+
"step": 207
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 24.0,
|
| 358 |
+
"learning_rate": 8.799999999999999e-07,
|
| 359 |
+
"loss": 6.9206,
|
| 360 |
+
"step": 216
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 24.0,
|
| 364 |
+
"eval_accuracy": 0.2512218963831867,
|
| 365 |
+
"eval_loss": 6.685108184814453,
|
| 366 |
+
"eval_runtime": 9.9405,
|
| 367 |
+
"eval_samples_per_second": 0.101,
|
| 368 |
+
"eval_steps_per_second": 0.101,
|
| 369 |
+
"step": 216
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 25.0,
|
| 373 |
+
"learning_rate": 8.75e-07,
|
| 374 |
+
"loss": 6.8588,
|
| 375 |
+
"step": 225
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 25.0,
|
| 379 |
+
"eval_accuracy": 0.2561094819159335,
|
| 380 |
+
"eval_loss": 6.621737957000732,
|
| 381 |
+
"eval_runtime": 3.7996,
|
| 382 |
+
"eval_samples_per_second": 0.263,
|
| 383 |
+
"eval_steps_per_second": 0.263,
|
| 384 |
+
"step": 225
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 26.0,
|
| 388 |
+
"learning_rate": 8.699999999999999e-07,
|
| 389 |
+
"loss": 6.7975,
|
| 390 |
+
"step": 234
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 26.0,
|
| 394 |
+
"eval_accuracy": 0.2590420332355816,
|
| 395 |
+
"eval_loss": 6.562064170837402,
|
| 396 |
+
"eval_runtime": 4.7111,
|
| 397 |
+
"eval_samples_per_second": 0.212,
|
| 398 |
+
"eval_steps_per_second": 0.212,
|
| 399 |
+
"step": 234
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 27.0,
|
| 403 |
+
"learning_rate": 8.65e-07,
|
| 404 |
+
"loss": 6.7355,
|
| 405 |
+
"step": 243
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 27.0,
|
| 409 |
+
"eval_accuracy": 0.26099706744868034,
|
| 410 |
+
"eval_loss": 6.5111308097839355,
|
| 411 |
+
"eval_runtime": 3.4887,
|
| 412 |
+
"eval_samples_per_second": 0.287,
|
| 413 |
+
"eval_steps_per_second": 0.287,
|
| 414 |
+
"step": 243
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 28.0,
|
| 418 |
+
"learning_rate": 8.599999999999999e-07,
|
| 419 |
+
"loss": 6.6928,
|
| 420 |
+
"step": 252
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 28.0,
|
| 424 |
+
"eval_accuracy": 0.26392961876832843,
|
| 425 |
+
"eval_loss": 6.462972640991211,
|
| 426 |
+
"eval_runtime": 10.4927,
|
| 427 |
+
"eval_samples_per_second": 0.095,
|
| 428 |
+
"eval_steps_per_second": 0.095,
|
| 429 |
+
"step": 252
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 29.0,
|
| 433 |
+
"learning_rate": 8.55e-07,
|
| 434 |
+
"loss": 6.6483,
|
| 435 |
+
"step": 261
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 29.0,
|
| 439 |
+
"eval_accuracy": 0.2697947214076246,
|
| 440 |
+
"eval_loss": 6.413713455200195,
|
| 441 |
+
"eval_runtime": 9.1985,
|
| 442 |
+
"eval_samples_per_second": 0.109,
|
| 443 |
+
"eval_steps_per_second": 0.109,
|
| 444 |
+
"step": 261
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 30.0,
|
| 448 |
+
"learning_rate": 8.499999999999999e-07,
|
| 449 |
+
"loss": 6.6169,
|
| 450 |
+
"step": 270
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 30.0,
|
| 454 |
+
"eval_accuracy": 0.27174975562072334,
|
| 455 |
+
"eval_loss": 6.367518424987793,
|
| 456 |
+
"eval_runtime": 4.5985,
|
| 457 |
+
"eval_samples_per_second": 0.217,
|
| 458 |
+
"eval_steps_per_second": 0.217,
|
| 459 |
+
"step": 270
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 31.0,
|
| 463 |
+
"learning_rate": 8.45e-07,
|
| 464 |
+
"loss": 6.5498,
|
| 465 |
+
"step": 279
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 31.0,
|
| 469 |
+
"eval_accuracy": 0.27468230694037143,
|
| 470 |
+
"eval_loss": 6.329223155975342,
|
| 471 |
+
"eval_runtime": 4.0947,
|
| 472 |
+
"eval_samples_per_second": 0.244,
|
| 473 |
+
"eval_steps_per_second": 0.244,
|
| 474 |
+
"step": 279
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 32.0,
|
| 478 |
+
"learning_rate": 8.399999999999999e-07,
|
| 479 |
+
"loss": 6.5385,
|
| 480 |
+
"step": 288
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 32.0,
|
| 484 |
+
"eval_accuracy": 0.27663734115347016,
|
| 485 |
+
"eval_loss": 6.296669960021973,
|
| 486 |
+
"eval_runtime": 3.6593,
|
| 487 |
+
"eval_samples_per_second": 0.273,
|
| 488 |
+
"eval_steps_per_second": 0.273,
|
| 489 |
+
"step": 288
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 33.0,
|
| 493 |
+
"learning_rate": 8.349999999999999e-07,
|
| 494 |
+
"loss": 6.5111,
|
| 495 |
+
"step": 297
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 33.0,
|
| 499 |
+
"eval_accuracy": 0.2785923753665689,
|
| 500 |
+
"eval_loss": 6.2592363357543945,
|
| 501 |
+
"eval_runtime": 4.3919,
|
| 502 |
+
"eval_samples_per_second": 0.228,
|
| 503 |
+
"eval_steps_per_second": 0.228,
|
| 504 |
+
"step": 297
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 34.0,
|
| 508 |
+
"learning_rate": 8.299999999999999e-07,
|
| 509 |
+
"loss": 6.4748,
|
| 510 |
+
"step": 306
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 34.0,
|
| 514 |
+
"eval_accuracy": 0.28152492668621704,
|
| 515 |
+
"eval_loss": 6.231565475463867,
|
| 516 |
+
"eval_runtime": 4.3688,
|
| 517 |
+
"eval_samples_per_second": 0.229,
|
| 518 |
+
"eval_steps_per_second": 0.229,
|
| 519 |
+
"step": 306
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 35.0,
|
| 523 |
+
"learning_rate": 8.249999999999999e-07,
|
| 524 |
+
"loss": 6.4575,
|
| 525 |
+
"step": 315
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 35.0,
|
| 529 |
+
"eval_accuracy": 0.28347996089931576,
|
| 530 |
+
"eval_loss": 6.209580421447754,
|
| 531 |
+
"eval_runtime": 3.5499,
|
| 532 |
+
"eval_samples_per_second": 0.282,
|
| 533 |
+
"eval_steps_per_second": 0.282,
|
| 534 |
+
"step": 315
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 36.0,
|
| 538 |
+
"learning_rate": 8.199999999999999e-07,
|
| 539 |
+
"loss": 6.4251,
|
| 540 |
+
"step": 324
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 36.0,
|
| 544 |
+
"eval_accuracy": 0.2844574780058651,
|
| 545 |
+
"eval_loss": 6.184682369232178,
|
| 546 |
+
"eval_runtime": 3.6926,
|
| 547 |
+
"eval_samples_per_second": 0.271,
|
| 548 |
+
"eval_steps_per_second": 0.271,
|
| 549 |
+
"step": 324
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 37.0,
|
| 553 |
+
"learning_rate": 8.149999999999999e-07,
|
| 554 |
+
"loss": 6.4096,
|
| 555 |
+
"step": 333
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 37.0,
|
| 559 |
+
"eval_accuracy": 0.2854349951124145,
|
| 560 |
+
"eval_loss": 6.161267280578613,
|
| 561 |
+
"eval_runtime": 3.9736,
|
| 562 |
+
"eval_samples_per_second": 0.252,
|
| 563 |
+
"eval_steps_per_second": 0.252,
|
| 564 |
+
"step": 333
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 38.0,
|
| 568 |
+
"learning_rate": 8.1e-07,
|
| 569 |
+
"loss": 6.3741,
|
| 570 |
+
"step": 342
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 38.0,
|
| 574 |
+
"eval_accuracy": 0.2854349951124145,
|
| 575 |
+
"eval_loss": 6.140196800231934,
|
| 576 |
+
"eval_runtime": 3.9847,
|
| 577 |
+
"eval_samples_per_second": 0.251,
|
| 578 |
+
"eval_steps_per_second": 0.251,
|
| 579 |
+
"step": 342
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 39.0,
|
| 583 |
+
"learning_rate": 8.05e-07,
|
| 584 |
+
"loss": 6.3645,
|
| 585 |
+
"step": 351
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 39.0,
|
| 589 |
+
"eval_accuracy": 0.2873900293255132,
|
| 590 |
+
"eval_loss": 6.128504753112793,
|
| 591 |
+
"eval_runtime": 3.6581,
|
| 592 |
+
"eval_samples_per_second": 0.273,
|
| 593 |
+
"eval_steps_per_second": 0.273,
|
| 594 |
+
"step": 351
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 40.0,
|
| 598 |
+
"learning_rate": 8e-07,
|
| 599 |
+
"loss": 6.3511,
|
| 600 |
+
"step": 360
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 40.0,
|
| 604 |
+
"eval_accuracy": 0.2873900293255132,
|
| 605 |
+
"eval_loss": 6.11335563659668,
|
| 606 |
+
"eval_runtime": 3.9362,
|
| 607 |
+
"eval_samples_per_second": 0.254,
|
| 608 |
+
"eval_steps_per_second": 0.254,
|
| 609 |
+
"step": 360
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 41.0,
|
| 613 |
+
"learning_rate": 7.95e-07,
|
| 614 |
+
"loss": 6.3254,
|
| 615 |
+
"step": 369
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 41.0,
|
| 619 |
+
"eval_accuracy": 0.2873900293255132,
|
| 620 |
+
"eval_loss": 6.095658779144287,
|
| 621 |
+
"eval_runtime": 3.8555,
|
| 622 |
+
"eval_samples_per_second": 0.259,
|
| 623 |
+
"eval_steps_per_second": 0.259,
|
| 624 |
+
"step": 369
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 42.0,
|
| 628 |
+
"learning_rate": 7.9e-07,
|
| 629 |
+
"loss": 6.3077,
|
| 630 |
+
"step": 378
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 42.0,
|
| 634 |
+
"eval_accuracy": 0.2873900293255132,
|
| 635 |
+
"eval_loss": 6.081845283508301,
|
| 636 |
+
"eval_runtime": 3.4864,
|
| 637 |
+
"eval_samples_per_second": 0.287,
|
| 638 |
+
"eval_steps_per_second": 0.287,
|
| 639 |
+
"step": 378
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 43.0,
|
| 643 |
+
"learning_rate": 7.85e-07,
|
| 644 |
+
"loss": 6.301,
|
| 645 |
+
"step": 387
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 43.0,
|
| 649 |
+
"eval_accuracy": 0.2873900293255132,
|
| 650 |
+
"eval_loss": 6.0687761306762695,
|
| 651 |
+
"eval_runtime": 4.7364,
|
| 652 |
+
"eval_samples_per_second": 0.211,
|
| 653 |
+
"eval_steps_per_second": 0.211,
|
| 654 |
+
"step": 387
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 44.0,
|
| 658 |
+
"learning_rate": 7.799999999999999e-07,
|
| 659 |
+
"loss": 6.2846,
|
| 660 |
+
"step": 396
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 44.0,
|
| 664 |
+
"eval_accuracy": 0.2883675464320626,
|
| 665 |
+
"eval_loss": 6.051051616668701,
|
| 666 |
+
"eval_runtime": 3.4251,
|
| 667 |
+
"eval_samples_per_second": 0.292,
|
| 668 |
+
"eval_steps_per_second": 0.292,
|
| 669 |
+
"step": 396
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 45.0,
|
| 673 |
+
"learning_rate": 7.75e-07,
|
| 674 |
+
"loss": 6.2739,
|
| 675 |
+
"step": 405
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 45.0,
|
| 679 |
+
"eval_accuracy": 0.28934506353861195,
|
| 680 |
+
"eval_loss": 6.039752006530762,
|
| 681 |
+
"eval_runtime": 4.3563,
|
| 682 |
+
"eval_samples_per_second": 0.23,
|
| 683 |
+
"eval_steps_per_second": 0.23,
|
| 684 |
+
"step": 405
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 46.0,
|
| 688 |
+
"learning_rate": 7.699999999999999e-07,
|
| 689 |
+
"loss": 6.2569,
|
| 690 |
+
"step": 414
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 46.0,
|
| 694 |
+
"eval_accuracy": 0.2913000977517107,
|
| 695 |
+
"eval_loss": 6.030134677886963,
|
| 696 |
+
"eval_runtime": 3.8175,
|
| 697 |
+
"eval_samples_per_second": 0.262,
|
| 698 |
+
"eval_steps_per_second": 0.262,
|
| 699 |
+
"step": 414
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 47.0,
|
| 703 |
+
"learning_rate": 7.65e-07,
|
| 704 |
+
"loss": 6.258,
|
| 705 |
+
"step": 423
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 47.0,
|
| 709 |
+
"eval_accuracy": 0.2913000977517107,
|
| 710 |
+
"eval_loss": 6.01760721206665,
|
| 711 |
+
"eval_runtime": 4.1737,
|
| 712 |
+
"eval_samples_per_second": 0.24,
|
| 713 |
+
"eval_steps_per_second": 0.24,
|
| 714 |
+
"step": 423
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 48.0,
|
| 718 |
+
"learning_rate": 7.599999999999999e-07,
|
| 719 |
+
"loss": 6.2273,
|
| 720 |
+
"step": 432
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 48.0,
|
| 724 |
+
"eval_accuracy": 0.2932551319648094,
|
| 725 |
+
"eval_loss": 6.003724575042725,
|
| 726 |
+
"eval_runtime": 3.8811,
|
| 727 |
+
"eval_samples_per_second": 0.258,
|
| 728 |
+
"eval_steps_per_second": 0.258,
|
| 729 |
+
"step": 432
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 49.0,
|
| 733 |
+
"learning_rate": 7.55e-07,
|
| 734 |
+
"loss": 6.2256,
|
| 735 |
+
"step": 441
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 49.0,
|
| 739 |
+
"eval_accuracy": 0.2932551319648094,
|
| 740 |
+
"eval_loss": 5.9961748123168945,
|
| 741 |
+
"eval_runtime": 4.331,
|
| 742 |
+
"eval_samples_per_second": 0.231,
|
| 743 |
+
"eval_steps_per_second": 0.231,
|
| 744 |
+
"step": 441
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 50.0,
|
| 748 |
+
"learning_rate": 7.5e-07,
|
| 749 |
+
"loss": 6.2065,
|
| 750 |
+
"step": 450
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 50.0,
|
| 754 |
+
"eval_accuracy": 0.2932551319648094,
|
| 755 |
+
"eval_loss": 5.986941337585449,
|
| 756 |
+
"eval_runtime": 4.6185,
|
| 757 |
+
"eval_samples_per_second": 0.217,
|
| 758 |
+
"eval_steps_per_second": 0.217,
|
| 759 |
+
"step": 450
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 51.0,
|
| 763 |
+
"learning_rate": 7.45e-07,
|
| 764 |
+
"loss": 6.1991,
|
| 765 |
+
"step": 459
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 51.0,
|
| 769 |
+
"eval_accuracy": 0.2932551319648094,
|
| 770 |
+
"eval_loss": 5.975553512573242,
|
| 771 |
+
"eval_runtime": 3.9164,
|
| 772 |
+
"eval_samples_per_second": 0.255,
|
| 773 |
+
"eval_steps_per_second": 0.255,
|
| 774 |
+
"step": 459
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 52.0,
|
| 778 |
+
"learning_rate": 7.4e-07,
|
| 779 |
+
"loss": 6.1895,
|
| 780 |
+
"step": 468
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 52.0,
|
| 784 |
+
"eval_accuracy": 0.29423264907135877,
|
| 785 |
+
"eval_loss": 5.968191623687744,
|
| 786 |
+
"eval_runtime": 4.4356,
|
| 787 |
+
"eval_samples_per_second": 0.225,
|
| 788 |
+
"eval_steps_per_second": 0.225,
|
| 789 |
+
"step": 468
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 53.0,
|
| 793 |
+
"learning_rate": 7.35e-07,
|
| 794 |
+
"loss": 6.1763,
|
| 795 |
+
"step": 477
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 53.0,
|
| 799 |
+
"eval_accuracy": 0.29423264907135877,
|
| 800 |
+
"eval_loss": 5.961073398590088,
|
| 801 |
+
"eval_runtime": 4.4638,
|
| 802 |
+
"eval_samples_per_second": 0.224,
|
| 803 |
+
"eval_steps_per_second": 0.224,
|
| 804 |
+
"step": 477
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 54.0,
|
| 808 |
+
"learning_rate": 7.3e-07,
|
| 809 |
+
"loss": 6.1734,
|
| 810 |
+
"step": 486
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 54.0,
|
| 814 |
+
"eval_accuracy": 0.29423264907135877,
|
| 815 |
+
"eval_loss": 5.953503608703613,
|
| 816 |
+
"eval_runtime": 3.7888,
|
| 817 |
+
"eval_samples_per_second": 0.264,
|
| 818 |
+
"eval_steps_per_second": 0.264,
|
| 819 |
+
"step": 486
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 55.0,
|
| 823 |
+
"learning_rate": 7.249999999999999e-07,
|
| 824 |
+
"loss": 6.1702,
|
| 825 |
+
"step": 495
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 55.0,
|
| 829 |
+
"eval_accuracy": 0.29423264907135877,
|
| 830 |
+
"eval_loss": 5.94571590423584,
|
| 831 |
+
"eval_runtime": 4.1672,
|
| 832 |
+
"eval_samples_per_second": 0.24,
|
| 833 |
+
"eval_steps_per_second": 0.24,
|
| 834 |
+
"step": 495
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 56.0,
|
| 838 |
+
"learning_rate": 7.2e-07,
|
| 839 |
+
"loss": 6.1556,
|
| 840 |
+
"step": 504
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 56.0,
|
| 844 |
+
"eval_accuracy": 0.29521016617790813,
|
| 845 |
+
"eval_loss": 5.937567710876465,
|
| 846 |
+
"eval_runtime": 4.244,
|
| 847 |
+
"eval_samples_per_second": 0.236,
|
| 848 |
+
"eval_steps_per_second": 0.236,
|
| 849 |
+
"step": 504
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 57.0,
|
| 853 |
+
"learning_rate": 7.149999999999999e-07,
|
| 854 |
+
"loss": 6.1481,
|
| 855 |
+
"step": 513
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 57.0,
|
| 859 |
+
"eval_accuracy": 0.29521016617790813,
|
| 860 |
+
"eval_loss": 5.930552005767822,
|
| 861 |
+
"eval_runtime": 4.4958,
|
| 862 |
+
"eval_samples_per_second": 0.222,
|
| 863 |
+
"eval_steps_per_second": 0.222,
|
| 864 |
+
"step": 513
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 58.0,
|
| 868 |
+
"learning_rate": 7.1e-07,
|
| 869 |
+
"loss": 6.1425,
|
| 870 |
+
"step": 522
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 58.0,
|
| 874 |
+
"eval_accuracy": 0.29423264907135877,
|
| 875 |
+
"eval_loss": 5.922237396240234,
|
| 876 |
+
"eval_runtime": 4.1228,
|
| 877 |
+
"eval_samples_per_second": 0.243,
|
| 878 |
+
"eval_steps_per_second": 0.243,
|
| 879 |
+
"step": 522
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 59.0,
|
| 883 |
+
"learning_rate": 7.049999999999999e-07,
|
| 884 |
+
"loss": 6.1416,
|
| 885 |
+
"step": 531
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 59.0,
|
| 889 |
+
"eval_accuracy": 0.29423264907135877,
|
| 890 |
+
"eval_loss": 5.91697883605957,
|
| 891 |
+
"eval_runtime": 4.4905,
|
| 892 |
+
"eval_samples_per_second": 0.223,
|
| 893 |
+
"eval_steps_per_second": 0.223,
|
| 894 |
+
"step": 531
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 60.0,
|
| 898 |
+
"learning_rate": 7e-07,
|
| 899 |
+
"loss": 6.1328,
|
| 900 |
+
"step": 540
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 60.0,
|
| 904 |
+
"eval_accuracy": 0.29423264907135877,
|
| 905 |
+
"eval_loss": 5.914181232452393,
|
| 906 |
+
"eval_runtime": 4.5555,
|
| 907 |
+
"eval_samples_per_second": 0.22,
|
| 908 |
+
"eval_steps_per_second": 0.22,
|
| 909 |
+
"step": 540
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 61.0,
|
| 913 |
+
"learning_rate": 6.949999999999999e-07,
|
| 914 |
+
"loss": 6.1176,
|
| 915 |
+
"step": 549
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 61.0,
|
| 919 |
+
"eval_accuracy": 0.29423264907135877,
|
| 920 |
+
"eval_loss": 5.906389236450195,
|
| 921 |
+
"eval_runtime": 4.1516,
|
| 922 |
+
"eval_samples_per_second": 0.241,
|
| 923 |
+
"eval_steps_per_second": 0.241,
|
| 924 |
+
"step": 549
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 62.0,
|
| 928 |
+
"learning_rate": 6.9e-07,
|
| 929 |
+
"loss": 6.1091,
|
| 930 |
+
"step": 558
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 62.0,
|
| 934 |
+
"eval_accuracy": 0.29423264907135877,
|
| 935 |
+
"eval_loss": 5.901010513305664,
|
| 936 |
+
"eval_runtime": 4.2358,
|
| 937 |
+
"eval_samples_per_second": 0.236,
|
| 938 |
+
"eval_steps_per_second": 0.236,
|
| 939 |
+
"step": 558
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 63.0,
|
| 943 |
+
"learning_rate": 6.85e-07,
|
| 944 |
+
"loss": 6.104,
|
| 945 |
+
"step": 567
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 63.0,
|
| 949 |
+
"eval_accuracy": 0.29423264907135877,
|
| 950 |
+
"eval_loss": 5.896124362945557,
|
| 951 |
+
"eval_runtime": 4.4857,
|
| 952 |
+
"eval_samples_per_second": 0.223,
|
| 953 |
+
"eval_steps_per_second": 0.223,
|
| 954 |
+
"step": 567
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 64.0,
|
| 958 |
+
"learning_rate": 6.800000000000001e-07,
|
| 959 |
+
"loss": 6.0986,
|
| 960 |
+
"step": 576
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 64.0,
|
| 964 |
+
"eval_accuracy": 0.29423264907135877,
|
| 965 |
+
"eval_loss": 5.891510486602783,
|
| 966 |
+
"eval_runtime": 4.4644,
|
| 967 |
+
"eval_samples_per_second": 0.224,
|
| 968 |
+
"eval_steps_per_second": 0.224,
|
| 969 |
+
"step": 576
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 65.0,
|
| 973 |
+
"learning_rate": 6.75e-07,
|
| 974 |
+
"loss": 6.089,
|
| 975 |
+
"step": 585
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 65.0,
|
| 979 |
+
"eval_accuracy": 0.29521016617790813,
|
| 980 |
+
"eval_loss": 5.885429382324219,
|
| 981 |
+
"eval_runtime": 3.9204,
|
| 982 |
+
"eval_samples_per_second": 0.255,
|
| 983 |
+
"eval_steps_per_second": 0.255,
|
| 984 |
+
"step": 585
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 66.0,
|
| 988 |
+
"learning_rate": 6.7e-07,
|
| 989 |
+
"loss": 6.0734,
|
| 990 |
+
"step": 594
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 66.0,
|
| 994 |
+
"eval_accuracy": 0.29423264907135877,
|
| 995 |
+
"eval_loss": 5.880984783172607,
|
| 996 |
+
"eval_runtime": 4.2715,
|
| 997 |
+
"eval_samples_per_second": 0.234,
|
| 998 |
+
"eval_steps_per_second": 0.234,
|
| 999 |
+
"step": 594
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 67.0,
|
| 1003 |
+
"learning_rate": 6.65e-07,
|
| 1004 |
+
"loss": 6.0905,
|
| 1005 |
+
"step": 603
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 67.0,
|
| 1009 |
+
"eval_accuracy": 0.29423264907135877,
|
| 1010 |
+
"eval_loss": 5.876211166381836,
|
| 1011 |
+
"eval_runtime": 4.4141,
|
| 1012 |
+
"eval_samples_per_second": 0.227,
|
| 1013 |
+
"eval_steps_per_second": 0.227,
|
| 1014 |
+
"step": 603
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 68.0,
|
| 1018 |
+
"learning_rate": 6.6e-07,
|
| 1019 |
+
"loss": 6.0701,
|
| 1020 |
+
"step": 612
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 68.0,
|
| 1024 |
+
"eval_accuracy": 0.2961876832844575,
|
| 1025 |
+
"eval_loss": 5.873714447021484,
|
| 1026 |
+
"eval_runtime": 4.3879,
|
| 1027 |
+
"eval_samples_per_second": 0.228,
|
| 1028 |
+
"eval_steps_per_second": 0.228,
|
| 1029 |
+
"step": 612
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 69.0,
|
| 1033 |
+
"learning_rate": 6.55e-07,
|
| 1034 |
+
"loss": 6.0595,
|
| 1035 |
+
"step": 621
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 69.0,
|
| 1039 |
+
"eval_accuracy": 0.2961876832844575,
|
| 1040 |
+
"eval_loss": 5.869368553161621,
|
| 1041 |
+
"eval_runtime": 4.2313,
|
| 1042 |
+
"eval_samples_per_second": 0.236,
|
| 1043 |
+
"eval_steps_per_second": 0.236,
|
| 1044 |
+
"step": 621
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 70.0,
|
| 1048 |
+
"learning_rate": 6.5e-07,
|
| 1049 |
+
"loss": 6.0616,
|
| 1050 |
+
"step": 630
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 70.0,
|
| 1054 |
+
"eval_accuracy": 0.2961876832844575,
|
| 1055 |
+
"eval_loss": 5.8661274909973145,
|
| 1056 |
+
"eval_runtime": 4.1245,
|
| 1057 |
+
"eval_samples_per_second": 0.242,
|
| 1058 |
+
"eval_steps_per_second": 0.242,
|
| 1059 |
+
"step": 630
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 71.0,
|
| 1063 |
+
"learning_rate": 6.45e-07,
|
| 1064 |
+
"loss": 6.0512,
|
| 1065 |
+
"step": 639
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 71.0,
|
| 1069 |
+
"eval_accuracy": 0.2961876832844575,
|
| 1070 |
+
"eval_loss": 5.863522052764893,
|
| 1071 |
+
"eval_runtime": 4.2876,
|
| 1072 |
+
"eval_samples_per_second": 0.233,
|
| 1073 |
+
"eval_steps_per_second": 0.233,
|
| 1074 |
+
"step": 639
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 72.0,
|
| 1078 |
+
"learning_rate": 6.4e-07,
|
| 1079 |
+
"loss": 6.0415,
|
| 1080 |
+
"step": 648
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 72.0,
|
| 1084 |
+
"eval_accuracy": 0.2961876832844575,
|
| 1085 |
+
"eval_loss": 5.861262321472168,
|
| 1086 |
+
"eval_runtime": 4.932,
|
| 1087 |
+
"eval_samples_per_second": 0.203,
|
| 1088 |
+
"eval_steps_per_second": 0.203,
|
| 1089 |
+
"step": 648
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 73.0,
|
| 1093 |
+
"learning_rate": 6.35e-07,
|
| 1094 |
+
"loss": 6.0391,
|
| 1095 |
+
"step": 657
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 73.0,
|
| 1099 |
+
"eval_accuracy": 0.2961876832844575,
|
| 1100 |
+
"eval_loss": 5.858279228210449,
|
| 1101 |
+
"eval_runtime": 4.4327,
|
| 1102 |
+
"eval_samples_per_second": 0.226,
|
| 1103 |
+
"eval_steps_per_second": 0.226,
|
| 1104 |
+
"step": 657
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 74.0,
|
| 1108 |
+
"learning_rate": 6.3e-07,
|
| 1109 |
+
"loss": 6.032,
|
| 1110 |
+
"step": 666
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 74.0,
|
| 1114 |
+
"eval_accuracy": 0.2961876832844575,
|
| 1115 |
+
"eval_loss": 5.854796886444092,
|
| 1116 |
+
"eval_runtime": 4.4253,
|
| 1117 |
+
"eval_samples_per_second": 0.226,
|
| 1118 |
+
"eval_steps_per_second": 0.226,
|
| 1119 |
+
"step": 666
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 75.0,
|
| 1123 |
+
"learning_rate": 6.249999999999999e-07,
|
| 1124 |
+
"loss": 6.0317,
|
| 1125 |
+
"step": 675
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 75.0,
|
| 1129 |
+
"eval_accuracy": 0.2961876832844575,
|
| 1130 |
+
"eval_loss": 5.851076126098633,
|
| 1131 |
+
"eval_runtime": 4.5175,
|
| 1132 |
+
"eval_samples_per_second": 0.221,
|
| 1133 |
+
"eval_steps_per_second": 0.221,
|
| 1134 |
+
"step": 675
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 76.0,
|
| 1138 |
+
"learning_rate": 6.2e-07,
|
| 1139 |
+
"loss": 6.0343,
|
| 1140 |
+
"step": 684
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 76.0,
|
| 1144 |
+
"eval_accuracy": 0.29716520039100686,
|
| 1145 |
+
"eval_loss": 5.847884178161621,
|
| 1146 |
+
"eval_runtime": 4.5177,
|
| 1147 |
+
"eval_samples_per_second": 0.221,
|
| 1148 |
+
"eval_steps_per_second": 0.221,
|
| 1149 |
+
"step": 684
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 77.0,
|
| 1153 |
+
"learning_rate": 6.149999999999999e-07,
|
| 1154 |
+
"loss": 6.0156,
|
| 1155 |
+
"step": 693
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 77.0,
|
| 1159 |
+
"eval_accuracy": 0.2981427174975562,
|
| 1160 |
+
"eval_loss": 5.843489646911621,
|
| 1161 |
+
"eval_runtime": 4.7261,
|
| 1162 |
+
"eval_samples_per_second": 0.212,
|
| 1163 |
+
"eval_steps_per_second": 0.212,
|
| 1164 |
+
"step": 693
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 78.0,
|
| 1168 |
+
"learning_rate": 6.1e-07,
|
| 1169 |
+
"loss": 6.0167,
|
| 1170 |
+
"step": 702
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 78.0,
|
| 1174 |
+
"eval_accuracy": 0.2981427174975562,
|
| 1175 |
+
"eval_loss": 5.8403215408325195,
|
| 1176 |
+
"eval_runtime": 4.3157,
|
| 1177 |
+
"eval_samples_per_second": 0.232,
|
| 1178 |
+
"eval_steps_per_second": 0.232,
|
| 1179 |
+
"step": 702
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 79.0,
|
| 1183 |
+
"learning_rate": 6.049999999999999e-07,
|
| 1184 |
+
"loss": 6.0052,
|
| 1185 |
+
"step": 711
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 79.0,
|
| 1189 |
+
"eval_accuracy": 0.2981427174975562,
|
| 1190 |
+
"eval_loss": 5.83635950088501,
|
| 1191 |
+
"eval_runtime": 4.3221,
|
| 1192 |
+
"eval_samples_per_second": 0.231,
|
| 1193 |
+
"eval_steps_per_second": 0.231,
|
| 1194 |
+
"step": 711
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 80.0,
|
| 1198 |
+
"learning_rate": 6e-07,
|
| 1199 |
+
"loss": 6.0057,
|
| 1200 |
+
"step": 720
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 80.0,
|
| 1204 |
+
"eval_accuracy": 0.2981427174975562,
|
| 1205 |
+
"eval_loss": 5.8336262702941895,
|
| 1206 |
+
"eval_runtime": 4.8317,
|
| 1207 |
+
"eval_samples_per_second": 0.207,
|
| 1208 |
+
"eval_steps_per_second": 0.207,
|
| 1209 |
+
"step": 720
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 81.0,
|
| 1213 |
+
"learning_rate": 5.949999999999999e-07,
|
| 1214 |
+
"loss": 6.0001,
|
| 1215 |
+
"step": 729
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 81.0,
|
| 1219 |
+
"eval_accuracy": 0.2981427174975562,
|
| 1220 |
+
"eval_loss": 5.8319411277771,
|
| 1221 |
+
"eval_runtime": 4.5299,
|
| 1222 |
+
"eval_samples_per_second": 0.221,
|
| 1223 |
+
"eval_steps_per_second": 0.221,
|
| 1224 |
+
"step": 729
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 82.0,
|
| 1228 |
+
"learning_rate": 5.9e-07,
|
| 1229 |
+
"loss": 6.001,
|
| 1230 |
+
"step": 738
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 82.0,
|
| 1234 |
+
"eval_accuracy": 0.2991202346041056,
|
| 1235 |
+
"eval_loss": 5.828815937042236,
|
| 1236 |
+
"eval_runtime": 4.4118,
|
| 1237 |
+
"eval_samples_per_second": 0.227,
|
| 1238 |
+
"eval_steps_per_second": 0.227,
|
| 1239 |
+
"step": 738
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 83.0,
|
| 1243 |
+
"learning_rate": 5.849999999999999e-07,
|
| 1244 |
+
"loss": 5.9905,
|
| 1245 |
+
"step": 747
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 83.0,
|
| 1249 |
+
"eval_accuracy": 0.3010752688172043,
|
| 1250 |
+
"eval_loss": 5.826550006866455,
|
| 1251 |
+
"eval_runtime": 4.3318,
|
| 1252 |
+
"eval_samples_per_second": 0.231,
|
| 1253 |
+
"eval_steps_per_second": 0.231,
|
| 1254 |
+
"step": 747
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 84.0,
|
| 1258 |
+
"learning_rate": 5.8e-07,
|
| 1259 |
+
"loss": 5.9906,
|
| 1260 |
+
"step": 756
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 84.0,
|
| 1264 |
+
"eval_accuracy": 0.3020527859237537,
|
| 1265 |
+
"eval_loss": 5.824170112609863,
|
| 1266 |
+
"eval_runtime": 4.4876,
|
| 1267 |
+
"eval_samples_per_second": 0.223,
|
| 1268 |
+
"eval_steps_per_second": 0.223,
|
| 1269 |
+
"step": 756
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 85.0,
|
| 1273 |
+
"learning_rate": 5.749999999999999e-07,
|
| 1274 |
+
"loss": 5.9862,
|
| 1275 |
+
"step": 765
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 85.0,
|
| 1279 |
+
"eval_accuracy": 0.3020527859237537,
|
| 1280 |
+
"eval_loss": 5.821638584136963,
|
| 1281 |
+
"eval_runtime": 4.7768,
|
| 1282 |
+
"eval_samples_per_second": 0.209,
|
| 1283 |
+
"eval_steps_per_second": 0.209,
|
| 1284 |
+
"step": 765
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 86.0,
|
| 1288 |
+
"learning_rate": 5.699999999999999e-07,
|
| 1289 |
+
"loss": 5.9829,
|
| 1290 |
+
"step": 774
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 86.0,
|
| 1294 |
+
"eval_accuracy": 0.30498533724340177,
|
| 1295 |
+
"eval_loss": 5.819102764129639,
|
| 1296 |
+
"eval_runtime": 4.2697,
|
| 1297 |
+
"eval_samples_per_second": 0.234,
|
| 1298 |
+
"eval_steps_per_second": 0.234,
|
| 1299 |
+
"step": 774
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 87.0,
|
| 1303 |
+
"learning_rate": 5.649999999999999e-07,
|
| 1304 |
+
"loss": 5.9725,
|
| 1305 |
+
"step": 783
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 87.0,
|
| 1309 |
+
"eval_accuracy": 0.3069403714565005,
|
| 1310 |
+
"eval_loss": 5.817332744598389,
|
| 1311 |
+
"eval_runtime": 4.3341,
|
| 1312 |
+
"eval_samples_per_second": 0.231,
|
| 1313 |
+
"eval_steps_per_second": 0.231,
|
| 1314 |
+
"step": 783
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 88.0,
|
| 1318 |
+
"learning_rate": 5.6e-07,
|
| 1319 |
+
"loss": 5.9795,
|
| 1320 |
+
"step": 792
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 88.0,
|
| 1324 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1325 |
+
"eval_loss": 5.814181804656982,
|
| 1326 |
+
"eval_runtime": 4.759,
|
| 1327 |
+
"eval_samples_per_second": 0.21,
|
| 1328 |
+
"eval_steps_per_second": 0.21,
|
| 1329 |
+
"step": 792
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 89.0,
|
| 1333 |
+
"learning_rate": 5.55e-07,
|
| 1334 |
+
"loss": 5.9695,
|
| 1335 |
+
"step": 801
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 89.0,
|
| 1339 |
+
"eval_accuracy": 0.3069403714565005,
|
| 1340 |
+
"eval_loss": 5.811527729034424,
|
| 1341 |
+
"eval_runtime": 4.3738,
|
| 1342 |
+
"eval_samples_per_second": 0.229,
|
| 1343 |
+
"eval_steps_per_second": 0.229,
|
| 1344 |
+
"step": 801
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 90.0,
|
| 1348 |
+
"learning_rate": 5.5e-07,
|
| 1349 |
+
"loss": 5.9607,
|
| 1350 |
+
"step": 810
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 90.0,
|
| 1354 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1355 |
+
"eval_loss": 5.808058261871338,
|
| 1356 |
+
"eval_runtime": 3.4007,
|
| 1357 |
+
"eval_samples_per_second": 0.294,
|
| 1358 |
+
"eval_steps_per_second": 0.294,
|
| 1359 |
+
"step": 810
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 91.0,
|
| 1363 |
+
"learning_rate": 5.45e-07,
|
| 1364 |
+
"loss": 5.9605,
|
| 1365 |
+
"step": 819
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 91.0,
|
| 1369 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1370 |
+
"eval_loss": 5.805708885192871,
|
| 1371 |
+
"eval_runtime": 3.8195,
|
| 1372 |
+
"eval_samples_per_second": 0.262,
|
| 1373 |
+
"eval_steps_per_second": 0.262,
|
| 1374 |
+
"step": 819
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 92.0,
|
| 1378 |
+
"learning_rate": 5.4e-07,
|
| 1379 |
+
"loss": 5.9591,
|
| 1380 |
+
"step": 828
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 92.0,
|
| 1384 |
+
"eval_accuracy": 0.30791788856304986,
|
| 1385 |
+
"eval_loss": 5.80366849899292,
|
| 1386 |
+
"eval_runtime": 4.5002,
|
| 1387 |
+
"eval_samples_per_second": 0.222,
|
| 1388 |
+
"eval_steps_per_second": 0.222,
|
| 1389 |
+
"step": 828
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 93.0,
|
| 1393 |
+
"learning_rate": 5.35e-07,
|
| 1394 |
+
"loss": 5.9481,
|
| 1395 |
+
"step": 837
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 93.0,
|
| 1399 |
+
"eval_accuracy": 0.3069403714565005,
|
| 1400 |
+
"eval_loss": 5.801074981689453,
|
| 1401 |
+
"eval_runtime": 4.3082,
|
| 1402 |
+
"eval_samples_per_second": 0.232,
|
| 1403 |
+
"eval_steps_per_second": 0.232,
|
| 1404 |
+
"step": 837
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 94.0,
|
| 1408 |
+
"learning_rate": 5.3e-07,
|
| 1409 |
+
"loss": 5.9501,
|
| 1410 |
+
"step": 846
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 94.0,
|
| 1414 |
+
"eval_accuracy": 0.30791788856304986,
|
| 1415 |
+
"eval_loss": 5.798312664031982,
|
| 1416 |
+
"eval_runtime": 4.9824,
|
| 1417 |
+
"eval_samples_per_second": 0.201,
|
| 1418 |
+
"eval_steps_per_second": 0.201,
|
| 1419 |
+
"step": 846
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 95.0,
|
| 1423 |
+
"learning_rate": 5.25e-07,
|
| 1424 |
+
"loss": 5.948,
|
| 1425 |
+
"step": 855
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 95.0,
|
| 1429 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1430 |
+
"eval_loss": 5.794301509857178,
|
| 1431 |
+
"eval_runtime": 3.717,
|
| 1432 |
+
"eval_samples_per_second": 0.269,
|
| 1433 |
+
"eval_steps_per_second": 0.269,
|
| 1434 |
+
"step": 855
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 96.0,
|
| 1438 |
+
"learning_rate": 5.2e-07,
|
| 1439 |
+
"loss": 5.9488,
|
| 1440 |
+
"step": 864
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 96.0,
|
| 1444 |
+
"eval_accuracy": 0.31085043988269795,
|
| 1445 |
+
"eval_loss": 5.790666580200195,
|
| 1446 |
+
"eval_runtime": 3.5092,
|
| 1447 |
+
"eval_samples_per_second": 0.285,
|
| 1448 |
+
"eval_steps_per_second": 0.285,
|
| 1449 |
+
"step": 864
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 97.0,
|
| 1453 |
+
"learning_rate": 5.149999999999999e-07,
|
| 1454 |
+
"loss": 5.9449,
|
| 1455 |
+
"step": 873
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 97.0,
|
| 1459 |
+
"eval_accuracy": 0.3118279569892473,
|
| 1460 |
+
"eval_loss": 5.788764476776123,
|
| 1461 |
+
"eval_runtime": 4.3467,
|
| 1462 |
+
"eval_samples_per_second": 0.23,
|
| 1463 |
+
"eval_steps_per_second": 0.23,
|
| 1464 |
+
"step": 873
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 98.0,
|
| 1468 |
+
"learning_rate": 5.1e-07,
|
| 1469 |
+
"loss": 5.9357,
|
| 1470 |
+
"step": 882
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 98.0,
|
| 1474 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1475 |
+
"eval_loss": 5.787247180938721,
|
| 1476 |
+
"eval_runtime": 4.2514,
|
| 1477 |
+
"eval_samples_per_second": 0.235,
|
| 1478 |
+
"eval_steps_per_second": 0.235,
|
| 1479 |
+
"step": 882
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 99.0,
|
| 1483 |
+
"learning_rate": 5.049999999999999e-07,
|
| 1484 |
+
"loss": 5.9363,
|
| 1485 |
+
"step": 891
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 99.0,
|
| 1489 |
+
"eval_accuracy": 0.30791788856304986,
|
| 1490 |
+
"eval_loss": 5.783356666564941,
|
| 1491 |
+
"eval_runtime": 3.6762,
|
| 1492 |
+
"eval_samples_per_second": 0.272,
|
| 1493 |
+
"eval_steps_per_second": 0.272,
|
| 1494 |
+
"step": 891
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 100.0,
|
| 1498 |
+
"learning_rate": 5e-07,
|
| 1499 |
+
"loss": 5.9368,
|
| 1500 |
+
"step": 900
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 100.0,
|
| 1504 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1505 |
+
"eval_loss": 5.781695365905762,
|
| 1506 |
+
"eval_runtime": 4.5979,
|
| 1507 |
+
"eval_samples_per_second": 0.217,
|
| 1508 |
+
"eval_steps_per_second": 0.217,
|
| 1509 |
+
"step": 900
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 101.0,
|
| 1513 |
+
"learning_rate": 4.95e-07,
|
| 1514 |
+
"loss": 5.9215,
|
| 1515 |
+
"step": 909
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 101.0,
|
| 1519 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1520 |
+
"eval_loss": 5.779068946838379,
|
| 1521 |
+
"eval_runtime": 4.3432,
|
| 1522 |
+
"eval_samples_per_second": 0.23,
|
| 1523 |
+
"eval_steps_per_second": 0.23,
|
| 1524 |
+
"step": 909
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 102.0,
|
| 1528 |
+
"learning_rate": 4.9e-07,
|
| 1529 |
+
"loss": 5.9264,
|
| 1530 |
+
"step": 918
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 102.0,
|
| 1534 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1535 |
+
"eval_loss": 5.77487325668335,
|
| 1536 |
+
"eval_runtime": 3.8835,
|
| 1537 |
+
"eval_samples_per_second": 0.258,
|
| 1538 |
+
"eval_steps_per_second": 0.258,
|
| 1539 |
+
"step": 918
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 103.0,
|
| 1543 |
+
"learning_rate": 4.85e-07,
|
| 1544 |
+
"loss": 5.9126,
|
| 1545 |
+
"step": 927
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 103.0,
|
| 1549 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1550 |
+
"eval_loss": 5.772488594055176,
|
| 1551 |
+
"eval_runtime": 4.0354,
|
| 1552 |
+
"eval_samples_per_second": 0.248,
|
| 1553 |
+
"eval_steps_per_second": 0.248,
|
| 1554 |
+
"step": 927
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 104.0,
|
| 1558 |
+
"learning_rate": 4.8e-07,
|
| 1559 |
+
"loss": 5.9163,
|
| 1560 |
+
"step": 936
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 104.0,
|
| 1564 |
+
"eval_accuracy": 0.3069403714565005,
|
| 1565 |
+
"eval_loss": 5.769767761230469,
|
| 1566 |
+
"eval_runtime": 3.6621,
|
| 1567 |
+
"eval_samples_per_second": 0.273,
|
| 1568 |
+
"eval_steps_per_second": 0.273,
|
| 1569 |
+
"step": 936
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 105.0,
|
| 1573 |
+
"learning_rate": 4.7499999999999995e-07,
|
| 1574 |
+
"loss": 5.9183,
|
| 1575 |
+
"step": 945
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 105.0,
|
| 1579 |
+
"eval_accuracy": 0.3069403714565005,
|
| 1580 |
+
"eval_loss": 5.767302989959717,
|
| 1581 |
+
"eval_runtime": 4.3085,
|
| 1582 |
+
"eval_samples_per_second": 0.232,
|
| 1583 |
+
"eval_steps_per_second": 0.232,
|
| 1584 |
+
"step": 945
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 106.0,
|
| 1588 |
+
"learning_rate": 4.6999999999999995e-07,
|
| 1589 |
+
"loss": 5.9154,
|
| 1590 |
+
"step": 954
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 106.0,
|
| 1594 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1595 |
+
"eval_loss": 5.763917446136475,
|
| 1596 |
+
"eval_runtime": 4.2316,
|
| 1597 |
+
"eval_samples_per_second": 0.236,
|
| 1598 |
+
"eval_steps_per_second": 0.236,
|
| 1599 |
+
"step": 954
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 107.0,
|
| 1603 |
+
"learning_rate": 4.65e-07,
|
| 1604 |
+
"loss": 5.9149,
|
| 1605 |
+
"step": 963
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 107.0,
|
| 1609 |
+
"eval_accuracy": 0.30791788856304986,
|
| 1610 |
+
"eval_loss": 5.76121187210083,
|
| 1611 |
+
"eval_runtime": 4.5023,
|
| 1612 |
+
"eval_samples_per_second": 0.222,
|
| 1613 |
+
"eval_steps_per_second": 0.222,
|
| 1614 |
+
"step": 963
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 108.0,
|
| 1618 |
+
"learning_rate": 4.6e-07,
|
| 1619 |
+
"loss": 5.9082,
|
| 1620 |
+
"step": 972
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 108.0,
|
| 1624 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1625 |
+
"eval_loss": 5.757934093475342,
|
| 1626 |
+
"eval_runtime": 4.5081,
|
| 1627 |
+
"eval_samples_per_second": 0.222,
|
| 1628 |
+
"eval_steps_per_second": 0.222,
|
| 1629 |
+
"step": 972
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 109.0,
|
| 1633 |
+
"learning_rate": 4.55e-07,
|
| 1634 |
+
"loss": 5.9051,
|
| 1635 |
+
"step": 981
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 109.0,
|
| 1639 |
+
"eval_accuracy": 0.30791788856304986,
|
| 1640 |
+
"eval_loss": 5.755875110626221,
|
| 1641 |
+
"eval_runtime": 4.207,
|
| 1642 |
+
"eval_samples_per_second": 0.238,
|
| 1643 |
+
"eval_steps_per_second": 0.238,
|
| 1644 |
+
"step": 981
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 110.0,
|
| 1648 |
+
"learning_rate": 4.5e-07,
|
| 1649 |
+
"loss": 5.908,
|
| 1650 |
+
"step": 990
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 110.0,
|
| 1654 |
+
"eval_accuracy": 0.30791788856304986,
|
| 1655 |
+
"eval_loss": 5.753073692321777,
|
| 1656 |
+
"eval_runtime": 4.5603,
|
| 1657 |
+
"eval_samples_per_second": 0.219,
|
| 1658 |
+
"eval_steps_per_second": 0.219,
|
| 1659 |
+
"step": 990
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 111.0,
|
| 1663 |
+
"learning_rate": 4.45e-07,
|
| 1664 |
+
"loss": 5.901,
|
| 1665 |
+
"step": 999
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 111.0,
|
| 1669 |
+
"eval_accuracy": 0.30791788856304986,
|
| 1670 |
+
"eval_loss": 5.750200271606445,
|
| 1671 |
+
"eval_runtime": 4.3073,
|
| 1672 |
+
"eval_samples_per_second": 0.232,
|
| 1673 |
+
"eval_steps_per_second": 0.232,
|
| 1674 |
+
"step": 999
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 112.0,
|
| 1678 |
+
"learning_rate": 4.3999999999999997e-07,
|
| 1679 |
+
"loss": 5.9064,
|
| 1680 |
+
"step": 1008
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 112.0,
|
| 1684 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1685 |
+
"eval_loss": 5.748438835144043,
|
| 1686 |
+
"eval_runtime": 4.0892,
|
| 1687 |
+
"eval_samples_per_second": 0.245,
|
| 1688 |
+
"eval_steps_per_second": 0.245,
|
| 1689 |
+
"step": 1008
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 113.0,
|
| 1693 |
+
"learning_rate": 4.3499999999999996e-07,
|
| 1694 |
+
"loss": 5.895,
|
| 1695 |
+
"step": 1017
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 113.0,
|
| 1699 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1700 |
+
"eval_loss": 5.745972156524658,
|
| 1701 |
+
"eval_runtime": 3.6273,
|
| 1702 |
+
"eval_samples_per_second": 0.276,
|
| 1703 |
+
"eval_steps_per_second": 0.276,
|
| 1704 |
+
"step": 1017
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 114.0,
|
| 1708 |
+
"learning_rate": 4.2999999999999996e-07,
|
| 1709 |
+
"loss": 5.894,
|
| 1710 |
+
"step": 1026
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 114.0,
|
| 1714 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1715 |
+
"eval_loss": 5.743812561035156,
|
| 1716 |
+
"eval_runtime": 3.373,
|
| 1717 |
+
"eval_samples_per_second": 0.296,
|
| 1718 |
+
"eval_steps_per_second": 0.296,
|
| 1719 |
+
"step": 1026
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 115.0,
|
| 1723 |
+
"learning_rate": 4.2499999999999995e-07,
|
| 1724 |
+
"loss": 5.8809,
|
| 1725 |
+
"step": 1035
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 115.0,
|
| 1729 |
+
"eval_accuracy": 0.30791788856304986,
|
| 1730 |
+
"eval_loss": 5.741939067840576,
|
| 1731 |
+
"eval_runtime": 3.9496,
|
| 1732 |
+
"eval_samples_per_second": 0.253,
|
| 1733 |
+
"eval_steps_per_second": 0.253,
|
| 1734 |
+
"step": 1035
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 116.0,
|
| 1738 |
+
"learning_rate": 4.1999999999999995e-07,
|
| 1739 |
+
"loss": 5.8893,
|
| 1740 |
+
"step": 1044
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 116.0,
|
| 1744 |
+
"eval_accuracy": 0.30791788856304986,
|
| 1745 |
+
"eval_loss": 5.74192476272583,
|
| 1746 |
+
"eval_runtime": 4.7772,
|
| 1747 |
+
"eval_samples_per_second": 0.209,
|
| 1748 |
+
"eval_steps_per_second": 0.209,
|
| 1749 |
+
"step": 1044
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 117.0,
|
| 1753 |
+
"learning_rate": 4.1499999999999994e-07,
|
| 1754 |
+
"loss": 5.8874,
|
| 1755 |
+
"step": 1053
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 117.0,
|
| 1759 |
+
"eval_accuracy": 0.3069403714565005,
|
| 1760 |
+
"eval_loss": 5.739171504974365,
|
| 1761 |
+
"eval_runtime": 4.0083,
|
| 1762 |
+
"eval_samples_per_second": 0.249,
|
| 1763 |
+
"eval_steps_per_second": 0.249,
|
| 1764 |
+
"step": 1053
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 118.0,
|
| 1768 |
+
"learning_rate": 4.0999999999999994e-07,
|
| 1769 |
+
"loss": 5.8798,
|
| 1770 |
+
"step": 1062
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 118.0,
|
| 1774 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1775 |
+
"eval_loss": 5.73491907119751,
|
| 1776 |
+
"eval_runtime": 4.1773,
|
| 1777 |
+
"eval_samples_per_second": 0.239,
|
| 1778 |
+
"eval_steps_per_second": 0.239,
|
| 1779 |
+
"step": 1062
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 119.0,
|
| 1783 |
+
"learning_rate": 4.05e-07,
|
| 1784 |
+
"loss": 5.8826,
|
| 1785 |
+
"step": 1071
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 119.0,
|
| 1789 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1790 |
+
"eval_loss": 5.732446193695068,
|
| 1791 |
+
"eval_runtime": 4.4141,
|
| 1792 |
+
"eval_samples_per_second": 0.227,
|
| 1793 |
+
"eval_steps_per_second": 0.227,
|
| 1794 |
+
"step": 1071
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 120.0,
|
| 1798 |
+
"learning_rate": 4e-07,
|
| 1799 |
+
"loss": 5.8736,
|
| 1800 |
+
"step": 1080
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 120.0,
|
| 1804 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1805 |
+
"eval_loss": 5.729852199554443,
|
| 1806 |
+
"eval_runtime": 4.4117,
|
| 1807 |
+
"eval_samples_per_second": 0.227,
|
| 1808 |
+
"eval_steps_per_second": 0.227,
|
| 1809 |
+
"step": 1080
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 121.0,
|
| 1813 |
+
"learning_rate": 3.95e-07,
|
| 1814 |
+
"loss": 5.8751,
|
| 1815 |
+
"step": 1089
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 121.0,
|
| 1819 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1820 |
+
"eval_loss": 5.726986408233643,
|
| 1821 |
+
"eval_runtime": 4.9388,
|
| 1822 |
+
"eval_samples_per_second": 0.202,
|
| 1823 |
+
"eval_steps_per_second": 0.202,
|
| 1824 |
+
"step": 1089
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 122.0,
|
| 1828 |
+
"learning_rate": 3.8999999999999997e-07,
|
| 1829 |
+
"loss": 5.8699,
|
| 1830 |
+
"step": 1098
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 122.0,
|
| 1834 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1835 |
+
"eval_loss": 5.72526216506958,
|
| 1836 |
+
"eval_runtime": 4.0541,
|
| 1837 |
+
"eval_samples_per_second": 0.247,
|
| 1838 |
+
"eval_steps_per_second": 0.247,
|
| 1839 |
+
"step": 1098
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 123.0,
|
| 1843 |
+
"learning_rate": 3.8499999999999997e-07,
|
| 1844 |
+
"loss": 5.8802,
|
| 1845 |
+
"step": 1107
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 123.0,
|
| 1849 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1850 |
+
"eval_loss": 5.723138809204102,
|
| 1851 |
+
"eval_runtime": 4.1908,
|
| 1852 |
+
"eval_samples_per_second": 0.239,
|
| 1853 |
+
"eval_steps_per_second": 0.239,
|
| 1854 |
+
"step": 1107
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 124.0,
|
| 1858 |
+
"learning_rate": 3.7999999999999996e-07,
|
| 1859 |
+
"loss": 5.8707,
|
| 1860 |
+
"step": 1116
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 124.0,
|
| 1864 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1865 |
+
"eval_loss": 5.7239861488342285,
|
| 1866 |
+
"eval_runtime": 4.4694,
|
| 1867 |
+
"eval_samples_per_second": 0.224,
|
| 1868 |
+
"eval_steps_per_second": 0.224,
|
| 1869 |
+
"step": 1116
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 125.0,
|
| 1873 |
+
"learning_rate": 3.75e-07,
|
| 1874 |
+
"loss": 5.8653,
|
| 1875 |
+
"step": 1125
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 125.0,
|
| 1879 |
+
"eval_accuracy": 0.31085043988269795,
|
| 1880 |
+
"eval_loss": 5.723176002502441,
|
| 1881 |
+
"eval_runtime": 4.2543,
|
| 1882 |
+
"eval_samples_per_second": 0.235,
|
| 1883 |
+
"eval_steps_per_second": 0.235,
|
| 1884 |
+
"step": 1125
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 126.0,
|
| 1888 |
+
"learning_rate": 3.7e-07,
|
| 1889 |
+
"loss": 5.8693,
|
| 1890 |
+
"step": 1134
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 126.0,
|
| 1894 |
+
"eval_accuracy": 0.31085043988269795,
|
| 1895 |
+
"eval_loss": 5.7180495262146,
|
| 1896 |
+
"eval_runtime": 4.3031,
|
| 1897 |
+
"eval_samples_per_second": 0.232,
|
| 1898 |
+
"eval_steps_per_second": 0.232,
|
| 1899 |
+
"step": 1134
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 127.0,
|
| 1903 |
+
"learning_rate": 3.65e-07,
|
| 1904 |
+
"loss": 5.8662,
|
| 1905 |
+
"step": 1143
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 127.0,
|
| 1909 |
+
"eval_accuracy": 0.31085043988269795,
|
| 1910 |
+
"eval_loss": 5.714703559875488,
|
| 1911 |
+
"eval_runtime": 4.1167,
|
| 1912 |
+
"eval_samples_per_second": 0.243,
|
| 1913 |
+
"eval_steps_per_second": 0.243,
|
| 1914 |
+
"step": 1143
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 128.0,
|
| 1918 |
+
"learning_rate": 3.6e-07,
|
| 1919 |
+
"loss": 5.8539,
|
| 1920 |
+
"step": 1152
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 128.0,
|
| 1924 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1925 |
+
"eval_loss": 5.713204383850098,
|
| 1926 |
+
"eval_runtime": 4.0648,
|
| 1927 |
+
"eval_samples_per_second": 0.246,
|
| 1928 |
+
"eval_steps_per_second": 0.246,
|
| 1929 |
+
"step": 1152
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 129.0,
|
| 1933 |
+
"learning_rate": 3.55e-07,
|
| 1934 |
+
"loss": 5.8611,
|
| 1935 |
+
"step": 1161
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 129.0,
|
| 1939 |
+
"eval_accuracy": 0.3098729227761486,
|
| 1940 |
+
"eval_loss": 5.712745189666748,
|
| 1941 |
+
"eval_runtime": 4.0643,
|
| 1942 |
+
"eval_samples_per_second": 0.246,
|
| 1943 |
+
"eval_steps_per_second": 0.246,
|
| 1944 |
+
"step": 1161
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 130.0,
|
| 1948 |
+
"learning_rate": 3.5e-07,
|
| 1949 |
+
"loss": 5.8495,
|
| 1950 |
+
"step": 1170
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 130.0,
|
| 1954 |
+
"eval_accuracy": 0.30791788856304986,
|
| 1955 |
+
"eval_loss": 5.713523864746094,
|
| 1956 |
+
"eval_runtime": 4.0466,
|
| 1957 |
+
"eval_samples_per_second": 0.247,
|
| 1958 |
+
"eval_steps_per_second": 0.247,
|
| 1959 |
+
"step": 1170
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 131.0,
|
| 1963 |
+
"learning_rate": 3.45e-07,
|
| 1964 |
+
"loss": 5.8602,
|
| 1965 |
+
"step": 1179
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 131.0,
|
| 1969 |
+
"eval_accuracy": 0.3088954056695992,
|
| 1970 |
+
"eval_loss": 5.711104393005371,
|
| 1971 |
+
"eval_runtime": 3.9889,
|
| 1972 |
+
"eval_samples_per_second": 0.251,
|
| 1973 |
+
"eval_steps_per_second": 0.251,
|
| 1974 |
+
"step": 1179
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 132.0,
|
| 1978 |
+
"learning_rate": 3.4000000000000003e-07,
|
| 1979 |
+
"loss": 5.8512,
|
| 1980 |
+
"step": 1188
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 132.0,
|
| 1984 |
+
"eval_accuracy": 0.31085043988269795,
|
| 1985 |
+
"eval_loss": 5.7077484130859375,
|
| 1986 |
+
"eval_runtime": 3.6772,
|
| 1987 |
+
"eval_samples_per_second": 0.272,
|
| 1988 |
+
"eval_steps_per_second": 0.272,
|
| 1989 |
+
"step": 1188
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 133.0,
|
| 1993 |
+
"learning_rate": 3.35e-07,
|
| 1994 |
+
"loss": 5.8493,
|
| 1995 |
+
"step": 1197
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 133.0,
|
| 1999 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2000 |
+
"eval_loss": 5.704965591430664,
|
| 2001 |
+
"eval_runtime": 4.3621,
|
| 2002 |
+
"eval_samples_per_second": 0.229,
|
| 2003 |
+
"eval_steps_per_second": 0.229,
|
| 2004 |
+
"step": 1197
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 134.0,
|
| 2008 |
+
"learning_rate": 3.3e-07,
|
| 2009 |
+
"loss": 5.8477,
|
| 2010 |
+
"step": 1206
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 134.0,
|
| 2014 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2015 |
+
"eval_loss": 5.704052925109863,
|
| 2016 |
+
"eval_runtime": 4.4972,
|
| 2017 |
+
"eval_samples_per_second": 0.222,
|
| 2018 |
+
"eval_steps_per_second": 0.222,
|
| 2019 |
+
"step": 1206
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 135.0,
|
| 2023 |
+
"learning_rate": 3.25e-07,
|
| 2024 |
+
"loss": 5.8464,
|
| 2025 |
+
"step": 1215
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 135.0,
|
| 2029 |
+
"eval_accuracy": 0.3088954056695992,
|
| 2030 |
+
"eval_loss": 5.704152584075928,
|
| 2031 |
+
"eval_runtime": 3.8106,
|
| 2032 |
+
"eval_samples_per_second": 0.262,
|
| 2033 |
+
"eval_steps_per_second": 0.262,
|
| 2034 |
+
"step": 1215
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 136.0,
|
| 2038 |
+
"learning_rate": 3.2e-07,
|
| 2039 |
+
"loss": 5.8459,
|
| 2040 |
+
"step": 1224
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 136.0,
|
| 2044 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2045 |
+
"eval_loss": 5.702281951904297,
|
| 2046 |
+
"eval_runtime": 4.7548,
|
| 2047 |
+
"eval_samples_per_second": 0.21,
|
| 2048 |
+
"eval_steps_per_second": 0.21,
|
| 2049 |
+
"step": 1224
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 137.0,
|
| 2053 |
+
"learning_rate": 3.15e-07,
|
| 2054 |
+
"loss": 5.8475,
|
| 2055 |
+
"step": 1233
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 137.0,
|
| 2059 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2060 |
+
"eval_loss": 5.7000274658203125,
|
| 2061 |
+
"eval_runtime": 4.1681,
|
| 2062 |
+
"eval_samples_per_second": 0.24,
|
| 2063 |
+
"eval_steps_per_second": 0.24,
|
| 2064 |
+
"step": 1233
|
| 2065 |
+
},
|
| 2066 |
+
{
|
| 2067 |
+
"epoch": 138.0,
|
| 2068 |
+
"learning_rate": 3.1e-07,
|
| 2069 |
+
"loss": 5.8384,
|
| 2070 |
+
"step": 1242
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 138.0,
|
| 2074 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2075 |
+
"eval_loss": 5.698208332061768,
|
| 2076 |
+
"eval_runtime": 4.6023,
|
| 2077 |
+
"eval_samples_per_second": 0.217,
|
| 2078 |
+
"eval_steps_per_second": 0.217,
|
| 2079 |
+
"step": 1242
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 139.0,
|
| 2083 |
+
"learning_rate": 3.05e-07,
|
| 2084 |
+
"loss": 5.8453,
|
| 2085 |
+
"step": 1251
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 139.0,
|
| 2089 |
+
"eval_accuracy": 0.3088954056695992,
|
| 2090 |
+
"eval_loss": 5.697631359100342,
|
| 2091 |
+
"eval_runtime": 3.8429,
|
| 2092 |
+
"eval_samples_per_second": 0.26,
|
| 2093 |
+
"eval_steps_per_second": 0.26,
|
| 2094 |
+
"step": 1251
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 140.0,
|
| 2098 |
+
"learning_rate": 3e-07,
|
| 2099 |
+
"loss": 5.8441,
|
| 2100 |
+
"step": 1260
|
| 2101 |
+
},
|
| 2102 |
+
{
|
| 2103 |
+
"epoch": 140.0,
|
| 2104 |
+
"eval_accuracy": 0.3088954056695992,
|
| 2105 |
+
"eval_loss": 5.697261810302734,
|
| 2106 |
+
"eval_runtime": 4.2789,
|
| 2107 |
+
"eval_samples_per_second": 0.234,
|
| 2108 |
+
"eval_steps_per_second": 0.234,
|
| 2109 |
+
"step": 1260
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 141.0,
|
| 2113 |
+
"learning_rate": 2.95e-07,
|
| 2114 |
+
"loss": 5.838,
|
| 2115 |
+
"step": 1269
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 141.0,
|
| 2119 |
+
"eval_accuracy": 0.30791788856304986,
|
| 2120 |
+
"eval_loss": 5.697081565856934,
|
| 2121 |
+
"eval_runtime": 4.6381,
|
| 2122 |
+
"eval_samples_per_second": 0.216,
|
| 2123 |
+
"eval_steps_per_second": 0.216,
|
| 2124 |
+
"step": 1269
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 142.0,
|
| 2128 |
+
"learning_rate": 2.9e-07,
|
| 2129 |
+
"loss": 5.8463,
|
| 2130 |
+
"step": 1278
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 142.0,
|
| 2134 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2135 |
+
"eval_loss": 5.695047855377197,
|
| 2136 |
+
"eval_runtime": 4.2254,
|
| 2137 |
+
"eval_samples_per_second": 0.237,
|
| 2138 |
+
"eval_steps_per_second": 0.237,
|
| 2139 |
+
"step": 1278
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 143.0,
|
| 2143 |
+
"learning_rate": 2.8499999999999997e-07,
|
| 2144 |
+
"loss": 5.8385,
|
| 2145 |
+
"step": 1287
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 143.0,
|
| 2149 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2150 |
+
"eval_loss": 5.692078113555908,
|
| 2151 |
+
"eval_runtime": 3.9314,
|
| 2152 |
+
"eval_samples_per_second": 0.254,
|
| 2153 |
+
"eval_steps_per_second": 0.254,
|
| 2154 |
+
"step": 1287
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 144.0,
|
| 2158 |
+
"learning_rate": 2.8e-07,
|
| 2159 |
+
"loss": 5.8354,
|
| 2160 |
+
"step": 1296
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 144.0,
|
| 2164 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2165 |
+
"eval_loss": 5.69093656539917,
|
| 2166 |
+
"eval_runtime": 4.4756,
|
| 2167 |
+
"eval_samples_per_second": 0.223,
|
| 2168 |
+
"eval_steps_per_second": 0.223,
|
| 2169 |
+
"step": 1296
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 145.0,
|
| 2173 |
+
"learning_rate": 2.75e-07,
|
| 2174 |
+
"loss": 5.8283,
|
| 2175 |
+
"step": 1305
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 145.0,
|
| 2179 |
+
"eval_accuracy": 0.30791788856304986,
|
| 2180 |
+
"eval_loss": 5.690831184387207,
|
| 2181 |
+
"eval_runtime": 4.2673,
|
| 2182 |
+
"eval_samples_per_second": 0.234,
|
| 2183 |
+
"eval_steps_per_second": 0.234,
|
| 2184 |
+
"step": 1305
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 146.0,
|
| 2188 |
+
"learning_rate": 2.7e-07,
|
| 2189 |
+
"loss": 5.8363,
|
| 2190 |
+
"step": 1314
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 146.0,
|
| 2194 |
+
"eval_accuracy": 0.30791788856304986,
|
| 2195 |
+
"eval_loss": 5.6901655197143555,
|
| 2196 |
+
"eval_runtime": 4.2991,
|
| 2197 |
+
"eval_samples_per_second": 0.233,
|
| 2198 |
+
"eval_steps_per_second": 0.233,
|
| 2199 |
+
"step": 1314
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 147.0,
|
| 2203 |
+
"learning_rate": 2.65e-07,
|
| 2204 |
+
"loss": 5.8433,
|
| 2205 |
+
"step": 1323
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 147.0,
|
| 2209 |
+
"eval_accuracy": 0.30791788856304986,
|
| 2210 |
+
"eval_loss": 5.689047813415527,
|
| 2211 |
+
"eval_runtime": 4.5899,
|
| 2212 |
+
"eval_samples_per_second": 0.218,
|
| 2213 |
+
"eval_steps_per_second": 0.218,
|
| 2214 |
+
"step": 1323
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 148.0,
|
| 2218 |
+
"learning_rate": 2.6e-07,
|
| 2219 |
+
"loss": 5.8302,
|
| 2220 |
+
"step": 1332
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 148.0,
|
| 2224 |
+
"eval_accuracy": 0.30791788856304986,
|
| 2225 |
+
"eval_loss": 5.689029216766357,
|
| 2226 |
+
"eval_runtime": 4.2306,
|
| 2227 |
+
"eval_samples_per_second": 0.236,
|
| 2228 |
+
"eval_steps_per_second": 0.236,
|
| 2229 |
+
"step": 1332
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 149.0,
|
| 2233 |
+
"learning_rate": 2.55e-07,
|
| 2234 |
+
"loss": 5.8276,
|
| 2235 |
+
"step": 1341
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 149.0,
|
| 2239 |
+
"eval_accuracy": 0.3088954056695992,
|
| 2240 |
+
"eval_loss": 5.688091278076172,
|
| 2241 |
+
"eval_runtime": 4.1426,
|
| 2242 |
+
"eval_samples_per_second": 0.241,
|
| 2243 |
+
"eval_steps_per_second": 0.241,
|
| 2244 |
+
"step": 1341
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 150.0,
|
| 2248 |
+
"learning_rate": 2.5e-07,
|
| 2249 |
+
"loss": 5.8366,
|
| 2250 |
+
"step": 1350
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 150.0,
|
| 2254 |
+
"eval_accuracy": 0.3088954056695992,
|
| 2255 |
+
"eval_loss": 5.686432838439941,
|
| 2256 |
+
"eval_runtime": 4.4478,
|
| 2257 |
+
"eval_samples_per_second": 0.225,
|
| 2258 |
+
"eval_steps_per_second": 0.225,
|
| 2259 |
+
"step": 1350
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 151.0,
|
| 2263 |
+
"learning_rate": 2.45e-07,
|
| 2264 |
+
"loss": 5.826,
|
| 2265 |
+
"step": 1359
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 151.0,
|
| 2269 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2270 |
+
"eval_loss": 5.685244083404541,
|
| 2271 |
+
"eval_runtime": 4.0347,
|
| 2272 |
+
"eval_samples_per_second": 0.248,
|
| 2273 |
+
"eval_steps_per_second": 0.248,
|
| 2274 |
+
"step": 1359
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 152.0,
|
| 2278 |
+
"learning_rate": 2.4e-07,
|
| 2279 |
+
"loss": 5.8293,
|
| 2280 |
+
"step": 1368
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 152.0,
|
| 2284 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2285 |
+
"eval_loss": 5.68437385559082,
|
| 2286 |
+
"eval_runtime": 3.7004,
|
| 2287 |
+
"eval_samples_per_second": 0.27,
|
| 2288 |
+
"eval_steps_per_second": 0.27,
|
| 2289 |
+
"step": 1368
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 153.0,
|
| 2293 |
+
"learning_rate": 2.3499999999999997e-07,
|
| 2294 |
+
"loss": 5.8278,
|
| 2295 |
+
"step": 1377
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 153.0,
|
| 2299 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2300 |
+
"eval_loss": 5.683380603790283,
|
| 2301 |
+
"eval_runtime": 4.0294,
|
| 2302 |
+
"eval_samples_per_second": 0.248,
|
| 2303 |
+
"eval_steps_per_second": 0.248,
|
| 2304 |
+
"step": 1377
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 154.0,
|
| 2308 |
+
"learning_rate": 2.3e-07,
|
| 2309 |
+
"loss": 5.8239,
|
| 2310 |
+
"step": 1386
|
| 2311 |
+
},
|
| 2312 |
+
{
|
| 2313 |
+
"epoch": 154.0,
|
| 2314 |
+
"eval_accuracy": 0.3088954056695992,
|
| 2315 |
+
"eval_loss": 5.683019638061523,
|
| 2316 |
+
"eval_runtime": 4.1012,
|
| 2317 |
+
"eval_samples_per_second": 0.244,
|
| 2318 |
+
"eval_steps_per_second": 0.244,
|
| 2319 |
+
"step": 1386
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 155.0,
|
| 2323 |
+
"learning_rate": 2.25e-07,
|
| 2324 |
+
"loss": 5.8262,
|
| 2325 |
+
"step": 1395
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 155.0,
|
| 2329 |
+
"eval_accuracy": 0.3088954056695992,
|
| 2330 |
+
"eval_loss": 5.681789875030518,
|
| 2331 |
+
"eval_runtime": 4.3436,
|
| 2332 |
+
"eval_samples_per_second": 0.23,
|
| 2333 |
+
"eval_steps_per_second": 0.23,
|
| 2334 |
+
"step": 1395
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 156.0,
|
| 2338 |
+
"learning_rate": 2.1999999999999998e-07,
|
| 2339 |
+
"loss": 5.8253,
|
| 2340 |
+
"step": 1404
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 156.0,
|
| 2344 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2345 |
+
"eval_loss": 5.680798530578613,
|
| 2346 |
+
"eval_runtime": 3.5916,
|
| 2347 |
+
"eval_samples_per_second": 0.278,
|
| 2348 |
+
"eval_steps_per_second": 0.278,
|
| 2349 |
+
"step": 1404
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 157.0,
|
| 2353 |
+
"learning_rate": 2.1499999999999998e-07,
|
| 2354 |
+
"loss": 5.8169,
|
| 2355 |
+
"step": 1413
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 157.0,
|
| 2359 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2360 |
+
"eval_loss": 5.679258346557617,
|
| 2361 |
+
"eval_runtime": 5.377,
|
| 2362 |
+
"eval_samples_per_second": 0.186,
|
| 2363 |
+
"eval_steps_per_second": 0.186,
|
| 2364 |
+
"step": 1413
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 158.0,
|
| 2368 |
+
"learning_rate": 2.0999999999999997e-07,
|
| 2369 |
+
"loss": 5.8201,
|
| 2370 |
+
"step": 1422
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 158.0,
|
| 2374 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2375 |
+
"eval_loss": 5.679547309875488,
|
| 2376 |
+
"eval_runtime": 3.4728,
|
| 2377 |
+
"eval_samples_per_second": 0.288,
|
| 2378 |
+
"eval_steps_per_second": 0.288,
|
| 2379 |
+
"step": 1422
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 159.0,
|
| 2383 |
+
"learning_rate": 2.0499999999999997e-07,
|
| 2384 |
+
"loss": 5.8077,
|
| 2385 |
+
"step": 1431
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 159.0,
|
| 2389 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2390 |
+
"eval_loss": 5.6798624992370605,
|
| 2391 |
+
"eval_runtime": 4.4591,
|
| 2392 |
+
"eval_samples_per_second": 0.224,
|
| 2393 |
+
"eval_steps_per_second": 0.224,
|
| 2394 |
+
"step": 1431
|
| 2395 |
+
},
|
| 2396 |
+
{
|
| 2397 |
+
"epoch": 160.0,
|
| 2398 |
+
"learning_rate": 2e-07,
|
| 2399 |
+
"loss": 5.8222,
|
| 2400 |
+
"step": 1440
|
| 2401 |
+
},
|
| 2402 |
+
{
|
| 2403 |
+
"epoch": 160.0,
|
| 2404 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2405 |
+
"eval_loss": 5.678928852081299,
|
| 2406 |
+
"eval_runtime": 3.9956,
|
| 2407 |
+
"eval_samples_per_second": 0.25,
|
| 2408 |
+
"eval_steps_per_second": 0.25,
|
| 2409 |
+
"step": 1440
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 161.0,
|
| 2413 |
+
"learning_rate": 1.9499999999999999e-07,
|
| 2414 |
+
"loss": 5.8191,
|
| 2415 |
+
"step": 1449
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 161.0,
|
| 2419 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2420 |
+
"eval_loss": 5.677757263183594,
|
| 2421 |
+
"eval_runtime": 3.5865,
|
| 2422 |
+
"eval_samples_per_second": 0.279,
|
| 2423 |
+
"eval_steps_per_second": 0.279,
|
| 2424 |
+
"step": 1449
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 162.0,
|
| 2428 |
+
"learning_rate": 1.8999999999999998e-07,
|
| 2429 |
+
"loss": 5.83,
|
| 2430 |
+
"step": 1458
|
| 2431 |
+
},
|
| 2432 |
+
{
|
| 2433 |
+
"epoch": 162.0,
|
| 2434 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2435 |
+
"eval_loss": 5.676819801330566,
|
| 2436 |
+
"eval_runtime": 4.7244,
|
| 2437 |
+
"eval_samples_per_second": 0.212,
|
| 2438 |
+
"eval_steps_per_second": 0.212,
|
| 2439 |
+
"step": 1458
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 163.0,
|
| 2443 |
+
"learning_rate": 1.85e-07,
|
| 2444 |
+
"loss": 5.8183,
|
| 2445 |
+
"step": 1467
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 163.0,
|
| 2449 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2450 |
+
"eval_loss": 5.675673484802246,
|
| 2451 |
+
"eval_runtime": 4.423,
|
| 2452 |
+
"eval_samples_per_second": 0.226,
|
| 2453 |
+
"eval_steps_per_second": 0.226,
|
| 2454 |
+
"step": 1467
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 164.0,
|
| 2458 |
+
"learning_rate": 1.8e-07,
|
| 2459 |
+
"loss": 5.8124,
|
| 2460 |
+
"step": 1476
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 164.0,
|
| 2464 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2465 |
+
"eval_loss": 5.674676418304443,
|
| 2466 |
+
"eval_runtime": 4.507,
|
| 2467 |
+
"eval_samples_per_second": 0.222,
|
| 2468 |
+
"eval_steps_per_second": 0.222,
|
| 2469 |
+
"step": 1476
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 165.0,
|
| 2473 |
+
"learning_rate": 1.75e-07,
|
| 2474 |
+
"loss": 5.8119,
|
| 2475 |
+
"step": 1485
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 165.0,
|
| 2479 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2480 |
+
"eval_loss": 5.674499034881592,
|
| 2481 |
+
"eval_runtime": 4.5793,
|
| 2482 |
+
"eval_samples_per_second": 0.218,
|
| 2483 |
+
"eval_steps_per_second": 0.218,
|
| 2484 |
+
"step": 1485
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 166.0,
|
| 2488 |
+
"learning_rate": 1.7000000000000001e-07,
|
| 2489 |
+
"loss": 5.821,
|
| 2490 |
+
"step": 1494
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 166.0,
|
| 2494 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2495 |
+
"eval_loss": 5.674384117126465,
|
| 2496 |
+
"eval_runtime": 4.1895,
|
| 2497 |
+
"eval_samples_per_second": 0.239,
|
| 2498 |
+
"eval_steps_per_second": 0.239,
|
| 2499 |
+
"step": 1494
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 167.0,
|
| 2503 |
+
"learning_rate": 1.65e-07,
|
| 2504 |
+
"loss": 5.807,
|
| 2505 |
+
"step": 1503
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 167.0,
|
| 2509 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2510 |
+
"eval_loss": 5.6736602783203125,
|
| 2511 |
+
"eval_runtime": 3.6058,
|
| 2512 |
+
"eval_samples_per_second": 0.277,
|
| 2513 |
+
"eval_steps_per_second": 0.277,
|
| 2514 |
+
"step": 1503
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 168.0,
|
| 2518 |
+
"learning_rate": 1.6e-07,
|
| 2519 |
+
"loss": 5.8177,
|
| 2520 |
+
"step": 1512
|
| 2521 |
+
},
|
| 2522 |
+
{
|
| 2523 |
+
"epoch": 168.0,
|
| 2524 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2525 |
+
"eval_loss": 5.672515392303467,
|
| 2526 |
+
"eval_runtime": 4.5742,
|
| 2527 |
+
"eval_samples_per_second": 0.219,
|
| 2528 |
+
"eval_steps_per_second": 0.219,
|
| 2529 |
+
"step": 1512
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 169.0,
|
| 2533 |
+
"learning_rate": 1.55e-07,
|
| 2534 |
+
"loss": 5.8046,
|
| 2535 |
+
"step": 1521
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 169.0,
|
| 2539 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2540 |
+
"eval_loss": 5.671017169952393,
|
| 2541 |
+
"eval_runtime": 3.599,
|
| 2542 |
+
"eval_samples_per_second": 0.278,
|
| 2543 |
+
"eval_steps_per_second": 0.278,
|
| 2544 |
+
"step": 1521
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 170.0,
|
| 2548 |
+
"learning_rate": 1.5e-07,
|
| 2549 |
+
"loss": 5.8093,
|
| 2550 |
+
"step": 1530
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 170.0,
|
| 2554 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2555 |
+
"eval_loss": 5.670760154724121,
|
| 2556 |
+
"eval_runtime": 3.7495,
|
| 2557 |
+
"eval_samples_per_second": 0.267,
|
| 2558 |
+
"eval_steps_per_second": 0.267,
|
| 2559 |
+
"step": 1530
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 171.0,
|
| 2563 |
+
"learning_rate": 1.45e-07,
|
| 2564 |
+
"loss": 5.8145,
|
| 2565 |
+
"step": 1539
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 171.0,
|
| 2569 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2570 |
+
"eval_loss": 5.6709699630737305,
|
| 2571 |
+
"eval_runtime": 4.4875,
|
| 2572 |
+
"eval_samples_per_second": 0.223,
|
| 2573 |
+
"eval_steps_per_second": 0.223,
|
| 2574 |
+
"step": 1539
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 172.0,
|
| 2578 |
+
"learning_rate": 1.4e-07,
|
| 2579 |
+
"loss": 5.803,
|
| 2580 |
+
"step": 1548
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 172.0,
|
| 2584 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2585 |
+
"eval_loss": 5.67042350769043,
|
| 2586 |
+
"eval_runtime": 4.2512,
|
| 2587 |
+
"eval_samples_per_second": 0.235,
|
| 2588 |
+
"eval_steps_per_second": 0.235,
|
| 2589 |
+
"step": 1548
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 173.0,
|
| 2593 |
+
"learning_rate": 1.35e-07,
|
| 2594 |
+
"loss": 5.8038,
|
| 2595 |
+
"step": 1557
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 173.0,
|
| 2599 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2600 |
+
"eval_loss": 5.6697306632995605,
|
| 2601 |
+
"eval_runtime": 3.7575,
|
| 2602 |
+
"eval_samples_per_second": 0.266,
|
| 2603 |
+
"eval_steps_per_second": 0.266,
|
| 2604 |
+
"step": 1557
|
| 2605 |
+
},
|
| 2606 |
+
{
|
| 2607 |
+
"epoch": 174.0,
|
| 2608 |
+
"learning_rate": 1.3e-07,
|
| 2609 |
+
"loss": 5.807,
|
| 2610 |
+
"step": 1566
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 174.0,
|
| 2614 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2615 |
+
"eval_loss": 5.668872356414795,
|
| 2616 |
+
"eval_runtime": 3.653,
|
| 2617 |
+
"eval_samples_per_second": 0.274,
|
| 2618 |
+
"eval_steps_per_second": 0.274,
|
| 2619 |
+
"step": 1566
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 175.0,
|
| 2623 |
+
"learning_rate": 1.25e-07,
|
| 2624 |
+
"loss": 5.7974,
|
| 2625 |
+
"step": 1575
|
| 2626 |
+
},
|
| 2627 |
+
{
|
| 2628 |
+
"epoch": 175.0,
|
| 2629 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2630 |
+
"eval_loss": 5.668323993682861,
|
| 2631 |
+
"eval_runtime": 3.3471,
|
| 2632 |
+
"eval_samples_per_second": 0.299,
|
| 2633 |
+
"eval_steps_per_second": 0.299,
|
| 2634 |
+
"step": 1575
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 176.0,
|
| 2638 |
+
"learning_rate": 1.2e-07,
|
| 2639 |
+
"loss": 5.8089,
|
| 2640 |
+
"step": 1584
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 176.0,
|
| 2644 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2645 |
+
"eval_loss": 5.66883659362793,
|
| 2646 |
+
"eval_runtime": 3.7471,
|
| 2647 |
+
"eval_samples_per_second": 0.267,
|
| 2648 |
+
"eval_steps_per_second": 0.267,
|
| 2649 |
+
"step": 1584
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 177.0,
|
| 2653 |
+
"learning_rate": 1.15e-07,
|
| 2654 |
+
"loss": 5.8067,
|
| 2655 |
+
"step": 1593
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 177.0,
|
| 2659 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2660 |
+
"eval_loss": 5.669277667999268,
|
| 2661 |
+
"eval_runtime": 3.325,
|
| 2662 |
+
"eval_samples_per_second": 0.301,
|
| 2663 |
+
"eval_steps_per_second": 0.301,
|
| 2664 |
+
"step": 1593
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 178.0,
|
| 2668 |
+
"learning_rate": 1.0999999999999999e-07,
|
| 2669 |
+
"loss": 5.8092,
|
| 2670 |
+
"step": 1602
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 178.0,
|
| 2674 |
+
"eval_accuracy": 0.31085043988269795,
|
| 2675 |
+
"eval_loss": 5.669507026672363,
|
| 2676 |
+
"eval_runtime": 3.3536,
|
| 2677 |
+
"eval_samples_per_second": 0.298,
|
| 2678 |
+
"eval_steps_per_second": 0.298,
|
| 2679 |
+
"step": 1602
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 179.0,
|
| 2683 |
+
"learning_rate": 1.0499999999999999e-07,
|
| 2684 |
+
"loss": 5.8047,
|
| 2685 |
+
"step": 1611
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 179.0,
|
| 2689 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2690 |
+
"eval_loss": 5.668713092803955,
|
| 2691 |
+
"eval_runtime": 4.0451,
|
| 2692 |
+
"eval_samples_per_second": 0.247,
|
| 2693 |
+
"eval_steps_per_second": 0.247,
|
| 2694 |
+
"step": 1611
|
| 2695 |
+
},
|
| 2696 |
+
{
|
| 2697 |
+
"epoch": 180.0,
|
| 2698 |
+
"learning_rate": 1e-07,
|
| 2699 |
+
"loss": 5.8007,
|
| 2700 |
+
"step": 1620
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 180.0,
|
| 2704 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2705 |
+
"eval_loss": 5.667922019958496,
|
| 2706 |
+
"eval_runtime": 4.4531,
|
| 2707 |
+
"eval_samples_per_second": 0.225,
|
| 2708 |
+
"eval_steps_per_second": 0.225,
|
| 2709 |
+
"step": 1620
|
| 2710 |
+
},
|
| 2711 |
+
{
|
| 2712 |
+
"epoch": 181.0,
|
| 2713 |
+
"learning_rate": 9.499999999999999e-08,
|
| 2714 |
+
"loss": 5.8041,
|
| 2715 |
+
"step": 1629
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 181.0,
|
| 2719 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2720 |
+
"eval_loss": 5.667240619659424,
|
| 2721 |
+
"eval_runtime": 4.9291,
|
| 2722 |
+
"eval_samples_per_second": 0.203,
|
| 2723 |
+
"eval_steps_per_second": 0.203,
|
| 2724 |
+
"step": 1629
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 182.0,
|
| 2728 |
+
"learning_rate": 9e-08,
|
| 2729 |
+
"loss": 5.8072,
|
| 2730 |
+
"step": 1638
|
| 2731 |
+
},
|
| 2732 |
+
{
|
| 2733 |
+
"epoch": 182.0,
|
| 2734 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2735 |
+
"eval_loss": 5.666664123535156,
|
| 2736 |
+
"eval_runtime": 3.1517,
|
| 2737 |
+
"eval_samples_per_second": 0.317,
|
| 2738 |
+
"eval_steps_per_second": 0.317,
|
| 2739 |
+
"step": 1638
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 183.0,
|
| 2743 |
+
"learning_rate": 8.500000000000001e-08,
|
| 2744 |
+
"loss": 5.8093,
|
| 2745 |
+
"step": 1647
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 183.0,
|
| 2749 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2750 |
+
"eval_loss": 5.66619873046875,
|
| 2751 |
+
"eval_runtime": 3.528,
|
| 2752 |
+
"eval_samples_per_second": 0.283,
|
| 2753 |
+
"eval_steps_per_second": 0.283,
|
| 2754 |
+
"step": 1647
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 184.0,
|
| 2758 |
+
"learning_rate": 8e-08,
|
| 2759 |
+
"loss": 5.7948,
|
| 2760 |
+
"step": 1656
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 184.0,
|
| 2764 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2765 |
+
"eval_loss": 5.665794372558594,
|
| 2766 |
+
"eval_runtime": 3.2036,
|
| 2767 |
+
"eval_samples_per_second": 0.312,
|
| 2768 |
+
"eval_steps_per_second": 0.312,
|
| 2769 |
+
"step": 1656
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 185.0,
|
| 2773 |
+
"learning_rate": 7.5e-08,
|
| 2774 |
+
"loss": 5.7968,
|
| 2775 |
+
"step": 1665
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 185.0,
|
| 2779 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2780 |
+
"eval_loss": 5.665581703186035,
|
| 2781 |
+
"eval_runtime": 4.064,
|
| 2782 |
+
"eval_samples_per_second": 0.246,
|
| 2783 |
+
"eval_steps_per_second": 0.246,
|
| 2784 |
+
"step": 1665
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"epoch": 186.0,
|
| 2788 |
+
"learning_rate": 7e-08,
|
| 2789 |
+
"loss": 5.8033,
|
| 2790 |
+
"step": 1674
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 186.0,
|
| 2794 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2795 |
+
"eval_loss": 5.665287494659424,
|
| 2796 |
+
"eval_runtime": 3.7071,
|
| 2797 |
+
"eval_samples_per_second": 0.27,
|
| 2798 |
+
"eval_steps_per_second": 0.27,
|
| 2799 |
+
"step": 1674
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 187.0,
|
| 2803 |
+
"learning_rate": 6.5e-08,
|
| 2804 |
+
"loss": 5.8031,
|
| 2805 |
+
"step": 1683
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 187.0,
|
| 2809 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2810 |
+
"eval_loss": 5.665118217468262,
|
| 2811 |
+
"eval_runtime": 3.8502,
|
| 2812 |
+
"eval_samples_per_second": 0.26,
|
| 2813 |
+
"eval_steps_per_second": 0.26,
|
| 2814 |
+
"step": 1683
|
| 2815 |
+
},
|
| 2816 |
+
{
|
| 2817 |
+
"epoch": 188.0,
|
| 2818 |
+
"learning_rate": 6e-08,
|
| 2819 |
+
"loss": 5.7953,
|
| 2820 |
+
"step": 1692
|
| 2821 |
+
},
|
| 2822 |
+
{
|
| 2823 |
+
"epoch": 188.0,
|
| 2824 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2825 |
+
"eval_loss": 5.664963722229004,
|
| 2826 |
+
"eval_runtime": 3.5099,
|
| 2827 |
+
"eval_samples_per_second": 0.285,
|
| 2828 |
+
"eval_steps_per_second": 0.285,
|
| 2829 |
+
"step": 1692
|
| 2830 |
+
},
|
| 2831 |
+
{
|
| 2832 |
+
"epoch": 189.0,
|
| 2833 |
+
"learning_rate": 5.4999999999999996e-08,
|
| 2834 |
+
"loss": 5.8085,
|
| 2835 |
+
"step": 1701
|
| 2836 |
+
},
|
| 2837 |
+
{
|
| 2838 |
+
"epoch": 189.0,
|
| 2839 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2840 |
+
"eval_loss": 5.6647257804870605,
|
| 2841 |
+
"eval_runtime": 3.1431,
|
| 2842 |
+
"eval_samples_per_second": 0.318,
|
| 2843 |
+
"eval_steps_per_second": 0.318,
|
| 2844 |
+
"step": 1701
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 190.0,
|
| 2848 |
+
"learning_rate": 5e-08,
|
| 2849 |
+
"loss": 5.8021,
|
| 2850 |
+
"step": 1710
|
| 2851 |
+
},
|
| 2852 |
+
{
|
| 2853 |
+
"epoch": 190.0,
|
| 2854 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2855 |
+
"eval_loss": 5.664588928222656,
|
| 2856 |
+
"eval_runtime": 3.9583,
|
| 2857 |
+
"eval_samples_per_second": 0.253,
|
| 2858 |
+
"eval_steps_per_second": 0.253,
|
| 2859 |
+
"step": 1710
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 191.0,
|
| 2863 |
+
"learning_rate": 4.5e-08,
|
| 2864 |
+
"loss": 5.7995,
|
| 2865 |
+
"step": 1719
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 191.0,
|
| 2869 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2870 |
+
"eval_loss": 5.664261817932129,
|
| 2871 |
+
"eval_runtime": 3.9361,
|
| 2872 |
+
"eval_samples_per_second": 0.254,
|
| 2873 |
+
"eval_steps_per_second": 0.254,
|
| 2874 |
+
"step": 1719
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 192.0,
|
| 2878 |
+
"learning_rate": 4e-08,
|
| 2879 |
+
"loss": 5.8057,
|
| 2880 |
+
"step": 1728
|
| 2881 |
+
},
|
| 2882 |
+
{
|
| 2883 |
+
"epoch": 192.0,
|
| 2884 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2885 |
+
"eval_loss": 5.664216041564941,
|
| 2886 |
+
"eval_runtime": 3.9765,
|
| 2887 |
+
"eval_samples_per_second": 0.251,
|
| 2888 |
+
"eval_steps_per_second": 0.251,
|
| 2889 |
+
"step": 1728
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 193.0,
|
| 2893 |
+
"learning_rate": 3.5e-08,
|
| 2894 |
+
"loss": 5.7989,
|
| 2895 |
+
"step": 1737
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 193.0,
|
| 2899 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2900 |
+
"eval_loss": 5.664193153381348,
|
| 2901 |
+
"eval_runtime": 3.0391,
|
| 2902 |
+
"eval_samples_per_second": 0.329,
|
| 2903 |
+
"eval_steps_per_second": 0.329,
|
| 2904 |
+
"step": 1737
|
| 2905 |
+
},
|
| 2906 |
+
{
|
| 2907 |
+
"epoch": 194.0,
|
| 2908 |
+
"learning_rate": 3e-08,
|
| 2909 |
+
"loss": 5.7977,
|
| 2910 |
+
"step": 1746
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 194.0,
|
| 2914 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2915 |
+
"eval_loss": 5.664215564727783,
|
| 2916 |
+
"eval_runtime": 3.3818,
|
| 2917 |
+
"eval_samples_per_second": 0.296,
|
| 2918 |
+
"eval_steps_per_second": 0.296,
|
| 2919 |
+
"step": 1746
|
| 2920 |
+
},
|
| 2921 |
+
{
|
| 2922 |
+
"epoch": 195.0,
|
| 2923 |
+
"learning_rate": 2.5e-08,
|
| 2924 |
+
"loss": 5.8009,
|
| 2925 |
+
"step": 1755
|
| 2926 |
+
},
|
| 2927 |
+
{
|
| 2928 |
+
"epoch": 195.0,
|
| 2929 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2930 |
+
"eval_loss": 5.664393901824951,
|
| 2931 |
+
"eval_runtime": 3.3657,
|
| 2932 |
+
"eval_samples_per_second": 0.297,
|
| 2933 |
+
"eval_steps_per_second": 0.297,
|
| 2934 |
+
"step": 1755
|
| 2935 |
+
},
|
| 2936 |
+
{
|
| 2937 |
+
"epoch": 196.0,
|
| 2938 |
+
"learning_rate": 2e-08,
|
| 2939 |
+
"loss": 5.7988,
|
| 2940 |
+
"step": 1764
|
| 2941 |
+
},
|
| 2942 |
+
{
|
| 2943 |
+
"epoch": 196.0,
|
| 2944 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2945 |
+
"eval_loss": 5.6645026206970215,
|
| 2946 |
+
"eval_runtime": 3.9312,
|
| 2947 |
+
"eval_samples_per_second": 0.254,
|
| 2948 |
+
"eval_steps_per_second": 0.254,
|
| 2949 |
+
"step": 1764
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 197.0,
|
| 2953 |
+
"learning_rate": 1.5e-08,
|
| 2954 |
+
"loss": 5.8016,
|
| 2955 |
+
"step": 1773
|
| 2956 |
+
},
|
| 2957 |
+
{
|
| 2958 |
+
"epoch": 197.0,
|
| 2959 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2960 |
+
"eval_loss": 5.664463043212891,
|
| 2961 |
+
"eval_runtime": 3.7271,
|
| 2962 |
+
"eval_samples_per_second": 0.268,
|
| 2963 |
+
"eval_steps_per_second": 0.268,
|
| 2964 |
+
"step": 1773
|
| 2965 |
+
},
|
| 2966 |
+
{
|
| 2967 |
+
"epoch": 198.0,
|
| 2968 |
+
"learning_rate": 1e-08,
|
| 2969 |
+
"loss": 5.7929,
|
| 2970 |
+
"step": 1782
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 198.0,
|
| 2974 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2975 |
+
"eval_loss": 5.664458751678467,
|
| 2976 |
+
"eval_runtime": 3.5464,
|
| 2977 |
+
"eval_samples_per_second": 0.282,
|
| 2978 |
+
"eval_steps_per_second": 0.282,
|
| 2979 |
+
"step": 1782
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 199.0,
|
| 2983 |
+
"learning_rate": 5e-09,
|
| 2984 |
+
"loss": 5.7973,
|
| 2985 |
+
"step": 1791
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 199.0,
|
| 2989 |
+
"eval_accuracy": 0.3098729227761486,
|
| 2990 |
+
"eval_loss": 5.664454936981201,
|
| 2991 |
+
"eval_runtime": 3.6778,
|
| 2992 |
+
"eval_samples_per_second": 0.272,
|
| 2993 |
+
"eval_steps_per_second": 0.272,
|
| 2994 |
+
"step": 1791
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 200.0,
|
| 2998 |
+
"learning_rate": 0.0,
|
| 2999 |
+
"loss": 5.8022,
|
| 3000 |
+
"step": 1800
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 200.0,
|
| 3004 |
+
"eval_accuracy": 0.3098729227761486,
|
| 3005 |
+
"eval_loss": 5.664452075958252,
|
| 3006 |
+
"eval_runtime": 3.5398,
|
| 3007 |
+
"eval_samples_per_second": 0.283,
|
| 3008 |
+
"eval_steps_per_second": 0.283,
|
| 3009 |
+
"step": 1800
|
| 3010 |
+
},
|
| 3011 |
+
{
|
| 3012 |
+
"epoch": 200.0,
|
| 3013 |
+
"step": 1800,
|
| 3014 |
+
"total_flos": 7834231111680000.0,
|
| 3015 |
+
"train_loss": 6.335995424058702,
|
| 3016 |
+
"train_runtime": 30730.8013,
|
| 3017 |
+
"train_samples_per_second": 0.059,
|
| 3018 |
+
"train_steps_per_second": 0.059
|
| 3019 |
+
}
|
| 3020 |
+
],
|
| 3021 |
+
"logging_steps": 500,
|
| 3022 |
+
"max_steps": 1800,
|
| 3023 |
+
"num_input_tokens_seen": 0,
|
| 3024 |
+
"num_train_epochs": 200,
|
| 3025 |
+
"save_steps": 500,
|
| 3026 |
+
"total_flos": 7834231111680000.0,
|
| 3027 |
+
"train_batch_size": 1,
|
| 3028 |
+
"trial_name": null,
|
| 3029 |
+
"trial_params": null
|
| 3030 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0309161ae26e875243c2ce765be6653533db2b1aac04b580c4fe0fc6ebc70128
|
| 3 |
+
size 4728
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|