End of training
Browse files- README.md +194 -0
- all_results.json +12 -0
- config.json +58 -0
- eval_results.json +8 -0
- model.safetensors +3 -0
- preprocessor_config.json +22 -0
- runs/Dec05_21-10-09_DESKTOP-SKBE9FB/events.out.tfevents.1733454610.DESKTOP-SKBE9FB.21152.0 +3 -0
- runs/Dec05_21-11-59_DESKTOP-SKBE9FB/events.out.tfevents.1733454720.DESKTOP-SKBE9FB.18648.0 +3 -0
- runs/Dec05_21-11-59_DESKTOP-SKBE9FB/events.out.tfevents.1733458229.DESKTOP-SKBE9FB.18648.1 +3 -0
- train_results.json +7 -0
- trainer_state.json +2811 -0
- training_args.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,194 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: microsoft/swinv2-tiny-patch4-window8-256
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
datasets:
|
| 7 |
+
- imagefolder
|
| 8 |
+
metrics:
|
| 9 |
+
- accuracy
|
| 10 |
+
model-index:
|
| 11 |
+
- name: Swin-dmae-DA5-N-Colab
|
| 12 |
+
results:
|
| 13 |
+
- task:
|
| 14 |
+
name: Image Classification
|
| 15 |
+
type: image-classification
|
| 16 |
+
dataset:
|
| 17 |
+
name: imagefolder
|
| 18 |
+
type: imagefolder
|
| 19 |
+
config: default
|
| 20 |
+
split: validation
|
| 21 |
+
args: default
|
| 22 |
+
metrics:
|
| 23 |
+
- name: Accuracy
|
| 24 |
+
type: accuracy
|
| 25 |
+
value: 0.78125
|
| 26 |
+
---
|
| 27 |
+
|
| 28 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 29 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 30 |
+
|
| 31 |
+
# Swin-dmae-DA5-N-Colab
|
| 32 |
+
|
| 33 |
+
This model is a fine-tuned version of [microsoft/swinv2-tiny-patch4-window8-256](https://huggingface.co/microsoft/swinv2-tiny-patch4-window8-256) on the imagefolder dataset.
|
| 34 |
+
It achieves the following results on the evaluation set:
|
| 35 |
+
- Loss: 1.4370
|
| 36 |
+
- Accuracy: 0.7812
|
| 37 |
+
|
| 38 |
+
## Model description
|
| 39 |
+
|
| 40 |
+
More information needed
|
| 41 |
+
|
| 42 |
+
## Intended uses & limitations
|
| 43 |
+
|
| 44 |
+
More information needed
|
| 45 |
+
|
| 46 |
+
## Training and evaluation data
|
| 47 |
+
|
| 48 |
+
More information needed
|
| 49 |
+
|
| 50 |
+
## Training procedure
|
| 51 |
+
|
| 52 |
+
### Training hyperparameters
|
| 53 |
+
|
| 54 |
+
The following hyperparameters were used during training:
|
| 55 |
+
- learning_rate: 4e-05
|
| 56 |
+
- train_batch_size: 16
|
| 57 |
+
- eval_batch_size: 16
|
| 58 |
+
- seed: 42
|
| 59 |
+
- gradient_accumulation_steps: 4
|
| 60 |
+
- total_train_batch_size: 64
|
| 61 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 62 |
+
- lr_scheduler_type: linear
|
| 63 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 64 |
+
- num_epochs: 120
|
| 65 |
+
|
| 66 |
+
### Training results
|
| 67 |
+
|
| 68 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 69 |
+
|:-------------:|:------:|:----:|:---------------:|:--------:|
|
| 70 |
+
| 6.8626 | 0.97 | 24 | 7.8026 | 0.125 |
|
| 71 |
+
| 6.6392 | 1.98 | 49 | 7.5823 | 0.125 |
|
| 72 |
+
| 6.3514 | 2.99 | 74 | 6.6653 | 0.125 |
|
| 73 |
+
| 5.3109 | 4.0 | 99 | 5.2798 | 0.125 |
|
| 74 |
+
| 3.7476 | 4.97 | 123 | 3.6186 | 0.125 |
|
| 75 |
+
| 2.7138 | 5.98 | 148 | 1.9377 | 0.125 |
|
| 76 |
+
| 1.4116 | 6.99 | 173 | 1.4680 | 0.125 |
|
| 77 |
+
| 1.3932 | 8.0 | 198 | 1.3819 | 0.5 |
|
| 78 |
+
| 1.2566 | 8.97 | 222 | 1.5912 | 0.1562 |
|
| 79 |
+
| 1.1332 | 9.98 | 247 | 1.3336 | 0.4375 |
|
| 80 |
+
| 0.9511 | 10.99 | 272 | 1.1492 | 0.2812 |
|
| 81 |
+
| 0.8905 | 12.0 | 297 | 1.0571 | 0.5312 |
|
| 82 |
+
| 0.8317 | 12.97 | 321 | 0.8474 | 0.6562 |
|
| 83 |
+
| 0.6611 | 13.98 | 346 | 0.7520 | 0.7188 |
|
| 84 |
+
| 0.5683 | 14.99 | 371 | 0.6656 | 0.75 |
|
| 85 |
+
| 0.569 | 16.0 | 396 | 0.8109 | 0.5312 |
|
| 86 |
+
| 0.4702 | 16.97 | 420 | 0.7036 | 0.625 |
|
| 87 |
+
| 0.4244 | 17.98 | 445 | 0.8169 | 0.6562 |
|
| 88 |
+
| 0.3483 | 18.99 | 470 | 0.7076 | 0.7188 |
|
| 89 |
+
| 0.3853 | 20.0 | 495 | 0.8644 | 0.7188 |
|
| 90 |
+
| 0.3038 | 20.97 | 519 | 0.8653 | 0.7188 |
|
| 91 |
+
| 0.2885 | 21.98 | 544 | 1.0444 | 0.7188 |
|
| 92 |
+
| 0.2014 | 22.99 | 569 | 1.0684 | 0.5938 |
|
| 93 |
+
| 0.2764 | 24.0 | 594 | 1.1422 | 0.6562 |
|
| 94 |
+
| 0.2493 | 24.97 | 618 | 1.1025 | 0.6875 |
|
| 95 |
+
| 0.2754 | 25.98 | 643 | 1.0371 | 0.7188 |
|
| 96 |
+
| 0.1793 | 26.99 | 668 | 1.1624 | 0.6562 |
|
| 97 |
+
| 0.1971 | 28.0 | 693 | 1.3177 | 0.6875 |
|
| 98 |
+
| 0.1881 | 28.97 | 717 | 1.2813 | 0.6875 |
|
| 99 |
+
| 0.167 | 29.98 | 742 | 1.5564 | 0.625 |
|
| 100 |
+
| 0.1872 | 30.99 | 767 | 1.3762 | 0.7188 |
|
| 101 |
+
| 0.1374 | 32.0 | 792 | 1.4407 | 0.625 |
|
| 102 |
+
| 0.1841 | 32.97 | 816 | 1.4038 | 0.6875 |
|
| 103 |
+
| 0.167 | 33.98 | 841 | 1.3769 | 0.6875 |
|
| 104 |
+
| 0.1614 | 34.99 | 866 | 1.5351 | 0.6562 |
|
| 105 |
+
| 0.1835 | 36.0 | 891 | 1.4466 | 0.6875 |
|
| 106 |
+
| 0.1917 | 36.97 | 915 | 1.3493 | 0.75 |
|
| 107 |
+
| 0.1171 | 37.98 | 940 | 1.4756 | 0.75 |
|
| 108 |
+
| 0.163 | 38.99 | 965 | 1.4373 | 0.6875 |
|
| 109 |
+
| 0.1688 | 40.0 | 990 | 1.4082 | 0.75 |
|
| 110 |
+
| 0.1318 | 40.97 | 1014 | 1.5907 | 0.6875 |
|
| 111 |
+
| 0.1107 | 41.98 | 1039 | 1.7462 | 0.6875 |
|
| 112 |
+
| 0.1064 | 42.99 | 1064 | 1.8704 | 0.5625 |
|
| 113 |
+
| 0.1423 | 44.0 | 1089 | 1.7155 | 0.5625 |
|
| 114 |
+
| 0.082 | 44.97 | 1113 | 1.5552 | 0.7188 |
|
| 115 |
+
| 0.1012 | 45.98 | 1138 | 1.4190 | 0.6875 |
|
| 116 |
+
| 0.1001 | 46.99 | 1163 | 1.6801 | 0.7188 |
|
| 117 |
+
| 0.1037 | 48.0 | 1188 | 1.6864 | 0.7188 |
|
| 118 |
+
| 0.1089 | 48.97 | 1212 | 1.5225 | 0.6875 |
|
| 119 |
+
| 0.0835 | 49.98 | 1237 | 1.9798 | 0.6875 |
|
| 120 |
+
| 0.0818 | 50.99 | 1262 | 1.7268 | 0.6562 |
|
| 121 |
+
| 0.1134 | 52.0 | 1287 | 1.5996 | 0.75 |
|
| 122 |
+
| 0.1115 | 52.97 | 1311 | 1.7281 | 0.6562 |
|
| 123 |
+
| 0.0929 | 53.98 | 1336 | 1.6346 | 0.75 |
|
| 124 |
+
| 0.0909 | 54.99 | 1361 | 1.4370 | 0.7812 |
|
| 125 |
+
| 0.1076 | 56.0 | 1386 | 1.5510 | 0.7812 |
|
| 126 |
+
| 0.0948 | 56.97 | 1410 | 1.6383 | 0.75 |
|
| 127 |
+
| 0.0914 | 57.98 | 1435 | 1.6938 | 0.6875 |
|
| 128 |
+
| 0.0598 | 58.99 | 1460 | 1.6291 | 0.75 |
|
| 129 |
+
| 0.0769 | 60.0 | 1485 | 1.6594 | 0.75 |
|
| 130 |
+
| 0.0894 | 60.97 | 1509 | 1.6302 | 0.7812 |
|
| 131 |
+
| 0.0999 | 61.98 | 1534 | 1.6562 | 0.7188 |
|
| 132 |
+
| 0.0759 | 62.99 | 1559 | 1.5989 | 0.75 |
|
| 133 |
+
| 0.102 | 64.0 | 1584 | 1.6602 | 0.7812 |
|
| 134 |
+
| 0.0864 | 64.97 | 1608 | 1.7386 | 0.7812 |
|
| 135 |
+
| 0.0722 | 65.98 | 1633 | 2.0495 | 0.7188 |
|
| 136 |
+
| 0.0956 | 66.99 | 1658 | 1.9749 | 0.6875 |
|
| 137 |
+
| 0.0698 | 68.0 | 1683 | 2.0090 | 0.6562 |
|
| 138 |
+
| 0.0635 | 68.97 | 1707 | 2.1600 | 0.625 |
|
| 139 |
+
| 0.0726 | 69.98 | 1732 | 1.8477 | 0.75 |
|
| 140 |
+
| 0.0905 | 70.99 | 1757 | 1.9970 | 0.7188 |
|
| 141 |
+
| 0.0955 | 72.0 | 1782 | 1.9001 | 0.75 |
|
| 142 |
+
| 0.0614 | 72.97 | 1806 | 1.9347 | 0.6562 |
|
| 143 |
+
| 0.0721 | 73.98 | 1831 | 1.9007 | 0.6875 |
|
| 144 |
+
| 0.0868 | 74.99 | 1856 | 2.0204 | 0.6562 |
|
| 145 |
+
| 0.0817 | 76.0 | 1881 | 1.9807 | 0.7188 |
|
| 146 |
+
| 0.0533 | 76.97 | 1905 | 1.9782 | 0.75 |
|
| 147 |
+
| 0.0682 | 77.98 | 1930 | 1.8320 | 0.75 |
|
| 148 |
+
| 0.078 | 78.99 | 1955 | 1.8351 | 0.7188 |
|
| 149 |
+
| 0.0991 | 80.0 | 1980 | 1.9694 | 0.7188 |
|
| 150 |
+
| 0.0601 | 80.97 | 2004 | 1.8795 | 0.7188 |
|
| 151 |
+
| 0.072 | 81.98 | 2029 | 2.0294 | 0.6562 |
|
| 152 |
+
| 0.0746 | 82.99 | 2054 | 1.8439 | 0.7188 |
|
| 153 |
+
| 0.0547 | 84.0 | 2079 | 1.9321 | 0.7188 |
|
| 154 |
+
| 0.0497 | 84.97 | 2103 | 1.8862 | 0.7812 |
|
| 155 |
+
| 0.0566 | 85.98 | 2128 | 2.0067 | 0.6562 |
|
| 156 |
+
| 0.0353 | 86.99 | 2153 | 2.0957 | 0.7188 |
|
| 157 |
+
| 0.0634 | 88.0 | 2178 | 2.1571 | 0.6562 |
|
| 158 |
+
| 0.0477 | 88.97 | 2202 | 2.0384 | 0.6875 |
|
| 159 |
+
| 0.0513 | 89.98 | 2227 | 1.9146 | 0.75 |
|
| 160 |
+
| 0.0717 | 90.99 | 2252 | 1.8838 | 0.7188 |
|
| 161 |
+
| 0.0644 | 92.0 | 2277 | 1.9186 | 0.6875 |
|
| 162 |
+
| 0.0848 | 92.97 | 2301 | 1.8828 | 0.7188 |
|
| 163 |
+
| 0.0393 | 93.98 | 2326 | 1.9442 | 0.7188 |
|
| 164 |
+
| 0.046 | 94.99 | 2351 | 1.8866 | 0.7188 |
|
| 165 |
+
| 0.0487 | 96.0 | 2376 | 1.9787 | 0.6875 |
|
| 166 |
+
| 0.074 | 96.97 | 2400 | 2.0081 | 0.6875 |
|
| 167 |
+
| 0.0435 | 97.98 | 2425 | 1.8839 | 0.75 |
|
| 168 |
+
| 0.0509 | 98.99 | 2450 | 1.9208 | 0.7188 |
|
| 169 |
+
| 0.0571 | 100.0 | 2475 | 1.9770 | 0.7188 |
|
| 170 |
+
| 0.0327 | 100.97 | 2499 | 1.9700 | 0.7188 |
|
| 171 |
+
| 0.0387 | 101.98 | 2524 | 1.9251 | 0.75 |
|
| 172 |
+
| 0.029 | 102.99 | 2549 | 1.9490 | 0.7188 |
|
| 173 |
+
| 0.0478 | 104.0 | 2574 | 1.9358 | 0.7188 |
|
| 174 |
+
| 0.0587 | 104.97 | 2598 | 1.9197 | 0.75 |
|
| 175 |
+
| 0.0523 | 105.98 | 2623 | 1.9309 | 0.7188 |
|
| 176 |
+
| 0.0581 | 106.99 | 2648 | 1.9829 | 0.7188 |
|
| 177 |
+
| 0.0352 | 108.0 | 2673 | 2.0047 | 0.6875 |
|
| 178 |
+
| 0.0373 | 108.97 | 2697 | 1.9897 | 0.7188 |
|
| 179 |
+
| 0.0258 | 109.98 | 2722 | 1.9384 | 0.7188 |
|
| 180 |
+
| 0.039 | 110.99 | 2747 | 1.9356 | 0.6875 |
|
| 181 |
+
| 0.0333 | 112.0 | 2772 | 1.9805 | 0.7188 |
|
| 182 |
+
| 0.0641 | 112.97 | 2796 | 1.9814 | 0.6875 |
|
| 183 |
+
| 0.0649 | 113.98 | 2821 | 1.9726 | 0.6875 |
|
| 184 |
+
| 0.0241 | 114.99 | 2846 | 1.9737 | 0.6875 |
|
| 185 |
+
| 0.0356 | 116.0 | 2871 | 1.9856 | 0.6875 |
|
| 186 |
+
| 0.0601 | 116.36 | 2880 | 1.9853 | 0.6875 |
|
| 187 |
+
|
| 188 |
+
|
| 189 |
+
### Framework versions
|
| 190 |
+
|
| 191 |
+
- Transformers 4.36.2
|
| 192 |
+
- Pytorch 2.1.2+cu118
|
| 193 |
+
- Datasets 2.16.1
|
| 194 |
+
- Tokenizers 0.15.0
|
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 116.36,
|
| 3 |
+
"eval_accuracy": 0.78125,
|
| 4 |
+
"eval_loss": 1.4369856119155884,
|
| 5 |
+
"eval_runtime": 0.4361,
|
| 6 |
+
"eval_samples_per_second": 73.378,
|
| 7 |
+
"eval_steps_per_second": 4.586,
|
| 8 |
+
"train_loss": 0.4532645735475752,
|
| 9 |
+
"train_runtime": 3508.1723,
|
| 10 |
+
"train_samples_per_second": 53.943,
|
| 11 |
+
"train_steps_per_second": 0.821
|
| 12 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/swinv2-tiny-patch4-window8-256",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"Swinv2ForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"depths": [
|
| 8 |
+
2,
|
| 9 |
+
2,
|
| 10 |
+
6,
|
| 11 |
+
2
|
| 12 |
+
],
|
| 13 |
+
"drop_path_rate": 0.1,
|
| 14 |
+
"embed_dim": 96,
|
| 15 |
+
"encoder_stride": 32,
|
| 16 |
+
"hidden_act": "gelu",
|
| 17 |
+
"hidden_dropout_prob": 0.0,
|
| 18 |
+
"hidden_size": 768,
|
| 19 |
+
"id2label": {
|
| 20 |
+
"0": "avanzada",
|
| 21 |
+
"1": "leve",
|
| 22 |
+
"2": "moderada",
|
| 23 |
+
"3": "no dmae"
|
| 24 |
+
},
|
| 25 |
+
"image_size": 256,
|
| 26 |
+
"initializer_range": 0.02,
|
| 27 |
+
"label2id": {
|
| 28 |
+
"avanzada": 0,
|
| 29 |
+
"leve": 1,
|
| 30 |
+
"moderada": 2,
|
| 31 |
+
"no dmae": 3
|
| 32 |
+
},
|
| 33 |
+
"layer_norm_eps": 1e-05,
|
| 34 |
+
"mlp_ratio": 4.0,
|
| 35 |
+
"model_type": "swinv2",
|
| 36 |
+
"num_channels": 3,
|
| 37 |
+
"num_heads": [
|
| 38 |
+
3,
|
| 39 |
+
6,
|
| 40 |
+
12,
|
| 41 |
+
24
|
| 42 |
+
],
|
| 43 |
+
"num_layers": 4,
|
| 44 |
+
"patch_size": 4,
|
| 45 |
+
"path_norm": true,
|
| 46 |
+
"pretrained_window_sizes": [
|
| 47 |
+
0,
|
| 48 |
+
0,
|
| 49 |
+
0,
|
| 50 |
+
0
|
| 51 |
+
],
|
| 52 |
+
"problem_type": "single_label_classification",
|
| 53 |
+
"qkv_bias": true,
|
| 54 |
+
"torch_dtype": "float32",
|
| 55 |
+
"transformers_version": "4.36.2",
|
| 56 |
+
"use_absolute_embeddings": false,
|
| 57 |
+
"window_size": 8
|
| 58 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 116.36,
|
| 3 |
+
"eval_accuracy": 0.78125,
|
| 4 |
+
"eval_loss": 1.4369856119155884,
|
| 5 |
+
"eval_runtime": 0.4361,
|
| 6 |
+
"eval_samples_per_second": 73.378,
|
| 7 |
+
"eval_steps_per_second": 4.586
|
| 8 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1ca25c34cdc4770ff4daf5db490c2ccd62bc40e6737cf15b15cd703da0d89682
|
| 3 |
+
size 110356296
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"do_resize": true,
|
| 5 |
+
"image_mean": [
|
| 6 |
+
0.485,
|
| 7 |
+
0.456,
|
| 8 |
+
0.406
|
| 9 |
+
],
|
| 10 |
+
"image_processor_type": "ViTImageProcessor",
|
| 11 |
+
"image_std": [
|
| 12 |
+
0.229,
|
| 13 |
+
0.224,
|
| 14 |
+
0.225
|
| 15 |
+
],
|
| 16 |
+
"resample": 3,
|
| 17 |
+
"rescale_factor": 0.00392156862745098,
|
| 18 |
+
"size": {
|
| 19 |
+
"height": 256,
|
| 20 |
+
"width": 256
|
| 21 |
+
}
|
| 22 |
+
}
|
runs/Dec05_21-10-09_DESKTOP-SKBE9FB/events.out.tfevents.1733454610.DESKTOP-SKBE9FB.21152.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4607809f0731b0ccb7e1b46614f827f0d15f37cb9cfee6ea50c9c26ea4deaee9
|
| 3 |
+
size 6309
|
runs/Dec05_21-11-59_DESKTOP-SKBE9FB/events.out.tfevents.1733454720.DESKTOP-SKBE9FB.18648.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:03c5e9367ba7a5320a012c42bb3128635d2e9cbc00c8582b85ec4eca71ed0209
|
| 3 |
+
size 88046
|
runs/Dec05_21-11-59_DESKTOP-SKBE9FB/events.out.tfevents.1733458229.DESKTOP-SKBE9FB.18648.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a170f0169c624b943bc76f2a386442a78925ca2657745af77f3be935b145cdfe
|
| 3 |
+
size 411
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 116.36,
|
| 3 |
+
"train_loss": 0.4532645735475752,
|
| 4 |
+
"train_runtime": 3508.1723,
|
| 5 |
+
"train_samples_per_second": 53.943,
|
| 6 |
+
"train_steps_per_second": 0.821
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2811 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.78125,
|
| 3 |
+
"best_model_checkpoint": "Swin-dmae-DA5-N-Colab\\checkpoint-1361",
|
| 4 |
+
"epoch": 116.36363636363636,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2880,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.4,
|
| 13 |
+
"learning_rate": 1.3888888888888892e-06,
|
| 14 |
+
"loss": 6.7613,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.81,
|
| 19 |
+
"learning_rate": 2.7777777777777783e-06,
|
| 20 |
+
"loss": 6.8626,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.97,
|
| 25 |
+
"eval_accuracy": 0.125,
|
| 26 |
+
"eval_loss": 7.802581787109375,
|
| 27 |
+
"eval_runtime": 0.4701,
|
| 28 |
+
"eval_samples_per_second": 68.069,
|
| 29 |
+
"eval_steps_per_second": 4.254,
|
| 30 |
+
"step": 24
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 1.21,
|
| 34 |
+
"learning_rate": 4.166666666666667e-06,
|
| 35 |
+
"loss": 6.8425,
|
| 36 |
+
"step": 30
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 1.62,
|
| 40 |
+
"learning_rate": 5.555555555555557e-06,
|
| 41 |
+
"loss": 6.6392,
|
| 42 |
+
"step": 40
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.98,
|
| 46 |
+
"eval_accuracy": 0.125,
|
| 47 |
+
"eval_loss": 7.5823211669921875,
|
| 48 |
+
"eval_runtime": 0.4001,
|
| 49 |
+
"eval_samples_per_second": 79.982,
|
| 50 |
+
"eval_steps_per_second": 4.999,
|
| 51 |
+
"step": 49
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 2.02,
|
| 55 |
+
"learning_rate": 6.944444444444445e-06,
|
| 56 |
+
"loss": 6.5894,
|
| 57 |
+
"step": 50
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 2.42,
|
| 61 |
+
"learning_rate": 8.333333333333334e-06,
|
| 62 |
+
"loss": 6.3711,
|
| 63 |
+
"step": 60
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.83,
|
| 67 |
+
"learning_rate": 9.722222222222223e-06,
|
| 68 |
+
"loss": 6.3514,
|
| 69 |
+
"step": 70
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 2.99,
|
| 73 |
+
"eval_accuracy": 0.125,
|
| 74 |
+
"eval_loss": 6.665298938751221,
|
| 75 |
+
"eval_runtime": 0.4046,
|
| 76 |
+
"eval_samples_per_second": 79.091,
|
| 77 |
+
"eval_steps_per_second": 4.943,
|
| 78 |
+
"step": 74
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 3.23,
|
| 82 |
+
"learning_rate": 1.1111111111111113e-05,
|
| 83 |
+
"loss": 5.8289,
|
| 84 |
+
"step": 80
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 3.64,
|
| 88 |
+
"learning_rate": 1.25e-05,
|
| 89 |
+
"loss": 5.3109,
|
| 90 |
+
"step": 90
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 4.0,
|
| 94 |
+
"eval_accuracy": 0.125,
|
| 95 |
+
"eval_loss": 5.279791355133057,
|
| 96 |
+
"eval_runtime": 0.4241,
|
| 97 |
+
"eval_samples_per_second": 75.455,
|
| 98 |
+
"eval_steps_per_second": 4.716,
|
| 99 |
+
"step": 99
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 4.04,
|
| 103 |
+
"learning_rate": 1.388888888888889e-05,
|
| 104 |
+
"loss": 4.9844,
|
| 105 |
+
"step": 100
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 4.44,
|
| 109 |
+
"learning_rate": 1.5277777777777777e-05,
|
| 110 |
+
"loss": 4.5015,
|
| 111 |
+
"step": 110
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 4.85,
|
| 115 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 116 |
+
"loss": 3.7476,
|
| 117 |
+
"step": 120
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 4.97,
|
| 121 |
+
"eval_accuracy": 0.125,
|
| 122 |
+
"eval_loss": 3.618558168411255,
|
| 123 |
+
"eval_runtime": 0.4101,
|
| 124 |
+
"eval_samples_per_second": 78.031,
|
| 125 |
+
"eval_steps_per_second": 4.877,
|
| 126 |
+
"step": 123
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 5.25,
|
| 130 |
+
"learning_rate": 1.8055555555555558e-05,
|
| 131 |
+
"loss": 3.1884,
|
| 132 |
+
"step": 130
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 5.66,
|
| 136 |
+
"learning_rate": 1.9444444444444445e-05,
|
| 137 |
+
"loss": 2.7138,
|
| 138 |
+
"step": 140
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 5.98,
|
| 142 |
+
"eval_accuracy": 0.125,
|
| 143 |
+
"eval_loss": 1.937683343887329,
|
| 144 |
+
"eval_runtime": 0.4036,
|
| 145 |
+
"eval_samples_per_second": 79.286,
|
| 146 |
+
"eval_steps_per_second": 4.955,
|
| 147 |
+
"step": 148
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 6.06,
|
| 151 |
+
"learning_rate": 2.0833333333333336e-05,
|
| 152 |
+
"loss": 2.1007,
|
| 153 |
+
"step": 150
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 6.46,
|
| 157 |
+
"learning_rate": 2.2222222222222227e-05,
|
| 158 |
+
"loss": 1.7134,
|
| 159 |
+
"step": 160
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 6.87,
|
| 163 |
+
"learning_rate": 2.3611111111111114e-05,
|
| 164 |
+
"loss": 1.4116,
|
| 165 |
+
"step": 170
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 6.99,
|
| 169 |
+
"eval_accuracy": 0.125,
|
| 170 |
+
"eval_loss": 1.4680219888687134,
|
| 171 |
+
"eval_runtime": 0.3951,
|
| 172 |
+
"eval_samples_per_second": 80.993,
|
| 173 |
+
"eval_steps_per_second": 5.062,
|
| 174 |
+
"step": 173
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 7.27,
|
| 178 |
+
"learning_rate": 2.5e-05,
|
| 179 |
+
"loss": 1.4216,
|
| 180 |
+
"step": 180
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 7.68,
|
| 184 |
+
"learning_rate": 2.6388888888888892e-05,
|
| 185 |
+
"loss": 1.3932,
|
| 186 |
+
"step": 190
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 8.0,
|
| 190 |
+
"eval_accuracy": 0.5,
|
| 191 |
+
"eval_loss": 1.3818670511245728,
|
| 192 |
+
"eval_runtime": 0.4901,
|
| 193 |
+
"eval_samples_per_second": 65.291,
|
| 194 |
+
"eval_steps_per_second": 4.081,
|
| 195 |
+
"step": 198
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 8.08,
|
| 199 |
+
"learning_rate": 2.777777777777778e-05,
|
| 200 |
+
"loss": 1.3888,
|
| 201 |
+
"step": 200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 8.48,
|
| 205 |
+
"learning_rate": 2.9166666666666666e-05,
|
| 206 |
+
"loss": 1.3077,
|
| 207 |
+
"step": 210
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 8.89,
|
| 211 |
+
"learning_rate": 3.0555555555555554e-05,
|
| 212 |
+
"loss": 1.2566,
|
| 213 |
+
"step": 220
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 8.97,
|
| 217 |
+
"eval_accuracy": 0.15625,
|
| 218 |
+
"eval_loss": 1.591198205947876,
|
| 219 |
+
"eval_runtime": 0.4171,
|
| 220 |
+
"eval_samples_per_second": 76.721,
|
| 221 |
+
"eval_steps_per_second": 4.795,
|
| 222 |
+
"step": 222
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 9.29,
|
| 226 |
+
"learning_rate": 3.194444444444445e-05,
|
| 227 |
+
"loss": 1.1679,
|
| 228 |
+
"step": 230
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 9.7,
|
| 232 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 233 |
+
"loss": 1.1332,
|
| 234 |
+
"step": 240
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 9.98,
|
| 238 |
+
"eval_accuracy": 0.4375,
|
| 239 |
+
"eval_loss": 1.3336095809936523,
|
| 240 |
+
"eval_runtime": 0.4176,
|
| 241 |
+
"eval_samples_per_second": 76.628,
|
| 242 |
+
"eval_steps_per_second": 4.789,
|
| 243 |
+
"step": 247
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 10.1,
|
| 247 |
+
"learning_rate": 3.472222222222223e-05,
|
| 248 |
+
"loss": 1.0699,
|
| 249 |
+
"step": 250
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 10.51,
|
| 253 |
+
"learning_rate": 3.6111111111111116e-05,
|
| 254 |
+
"loss": 0.9928,
|
| 255 |
+
"step": 260
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 10.91,
|
| 259 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 260 |
+
"loss": 0.9511,
|
| 261 |
+
"step": 270
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 10.99,
|
| 265 |
+
"eval_accuracy": 0.28125,
|
| 266 |
+
"eval_loss": 1.1492191553115845,
|
| 267 |
+
"eval_runtime": 0.4006,
|
| 268 |
+
"eval_samples_per_second": 79.881,
|
| 269 |
+
"eval_steps_per_second": 4.993,
|
| 270 |
+
"step": 272
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 11.31,
|
| 274 |
+
"learning_rate": 3.888888888888889e-05,
|
| 275 |
+
"loss": 0.9286,
|
| 276 |
+
"step": 280
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 11.72,
|
| 280 |
+
"learning_rate": 3.996913580246914e-05,
|
| 281 |
+
"loss": 0.8905,
|
| 282 |
+
"step": 290
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 12.0,
|
| 286 |
+
"eval_accuracy": 0.53125,
|
| 287 |
+
"eval_loss": 1.0571255683898926,
|
| 288 |
+
"eval_runtime": 0.4566,
|
| 289 |
+
"eval_samples_per_second": 70.082,
|
| 290 |
+
"eval_steps_per_second": 4.38,
|
| 291 |
+
"step": 297
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 12.12,
|
| 295 |
+
"learning_rate": 3.981481481481482e-05,
|
| 296 |
+
"loss": 0.8031,
|
| 297 |
+
"step": 300
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 12.53,
|
| 301 |
+
"learning_rate": 3.9660493827160496e-05,
|
| 302 |
+
"loss": 0.8253,
|
| 303 |
+
"step": 310
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 12.93,
|
| 307 |
+
"learning_rate": 3.950617283950617e-05,
|
| 308 |
+
"loss": 0.8317,
|
| 309 |
+
"step": 320
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 12.97,
|
| 313 |
+
"eval_accuracy": 0.65625,
|
| 314 |
+
"eval_loss": 0.8473750948905945,
|
| 315 |
+
"eval_runtime": 0.4201,
|
| 316 |
+
"eval_samples_per_second": 76.173,
|
| 317 |
+
"eval_steps_per_second": 4.761,
|
| 318 |
+
"step": 321
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 13.33,
|
| 322 |
+
"learning_rate": 3.935185185185186e-05,
|
| 323 |
+
"loss": 0.7445,
|
| 324 |
+
"step": 330
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 13.74,
|
| 328 |
+
"learning_rate": 3.9197530864197535e-05,
|
| 329 |
+
"loss": 0.6611,
|
| 330 |
+
"step": 340
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 13.98,
|
| 334 |
+
"eval_accuracy": 0.71875,
|
| 335 |
+
"eval_loss": 0.7519997954368591,
|
| 336 |
+
"eval_runtime": 0.4141,
|
| 337 |
+
"eval_samples_per_second": 77.277,
|
| 338 |
+
"eval_steps_per_second": 4.83,
|
| 339 |
+
"step": 346
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 14.14,
|
| 343 |
+
"learning_rate": 3.904320987654321e-05,
|
| 344 |
+
"loss": 0.6514,
|
| 345 |
+
"step": 350
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 14.55,
|
| 349 |
+
"learning_rate": 3.888888888888889e-05,
|
| 350 |
+
"loss": 0.6022,
|
| 351 |
+
"step": 360
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 14.95,
|
| 355 |
+
"learning_rate": 3.8734567901234575e-05,
|
| 356 |
+
"loss": 0.5683,
|
| 357 |
+
"step": 370
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 14.99,
|
| 361 |
+
"eval_accuracy": 0.75,
|
| 362 |
+
"eval_loss": 0.6655727624893188,
|
| 363 |
+
"eval_runtime": 0.4031,
|
| 364 |
+
"eval_samples_per_second": 79.386,
|
| 365 |
+
"eval_steps_per_second": 4.962,
|
| 366 |
+
"step": 371
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 15.35,
|
| 370 |
+
"learning_rate": 3.8580246913580246e-05,
|
| 371 |
+
"loss": 0.5743,
|
| 372 |
+
"step": 380
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 15.76,
|
| 376 |
+
"learning_rate": 3.842592592592593e-05,
|
| 377 |
+
"loss": 0.569,
|
| 378 |
+
"step": 390
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 16.0,
|
| 382 |
+
"eval_accuracy": 0.53125,
|
| 383 |
+
"eval_loss": 0.8109104037284851,
|
| 384 |
+
"eval_runtime": 0.4166,
|
| 385 |
+
"eval_samples_per_second": 76.812,
|
| 386 |
+
"eval_steps_per_second": 4.801,
|
| 387 |
+
"step": 396
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 16.16,
|
| 391 |
+
"learning_rate": 3.827160493827161e-05,
|
| 392 |
+
"loss": 0.4743,
|
| 393 |
+
"step": 400
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 16.57,
|
| 397 |
+
"learning_rate": 3.8117283950617286e-05,
|
| 398 |
+
"loss": 0.502,
|
| 399 |
+
"step": 410
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 16.97,
|
| 403 |
+
"learning_rate": 3.7962962962962964e-05,
|
| 404 |
+
"loss": 0.4702,
|
| 405 |
+
"step": 420
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 16.97,
|
| 409 |
+
"eval_accuracy": 0.625,
|
| 410 |
+
"eval_loss": 0.703583836555481,
|
| 411 |
+
"eval_runtime": 0.4076,
|
| 412 |
+
"eval_samples_per_second": 78.509,
|
| 413 |
+
"eval_steps_per_second": 4.907,
|
| 414 |
+
"step": 420
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 17.37,
|
| 418 |
+
"learning_rate": 3.780864197530865e-05,
|
| 419 |
+
"loss": 0.4441,
|
| 420 |
+
"step": 430
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 17.78,
|
| 424 |
+
"learning_rate": 3.7654320987654326e-05,
|
| 425 |
+
"loss": 0.4244,
|
| 426 |
+
"step": 440
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 17.98,
|
| 430 |
+
"eval_accuracy": 0.65625,
|
| 431 |
+
"eval_loss": 0.8169162273406982,
|
| 432 |
+
"eval_runtime": 0.4216,
|
| 433 |
+
"eval_samples_per_second": 75.9,
|
| 434 |
+
"eval_steps_per_second": 4.744,
|
| 435 |
+
"step": 445
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 18.18,
|
| 439 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 440 |
+
"loss": 0.4277,
|
| 441 |
+
"step": 450
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 18.59,
|
| 445 |
+
"learning_rate": 3.734567901234568e-05,
|
| 446 |
+
"loss": 0.3584,
|
| 447 |
+
"step": 460
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 18.99,
|
| 451 |
+
"learning_rate": 3.719135802469136e-05,
|
| 452 |
+
"loss": 0.3483,
|
| 453 |
+
"step": 470
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 18.99,
|
| 457 |
+
"eval_accuracy": 0.71875,
|
| 458 |
+
"eval_loss": 0.7075902223587036,
|
| 459 |
+
"eval_runtime": 0.4306,
|
| 460 |
+
"eval_samples_per_second": 74.312,
|
| 461 |
+
"eval_steps_per_second": 4.645,
|
| 462 |
+
"step": 470
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 19.39,
|
| 466 |
+
"learning_rate": 3.703703703703704e-05,
|
| 467 |
+
"loss": 0.3005,
|
| 468 |
+
"step": 480
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 19.8,
|
| 472 |
+
"learning_rate": 3.688271604938272e-05,
|
| 473 |
+
"loss": 0.3853,
|
| 474 |
+
"step": 490
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 20.0,
|
| 478 |
+
"eval_accuracy": 0.71875,
|
| 479 |
+
"eval_loss": 0.8644444942474365,
|
| 480 |
+
"eval_runtime": 0.4716,
|
| 481 |
+
"eval_samples_per_second": 67.852,
|
| 482 |
+
"eval_steps_per_second": 4.241,
|
| 483 |
+
"step": 495
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 20.2,
|
| 487 |
+
"learning_rate": 3.67283950617284e-05,
|
| 488 |
+
"loss": 0.307,
|
| 489 |
+
"step": 500
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 20.61,
|
| 493 |
+
"learning_rate": 3.6574074074074076e-05,
|
| 494 |
+
"loss": 0.3038,
|
| 495 |
+
"step": 510
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 20.97,
|
| 499 |
+
"eval_accuracy": 0.71875,
|
| 500 |
+
"eval_loss": 0.86528080701828,
|
| 501 |
+
"eval_runtime": 0.4181,
|
| 502 |
+
"eval_samples_per_second": 76.538,
|
| 503 |
+
"eval_steps_per_second": 4.784,
|
| 504 |
+
"step": 519
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 21.01,
|
| 508 |
+
"learning_rate": 3.641975308641976e-05,
|
| 509 |
+
"loss": 0.3084,
|
| 510 |
+
"step": 520
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 21.41,
|
| 514 |
+
"learning_rate": 3.626543209876543e-05,
|
| 515 |
+
"loss": 0.2967,
|
| 516 |
+
"step": 530
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 21.82,
|
| 520 |
+
"learning_rate": 3.6111111111111116e-05,
|
| 521 |
+
"loss": 0.2885,
|
| 522 |
+
"step": 540
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 21.98,
|
| 526 |
+
"eval_accuracy": 0.71875,
|
| 527 |
+
"eval_loss": 1.0443750619888306,
|
| 528 |
+
"eval_runtime": 0.4421,
|
| 529 |
+
"eval_samples_per_second": 72.382,
|
| 530 |
+
"eval_steps_per_second": 4.524,
|
| 531 |
+
"step": 544
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 22.22,
|
| 535 |
+
"learning_rate": 3.5956790123456794e-05,
|
| 536 |
+
"loss": 0.2644,
|
| 537 |
+
"step": 550
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 22.63,
|
| 541 |
+
"learning_rate": 3.580246913580247e-05,
|
| 542 |
+
"loss": 0.2014,
|
| 543 |
+
"step": 560
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 22.99,
|
| 547 |
+
"eval_accuracy": 0.59375,
|
| 548 |
+
"eval_loss": 1.06843900680542,
|
| 549 |
+
"eval_runtime": 0.3991,
|
| 550 |
+
"eval_samples_per_second": 80.182,
|
| 551 |
+
"eval_steps_per_second": 5.011,
|
| 552 |
+
"step": 569
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 23.03,
|
| 556 |
+
"learning_rate": 3.564814814814815e-05,
|
| 557 |
+
"loss": 0.309,
|
| 558 |
+
"step": 570
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 23.43,
|
| 562 |
+
"learning_rate": 3.5493827160493834e-05,
|
| 563 |
+
"loss": 0.2515,
|
| 564 |
+
"step": 580
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 23.84,
|
| 568 |
+
"learning_rate": 3.533950617283951e-05,
|
| 569 |
+
"loss": 0.2764,
|
| 570 |
+
"step": 590
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 24.0,
|
| 574 |
+
"eval_accuracy": 0.65625,
|
| 575 |
+
"eval_loss": 1.1421856880187988,
|
| 576 |
+
"eval_runtime": 0.4621,
|
| 577 |
+
"eval_samples_per_second": 69.248,
|
| 578 |
+
"eval_steps_per_second": 4.328,
|
| 579 |
+
"step": 594
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 24.24,
|
| 583 |
+
"learning_rate": 3.518518518518519e-05,
|
| 584 |
+
"loss": 0.2519,
|
| 585 |
+
"step": 600
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 24.65,
|
| 589 |
+
"learning_rate": 3.503086419753087e-05,
|
| 590 |
+
"loss": 0.2493,
|
| 591 |
+
"step": 610
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 24.97,
|
| 595 |
+
"eval_accuracy": 0.6875,
|
| 596 |
+
"eval_loss": 1.102545976638794,
|
| 597 |
+
"eval_runtime": 0.4256,
|
| 598 |
+
"eval_samples_per_second": 75.187,
|
| 599 |
+
"eval_steps_per_second": 4.699,
|
| 600 |
+
"step": 618
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 25.05,
|
| 604 |
+
"learning_rate": 3.4876543209876545e-05,
|
| 605 |
+
"loss": 0.2863,
|
| 606 |
+
"step": 620
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 25.45,
|
| 610 |
+
"learning_rate": 3.472222222222223e-05,
|
| 611 |
+
"loss": 0.2352,
|
| 612 |
+
"step": 630
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 25.86,
|
| 616 |
+
"learning_rate": 3.45679012345679e-05,
|
| 617 |
+
"loss": 0.2754,
|
| 618 |
+
"step": 640
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 25.98,
|
| 622 |
+
"eval_accuracy": 0.71875,
|
| 623 |
+
"eval_loss": 1.0370571613311768,
|
| 624 |
+
"eval_runtime": 0.4276,
|
| 625 |
+
"eval_samples_per_second": 74.835,
|
| 626 |
+
"eval_steps_per_second": 4.677,
|
| 627 |
+
"step": 643
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 26.26,
|
| 631 |
+
"learning_rate": 3.4413580246913584e-05,
|
| 632 |
+
"loss": 0.2549,
|
| 633 |
+
"step": 650
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 26.67,
|
| 637 |
+
"learning_rate": 3.425925925925926e-05,
|
| 638 |
+
"loss": 0.1793,
|
| 639 |
+
"step": 660
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 26.99,
|
| 643 |
+
"eval_accuracy": 0.65625,
|
| 644 |
+
"eval_loss": 1.1623690128326416,
|
| 645 |
+
"eval_runtime": 0.4181,
|
| 646 |
+
"eval_samples_per_second": 76.534,
|
| 647 |
+
"eval_steps_per_second": 4.783,
|
| 648 |
+
"step": 668
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 27.07,
|
| 652 |
+
"learning_rate": 3.410493827160494e-05,
|
| 653 |
+
"loss": 0.2305,
|
| 654 |
+
"step": 670
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 27.47,
|
| 658 |
+
"learning_rate": 3.395061728395062e-05,
|
| 659 |
+
"loss": 0.1836,
|
| 660 |
+
"step": 680
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 27.88,
|
| 664 |
+
"learning_rate": 3.37962962962963e-05,
|
| 665 |
+
"loss": 0.1971,
|
| 666 |
+
"step": 690
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 28.0,
|
| 670 |
+
"eval_accuracy": 0.6875,
|
| 671 |
+
"eval_loss": 1.3177158832550049,
|
| 672 |
+
"eval_runtime": 0.3941,
|
| 673 |
+
"eval_samples_per_second": 81.2,
|
| 674 |
+
"eval_steps_per_second": 5.075,
|
| 675 |
+
"step": 693
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 28.28,
|
| 679 |
+
"learning_rate": 3.364197530864198e-05,
|
| 680 |
+
"loss": 0.1514,
|
| 681 |
+
"step": 700
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 28.69,
|
| 685 |
+
"learning_rate": 3.348765432098766e-05,
|
| 686 |
+
"loss": 0.1881,
|
| 687 |
+
"step": 710
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 28.97,
|
| 691 |
+
"eval_accuracy": 0.6875,
|
| 692 |
+
"eval_loss": 1.2813467979431152,
|
| 693 |
+
"eval_runtime": 0.3961,
|
| 694 |
+
"eval_samples_per_second": 80.79,
|
| 695 |
+
"eval_steps_per_second": 5.049,
|
| 696 |
+
"step": 717
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 29.09,
|
| 700 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 701 |
+
"loss": 0.1854,
|
| 702 |
+
"step": 720
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 29.49,
|
| 706 |
+
"learning_rate": 3.317901234567901e-05,
|
| 707 |
+
"loss": 0.2045,
|
| 708 |
+
"step": 730
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 29.9,
|
| 712 |
+
"learning_rate": 3.30246913580247e-05,
|
| 713 |
+
"loss": 0.167,
|
| 714 |
+
"step": 740
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 29.98,
|
| 718 |
+
"eval_accuracy": 0.625,
|
| 719 |
+
"eval_loss": 1.5564466714859009,
|
| 720 |
+
"eval_runtime": 0.4011,
|
| 721 |
+
"eval_samples_per_second": 79.783,
|
| 722 |
+
"eval_steps_per_second": 4.986,
|
| 723 |
+
"step": 742
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 30.3,
|
| 727 |
+
"learning_rate": 3.2870370370370375e-05,
|
| 728 |
+
"loss": 0.1761,
|
| 729 |
+
"step": 750
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 30.71,
|
| 733 |
+
"learning_rate": 3.271604938271605e-05,
|
| 734 |
+
"loss": 0.1872,
|
| 735 |
+
"step": 760
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 30.99,
|
| 739 |
+
"eval_accuracy": 0.71875,
|
| 740 |
+
"eval_loss": 1.3761742115020752,
|
| 741 |
+
"eval_runtime": 0.4136,
|
| 742 |
+
"eval_samples_per_second": 77.369,
|
| 743 |
+
"eval_steps_per_second": 4.836,
|
| 744 |
+
"step": 767
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 31.11,
|
| 748 |
+
"learning_rate": 3.256172839506173e-05,
|
| 749 |
+
"loss": 0.1697,
|
| 750 |
+
"step": 770
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 31.52,
|
| 754 |
+
"learning_rate": 3.2407407407407415e-05,
|
| 755 |
+
"loss": 0.1806,
|
| 756 |
+
"step": 780
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 31.92,
|
| 760 |
+
"learning_rate": 3.2253086419753086e-05,
|
| 761 |
+
"loss": 0.1374,
|
| 762 |
+
"step": 790
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 32.0,
|
| 766 |
+
"eval_accuracy": 0.625,
|
| 767 |
+
"eval_loss": 1.4406580924987793,
|
| 768 |
+
"eval_runtime": 0.4051,
|
| 769 |
+
"eval_samples_per_second": 78.994,
|
| 770 |
+
"eval_steps_per_second": 4.937,
|
| 771 |
+
"step": 792
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 32.32,
|
| 775 |
+
"learning_rate": 3.209876543209877e-05,
|
| 776 |
+
"loss": 0.1848,
|
| 777 |
+
"step": 800
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 32.73,
|
| 781 |
+
"learning_rate": 3.194444444444445e-05,
|
| 782 |
+
"loss": 0.1841,
|
| 783 |
+
"step": 810
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 32.97,
|
| 787 |
+
"eval_accuracy": 0.6875,
|
| 788 |
+
"eval_loss": 1.4038010835647583,
|
| 789 |
+
"eval_runtime": 0.4176,
|
| 790 |
+
"eval_samples_per_second": 76.627,
|
| 791 |
+
"eval_steps_per_second": 4.789,
|
| 792 |
+
"step": 816
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 33.13,
|
| 796 |
+
"learning_rate": 3.1790123456790125e-05,
|
| 797 |
+
"loss": 0.1365,
|
| 798 |
+
"step": 820
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 33.54,
|
| 802 |
+
"learning_rate": 3.16358024691358e-05,
|
| 803 |
+
"loss": 0.1309,
|
| 804 |
+
"step": 830
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 33.94,
|
| 808 |
+
"learning_rate": 3.148148148148149e-05,
|
| 809 |
+
"loss": 0.167,
|
| 810 |
+
"step": 840
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 33.98,
|
| 814 |
+
"eval_accuracy": 0.6875,
|
| 815 |
+
"eval_loss": 1.3768980503082275,
|
| 816 |
+
"eval_runtime": 0.4141,
|
| 817 |
+
"eval_samples_per_second": 77.277,
|
| 818 |
+
"eval_steps_per_second": 4.83,
|
| 819 |
+
"step": 841
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 34.34,
|
| 823 |
+
"learning_rate": 3.1327160493827165e-05,
|
| 824 |
+
"loss": 0.1689,
|
| 825 |
+
"step": 850
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 34.75,
|
| 829 |
+
"learning_rate": 3.117283950617284e-05,
|
| 830 |
+
"loss": 0.1614,
|
| 831 |
+
"step": 860
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 34.99,
|
| 835 |
+
"eval_accuracy": 0.65625,
|
| 836 |
+
"eval_loss": 1.5350613594055176,
|
| 837 |
+
"eval_runtime": 0.4051,
|
| 838 |
+
"eval_samples_per_second": 78.994,
|
| 839 |
+
"eval_steps_per_second": 4.937,
|
| 840 |
+
"step": 866
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 35.15,
|
| 844 |
+
"learning_rate": 3.101851851851852e-05,
|
| 845 |
+
"loss": 0.1439,
|
| 846 |
+
"step": 870
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 35.56,
|
| 850 |
+
"learning_rate": 3.08641975308642e-05,
|
| 851 |
+
"loss": 0.1706,
|
| 852 |
+
"step": 880
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 35.96,
|
| 856 |
+
"learning_rate": 3.070987654320988e-05,
|
| 857 |
+
"loss": 0.1835,
|
| 858 |
+
"step": 890
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 36.0,
|
| 862 |
+
"eval_accuracy": 0.6875,
|
| 863 |
+
"eval_loss": 1.4466042518615723,
|
| 864 |
+
"eval_runtime": 0.4396,
|
| 865 |
+
"eval_samples_per_second": 72.793,
|
| 866 |
+
"eval_steps_per_second": 4.55,
|
| 867 |
+
"step": 891
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 36.36,
|
| 871 |
+
"learning_rate": 3.0555555555555554e-05,
|
| 872 |
+
"loss": 0.1558,
|
| 873 |
+
"step": 900
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 36.77,
|
| 877 |
+
"learning_rate": 3.0401234567901238e-05,
|
| 878 |
+
"loss": 0.1917,
|
| 879 |
+
"step": 910
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 36.97,
|
| 883 |
+
"eval_accuracy": 0.75,
|
| 884 |
+
"eval_loss": 1.349318265914917,
|
| 885 |
+
"eval_runtime": 0.4031,
|
| 886 |
+
"eval_samples_per_second": 79.387,
|
| 887 |
+
"eval_steps_per_second": 4.962,
|
| 888 |
+
"step": 915
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 37.17,
|
| 892 |
+
"learning_rate": 3.0246913580246916e-05,
|
| 893 |
+
"loss": 0.1761,
|
| 894 |
+
"step": 920
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 37.58,
|
| 898 |
+
"learning_rate": 3.0092592592592593e-05,
|
| 899 |
+
"loss": 0.1861,
|
| 900 |
+
"step": 930
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 37.98,
|
| 904 |
+
"learning_rate": 2.9938271604938275e-05,
|
| 905 |
+
"loss": 0.1171,
|
| 906 |
+
"step": 940
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 37.98,
|
| 910 |
+
"eval_accuracy": 0.75,
|
| 911 |
+
"eval_loss": 1.4756131172180176,
|
| 912 |
+
"eval_runtime": 0.4051,
|
| 913 |
+
"eval_samples_per_second": 78.994,
|
| 914 |
+
"eval_steps_per_second": 4.937,
|
| 915 |
+
"step": 940
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 38.38,
|
| 919 |
+
"learning_rate": 2.9783950617283956e-05,
|
| 920 |
+
"loss": 0.1234,
|
| 921 |
+
"step": 950
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 38.79,
|
| 925 |
+
"learning_rate": 2.962962962962963e-05,
|
| 926 |
+
"loss": 0.163,
|
| 927 |
+
"step": 960
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 38.99,
|
| 931 |
+
"eval_accuracy": 0.6875,
|
| 932 |
+
"eval_loss": 1.4373202323913574,
|
| 933 |
+
"eval_runtime": 0.4221,
|
| 934 |
+
"eval_samples_per_second": 75.809,
|
| 935 |
+
"eval_steps_per_second": 4.738,
|
| 936 |
+
"step": 965
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 39.19,
|
| 940 |
+
"learning_rate": 2.947530864197531e-05,
|
| 941 |
+
"loss": 0.1452,
|
| 942 |
+
"step": 970
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 39.6,
|
| 946 |
+
"learning_rate": 2.9320987654320992e-05,
|
| 947 |
+
"loss": 0.088,
|
| 948 |
+
"step": 980
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 40.0,
|
| 952 |
+
"learning_rate": 2.9166666666666666e-05,
|
| 953 |
+
"loss": 0.1688,
|
| 954 |
+
"step": 990
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 40.0,
|
| 958 |
+
"eval_accuracy": 0.75,
|
| 959 |
+
"eval_loss": 1.4082034826278687,
|
| 960 |
+
"eval_runtime": 0.4786,
|
| 961 |
+
"eval_samples_per_second": 66.859,
|
| 962 |
+
"eval_steps_per_second": 4.179,
|
| 963 |
+
"step": 990
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 40.4,
|
| 967 |
+
"learning_rate": 2.9012345679012347e-05,
|
| 968 |
+
"loss": 0.1359,
|
| 969 |
+
"step": 1000
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 40.81,
|
| 973 |
+
"learning_rate": 2.885802469135803e-05,
|
| 974 |
+
"loss": 0.1318,
|
| 975 |
+
"step": 1010
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 40.97,
|
| 979 |
+
"eval_accuracy": 0.6875,
|
| 980 |
+
"eval_loss": 1.5907220840454102,
|
| 981 |
+
"eval_runtime": 0.4061,
|
| 982 |
+
"eval_samples_per_second": 78.797,
|
| 983 |
+
"eval_steps_per_second": 4.925,
|
| 984 |
+
"step": 1014
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 41.21,
|
| 988 |
+
"learning_rate": 2.8703703703703706e-05,
|
| 989 |
+
"loss": 0.1547,
|
| 990 |
+
"step": 1020
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 41.62,
|
| 994 |
+
"learning_rate": 2.8549382716049384e-05,
|
| 995 |
+
"loss": 0.1107,
|
| 996 |
+
"step": 1030
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 41.98,
|
| 1000 |
+
"eval_accuracy": 0.6875,
|
| 1001 |
+
"eval_loss": 1.7461936473846436,
|
| 1002 |
+
"eval_runtime": 0.4081,
|
| 1003 |
+
"eval_samples_per_second": 78.414,
|
| 1004 |
+
"eval_steps_per_second": 4.901,
|
| 1005 |
+
"step": 1039
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 42.02,
|
| 1009 |
+
"learning_rate": 2.8395061728395065e-05,
|
| 1010 |
+
"loss": 0.1144,
|
| 1011 |
+
"step": 1040
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 42.42,
|
| 1015 |
+
"learning_rate": 2.8240740740740743e-05,
|
| 1016 |
+
"loss": 0.1239,
|
| 1017 |
+
"step": 1050
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 42.83,
|
| 1021 |
+
"learning_rate": 2.8086419753086424e-05,
|
| 1022 |
+
"loss": 0.1064,
|
| 1023 |
+
"step": 1060
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 42.99,
|
| 1027 |
+
"eval_accuracy": 0.5625,
|
| 1028 |
+
"eval_loss": 1.8704116344451904,
|
| 1029 |
+
"eval_runtime": 0.4081,
|
| 1030 |
+
"eval_samples_per_second": 78.414,
|
| 1031 |
+
"eval_steps_per_second": 4.901,
|
| 1032 |
+
"step": 1064
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 43.23,
|
| 1036 |
+
"learning_rate": 2.79320987654321e-05,
|
| 1037 |
+
"loss": 0.0978,
|
| 1038 |
+
"step": 1070
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 43.64,
|
| 1042 |
+
"learning_rate": 2.777777777777778e-05,
|
| 1043 |
+
"loss": 0.1423,
|
| 1044 |
+
"step": 1080
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 44.0,
|
| 1048 |
+
"eval_accuracy": 0.5625,
|
| 1049 |
+
"eval_loss": 1.715477705001831,
|
| 1050 |
+
"eval_runtime": 0.4051,
|
| 1051 |
+
"eval_samples_per_second": 78.994,
|
| 1052 |
+
"eval_steps_per_second": 4.937,
|
| 1053 |
+
"step": 1089
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 44.04,
|
| 1057 |
+
"learning_rate": 2.762345679012346e-05,
|
| 1058 |
+
"loss": 0.1031,
|
| 1059 |
+
"step": 1090
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 44.44,
|
| 1063 |
+
"learning_rate": 2.746913580246914e-05,
|
| 1064 |
+
"loss": 0.1473,
|
| 1065 |
+
"step": 1100
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 44.85,
|
| 1069 |
+
"learning_rate": 2.7314814814814816e-05,
|
| 1070 |
+
"loss": 0.082,
|
| 1071 |
+
"step": 1110
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 44.97,
|
| 1075 |
+
"eval_accuracy": 0.71875,
|
| 1076 |
+
"eval_loss": 1.5552070140838623,
|
| 1077 |
+
"eval_runtime": 0.3991,
|
| 1078 |
+
"eval_samples_per_second": 80.182,
|
| 1079 |
+
"eval_steps_per_second": 5.011,
|
| 1080 |
+
"step": 1113
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 45.25,
|
| 1084 |
+
"learning_rate": 2.7160493827160497e-05,
|
| 1085 |
+
"loss": 0.1206,
|
| 1086 |
+
"step": 1120
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 45.66,
|
| 1090 |
+
"learning_rate": 2.7006172839506174e-05,
|
| 1091 |
+
"loss": 0.1012,
|
| 1092 |
+
"step": 1130
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 45.98,
|
| 1096 |
+
"eval_accuracy": 0.6875,
|
| 1097 |
+
"eval_loss": 1.4189940690994263,
|
| 1098 |
+
"eval_runtime": 0.4136,
|
| 1099 |
+
"eval_samples_per_second": 77.37,
|
| 1100 |
+
"eval_steps_per_second": 4.836,
|
| 1101 |
+
"step": 1138
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 46.06,
|
| 1105 |
+
"learning_rate": 2.6851851851851852e-05,
|
| 1106 |
+
"loss": 0.1124,
|
| 1107 |
+
"step": 1140
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 46.46,
|
| 1111 |
+
"learning_rate": 2.6697530864197533e-05,
|
| 1112 |
+
"loss": 0.114,
|
| 1113 |
+
"step": 1150
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 46.87,
|
| 1117 |
+
"learning_rate": 2.654320987654321e-05,
|
| 1118 |
+
"loss": 0.1001,
|
| 1119 |
+
"step": 1160
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 46.99,
|
| 1123 |
+
"eval_accuracy": 0.71875,
|
| 1124 |
+
"eval_loss": 1.6800808906555176,
|
| 1125 |
+
"eval_runtime": 0.4021,
|
| 1126 |
+
"eval_samples_per_second": 79.584,
|
| 1127 |
+
"eval_steps_per_second": 4.974,
|
| 1128 |
+
"step": 1163
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 47.27,
|
| 1132 |
+
"learning_rate": 2.6388888888888892e-05,
|
| 1133 |
+
"loss": 0.1309,
|
| 1134 |
+
"step": 1170
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 47.68,
|
| 1138 |
+
"learning_rate": 2.623456790123457e-05,
|
| 1139 |
+
"loss": 0.1037,
|
| 1140 |
+
"step": 1180
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 48.0,
|
| 1144 |
+
"eval_accuracy": 0.71875,
|
| 1145 |
+
"eval_loss": 1.686426043510437,
|
| 1146 |
+
"eval_runtime": 0.4541,
|
| 1147 |
+
"eval_samples_per_second": 70.469,
|
| 1148 |
+
"eval_steps_per_second": 4.404,
|
| 1149 |
+
"step": 1188
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 48.08,
|
| 1153 |
+
"learning_rate": 2.6080246913580247e-05,
|
| 1154 |
+
"loss": 0.0909,
|
| 1155 |
+
"step": 1190
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 48.48,
|
| 1159 |
+
"learning_rate": 2.5925925925925928e-05,
|
| 1160 |
+
"loss": 0.1296,
|
| 1161 |
+
"step": 1200
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 48.89,
|
| 1165 |
+
"learning_rate": 2.577160493827161e-05,
|
| 1166 |
+
"loss": 0.1089,
|
| 1167 |
+
"step": 1210
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 48.97,
|
| 1171 |
+
"eval_accuracy": 0.6875,
|
| 1172 |
+
"eval_loss": 1.5225051641464233,
|
| 1173 |
+
"eval_runtime": 0.4051,
|
| 1174 |
+
"eval_samples_per_second": 78.995,
|
| 1175 |
+
"eval_steps_per_second": 4.937,
|
| 1176 |
+
"step": 1212
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 49.29,
|
| 1180 |
+
"learning_rate": 2.5617283950617284e-05,
|
| 1181 |
+
"loss": 0.1184,
|
| 1182 |
+
"step": 1220
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 49.7,
|
| 1186 |
+
"learning_rate": 2.5462962962962965e-05,
|
| 1187 |
+
"loss": 0.0835,
|
| 1188 |
+
"step": 1230
|
| 1189 |
+
},
|
| 1190 |
+
{
|
| 1191 |
+
"epoch": 49.98,
|
| 1192 |
+
"eval_accuracy": 0.6875,
|
| 1193 |
+
"eval_loss": 1.979781150817871,
|
| 1194 |
+
"eval_runtime": 0.4161,
|
| 1195 |
+
"eval_samples_per_second": 76.903,
|
| 1196 |
+
"eval_steps_per_second": 4.806,
|
| 1197 |
+
"step": 1237
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 50.1,
|
| 1201 |
+
"learning_rate": 2.5308641975308646e-05,
|
| 1202 |
+
"loss": 0.1079,
|
| 1203 |
+
"step": 1240
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 50.51,
|
| 1207 |
+
"learning_rate": 2.515432098765432e-05,
|
| 1208 |
+
"loss": 0.1204,
|
| 1209 |
+
"step": 1250
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 50.91,
|
| 1213 |
+
"learning_rate": 2.5e-05,
|
| 1214 |
+
"loss": 0.0818,
|
| 1215 |
+
"step": 1260
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 50.99,
|
| 1219 |
+
"eval_accuracy": 0.65625,
|
| 1220 |
+
"eval_loss": 1.726827621459961,
|
| 1221 |
+
"eval_runtime": 0.4051,
|
| 1222 |
+
"eval_samples_per_second": 78.995,
|
| 1223 |
+
"eval_steps_per_second": 4.937,
|
| 1224 |
+
"step": 1262
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 51.31,
|
| 1228 |
+
"learning_rate": 2.4845679012345682e-05,
|
| 1229 |
+
"loss": 0.0759,
|
| 1230 |
+
"step": 1270
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 51.72,
|
| 1234 |
+
"learning_rate": 2.469135802469136e-05,
|
| 1235 |
+
"loss": 0.1134,
|
| 1236 |
+
"step": 1280
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 52.0,
|
| 1240 |
+
"eval_accuracy": 0.75,
|
| 1241 |
+
"eval_loss": 1.599599838256836,
|
| 1242 |
+
"eval_runtime": 0.4061,
|
| 1243 |
+
"eval_samples_per_second": 78.8,
|
| 1244 |
+
"eval_steps_per_second": 4.925,
|
| 1245 |
+
"step": 1287
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 52.12,
|
| 1249 |
+
"learning_rate": 2.4537037037037038e-05,
|
| 1250 |
+
"loss": 0.0988,
|
| 1251 |
+
"step": 1290
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 52.53,
|
| 1255 |
+
"learning_rate": 2.438271604938272e-05,
|
| 1256 |
+
"loss": 0.0851,
|
| 1257 |
+
"step": 1300
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 52.93,
|
| 1261 |
+
"learning_rate": 2.4228395061728396e-05,
|
| 1262 |
+
"loss": 0.1115,
|
| 1263 |
+
"step": 1310
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 52.97,
|
| 1267 |
+
"eval_accuracy": 0.65625,
|
| 1268 |
+
"eval_loss": 1.7280734777450562,
|
| 1269 |
+
"eval_runtime": 0.4181,
|
| 1270 |
+
"eval_samples_per_second": 76.538,
|
| 1271 |
+
"eval_steps_per_second": 4.784,
|
| 1272 |
+
"step": 1311
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 53.33,
|
| 1276 |
+
"learning_rate": 2.4074074074074077e-05,
|
| 1277 |
+
"loss": 0.0995,
|
| 1278 |
+
"step": 1320
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 53.74,
|
| 1282 |
+
"learning_rate": 2.391975308641976e-05,
|
| 1283 |
+
"loss": 0.0929,
|
| 1284 |
+
"step": 1330
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 53.98,
|
| 1288 |
+
"eval_accuracy": 0.75,
|
| 1289 |
+
"eval_loss": 1.6346489191055298,
|
| 1290 |
+
"eval_runtime": 0.4061,
|
| 1291 |
+
"eval_samples_per_second": 78.8,
|
| 1292 |
+
"eval_steps_per_second": 4.925,
|
| 1293 |
+
"step": 1336
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 54.14,
|
| 1297 |
+
"learning_rate": 2.3765432098765433e-05,
|
| 1298 |
+
"loss": 0.1187,
|
| 1299 |
+
"step": 1340
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 54.55,
|
| 1303 |
+
"learning_rate": 2.3611111111111114e-05,
|
| 1304 |
+
"loss": 0.1477,
|
| 1305 |
+
"step": 1350
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 54.95,
|
| 1309 |
+
"learning_rate": 2.3456790123456795e-05,
|
| 1310 |
+
"loss": 0.0909,
|
| 1311 |
+
"step": 1360
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 54.99,
|
| 1315 |
+
"eval_accuracy": 0.78125,
|
| 1316 |
+
"eval_loss": 1.4369856119155884,
|
| 1317 |
+
"eval_runtime": 0.4041,
|
| 1318 |
+
"eval_samples_per_second": 79.19,
|
| 1319 |
+
"eval_steps_per_second": 4.949,
|
| 1320 |
+
"step": 1361
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 55.35,
|
| 1324 |
+
"learning_rate": 2.330246913580247e-05,
|
| 1325 |
+
"loss": 0.1106,
|
| 1326 |
+
"step": 1370
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 55.76,
|
| 1330 |
+
"learning_rate": 2.314814814814815e-05,
|
| 1331 |
+
"loss": 0.1076,
|
| 1332 |
+
"step": 1380
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 56.0,
|
| 1336 |
+
"eval_accuracy": 0.78125,
|
| 1337 |
+
"eval_loss": 1.550971508026123,
|
| 1338 |
+
"eval_runtime": 0.4091,
|
| 1339 |
+
"eval_samples_per_second": 78.222,
|
| 1340 |
+
"eval_steps_per_second": 4.889,
|
| 1341 |
+
"step": 1386
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 56.16,
|
| 1345 |
+
"learning_rate": 2.299382716049383e-05,
|
| 1346 |
+
"loss": 0.1301,
|
| 1347 |
+
"step": 1390
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 56.57,
|
| 1351 |
+
"learning_rate": 2.2839506172839506e-05,
|
| 1352 |
+
"loss": 0.0807,
|
| 1353 |
+
"step": 1400
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 56.97,
|
| 1357 |
+
"learning_rate": 2.2685185185185187e-05,
|
| 1358 |
+
"loss": 0.0948,
|
| 1359 |
+
"step": 1410
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 56.97,
|
| 1363 |
+
"eval_accuracy": 0.75,
|
| 1364 |
+
"eval_loss": 1.6382544040679932,
|
| 1365 |
+
"eval_runtime": 0.4751,
|
| 1366 |
+
"eval_samples_per_second": 67.353,
|
| 1367 |
+
"eval_steps_per_second": 4.21,
|
| 1368 |
+
"step": 1410
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 57.37,
|
| 1372 |
+
"learning_rate": 2.2530864197530865e-05,
|
| 1373 |
+
"loss": 0.0961,
|
| 1374 |
+
"step": 1420
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 57.78,
|
| 1378 |
+
"learning_rate": 2.2376543209876546e-05,
|
| 1379 |
+
"loss": 0.0914,
|
| 1380 |
+
"step": 1430
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 57.98,
|
| 1384 |
+
"eval_accuracy": 0.6875,
|
| 1385 |
+
"eval_loss": 1.6937541961669922,
|
| 1386 |
+
"eval_runtime": 0.3996,
|
| 1387 |
+
"eval_samples_per_second": 80.08,
|
| 1388 |
+
"eval_steps_per_second": 5.005,
|
| 1389 |
+
"step": 1435
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 58.18,
|
| 1393 |
+
"learning_rate": 2.2222222222222227e-05,
|
| 1394 |
+
"loss": 0.0716,
|
| 1395 |
+
"step": 1440
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 58.59,
|
| 1399 |
+
"learning_rate": 2.20679012345679e-05,
|
| 1400 |
+
"loss": 0.0911,
|
| 1401 |
+
"step": 1450
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 58.99,
|
| 1405 |
+
"learning_rate": 2.1913580246913582e-05,
|
| 1406 |
+
"loss": 0.0598,
|
| 1407 |
+
"step": 1460
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 58.99,
|
| 1411 |
+
"eval_accuracy": 0.75,
|
| 1412 |
+
"eval_loss": 1.6290702819824219,
|
| 1413 |
+
"eval_runtime": 0.3996,
|
| 1414 |
+
"eval_samples_per_second": 80.081,
|
| 1415 |
+
"eval_steps_per_second": 5.005,
|
| 1416 |
+
"step": 1460
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 59.39,
|
| 1420 |
+
"learning_rate": 2.1759259259259263e-05,
|
| 1421 |
+
"loss": 0.087,
|
| 1422 |
+
"step": 1470
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 59.8,
|
| 1426 |
+
"learning_rate": 2.1604938271604937e-05,
|
| 1427 |
+
"loss": 0.0769,
|
| 1428 |
+
"step": 1480
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 60.0,
|
| 1432 |
+
"eval_accuracy": 0.75,
|
| 1433 |
+
"eval_loss": 1.6593661308288574,
|
| 1434 |
+
"eval_runtime": 0.4171,
|
| 1435 |
+
"eval_samples_per_second": 76.721,
|
| 1436 |
+
"eval_steps_per_second": 4.795,
|
| 1437 |
+
"step": 1485
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 60.2,
|
| 1441 |
+
"learning_rate": 2.145061728395062e-05,
|
| 1442 |
+
"loss": 0.0964,
|
| 1443 |
+
"step": 1490
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 60.61,
|
| 1447 |
+
"learning_rate": 2.12962962962963e-05,
|
| 1448 |
+
"loss": 0.0894,
|
| 1449 |
+
"step": 1500
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 60.97,
|
| 1453 |
+
"eval_accuracy": 0.78125,
|
| 1454 |
+
"eval_loss": 1.630151629447937,
|
| 1455 |
+
"eval_runtime": 0.4706,
|
| 1456 |
+
"eval_samples_per_second": 67.996,
|
| 1457 |
+
"eval_steps_per_second": 4.25,
|
| 1458 |
+
"step": 1509
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 61.01,
|
| 1462 |
+
"learning_rate": 2.1141975308641974e-05,
|
| 1463 |
+
"loss": 0.0664,
|
| 1464 |
+
"step": 1510
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 61.41,
|
| 1468 |
+
"learning_rate": 2.0987654320987655e-05,
|
| 1469 |
+
"loss": 0.0701,
|
| 1470 |
+
"step": 1520
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 61.82,
|
| 1474 |
+
"learning_rate": 2.0833333333333336e-05,
|
| 1475 |
+
"loss": 0.0999,
|
| 1476 |
+
"step": 1530
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 61.98,
|
| 1480 |
+
"eval_accuracy": 0.71875,
|
| 1481 |
+
"eval_loss": 1.656170129776001,
|
| 1482 |
+
"eval_runtime": 0.7332,
|
| 1483 |
+
"eval_samples_per_second": 43.646,
|
| 1484 |
+
"eval_steps_per_second": 2.728,
|
| 1485 |
+
"step": 1534
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 62.22,
|
| 1489 |
+
"learning_rate": 2.0679012345679014e-05,
|
| 1490 |
+
"loss": 0.0559,
|
| 1491 |
+
"step": 1540
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 62.63,
|
| 1495 |
+
"learning_rate": 2.0524691358024695e-05,
|
| 1496 |
+
"loss": 0.0759,
|
| 1497 |
+
"step": 1550
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 62.99,
|
| 1501 |
+
"eval_accuracy": 0.75,
|
| 1502 |
+
"eval_loss": 1.598868489265442,
|
| 1503 |
+
"eval_runtime": 0.4032,
|
| 1504 |
+
"eval_samples_per_second": 79.364,
|
| 1505 |
+
"eval_steps_per_second": 4.96,
|
| 1506 |
+
"step": 1559
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 63.03,
|
| 1510 |
+
"learning_rate": 2.0370370370370372e-05,
|
| 1511 |
+
"loss": 0.1008,
|
| 1512 |
+
"step": 1560
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 63.43,
|
| 1516 |
+
"learning_rate": 2.021604938271605e-05,
|
| 1517 |
+
"loss": 0.06,
|
| 1518 |
+
"step": 1570
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 63.84,
|
| 1522 |
+
"learning_rate": 2.006172839506173e-05,
|
| 1523 |
+
"loss": 0.102,
|
| 1524 |
+
"step": 1580
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 64.0,
|
| 1528 |
+
"eval_accuracy": 0.78125,
|
| 1529 |
+
"eval_loss": 1.660170078277588,
|
| 1530 |
+
"eval_runtime": 0.4806,
|
| 1531 |
+
"eval_samples_per_second": 66.581,
|
| 1532 |
+
"eval_steps_per_second": 4.161,
|
| 1533 |
+
"step": 1584
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 64.24,
|
| 1537 |
+
"learning_rate": 1.990740740740741e-05,
|
| 1538 |
+
"loss": 0.0836,
|
| 1539 |
+
"step": 1590
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 64.65,
|
| 1543 |
+
"learning_rate": 1.9753086419753087e-05,
|
| 1544 |
+
"loss": 0.0864,
|
| 1545 |
+
"step": 1600
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 64.97,
|
| 1549 |
+
"eval_accuracy": 0.78125,
|
| 1550 |
+
"eval_loss": 1.7385599613189697,
|
| 1551 |
+
"eval_runtime": 0.4181,
|
| 1552 |
+
"eval_samples_per_second": 76.538,
|
| 1553 |
+
"eval_steps_per_second": 4.784,
|
| 1554 |
+
"step": 1608
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 65.05,
|
| 1558 |
+
"learning_rate": 1.9598765432098768e-05,
|
| 1559 |
+
"loss": 0.076,
|
| 1560 |
+
"step": 1610
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 65.45,
|
| 1564 |
+
"learning_rate": 1.9444444444444445e-05,
|
| 1565 |
+
"loss": 0.0763,
|
| 1566 |
+
"step": 1620
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 65.86,
|
| 1570 |
+
"learning_rate": 1.9290123456790123e-05,
|
| 1571 |
+
"loss": 0.0722,
|
| 1572 |
+
"step": 1630
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 65.98,
|
| 1576 |
+
"eval_accuracy": 0.71875,
|
| 1577 |
+
"eval_loss": 2.0494906902313232,
|
| 1578 |
+
"eval_runtime": 0.4086,
|
| 1579 |
+
"eval_samples_per_second": 78.317,
|
| 1580 |
+
"eval_steps_per_second": 4.895,
|
| 1581 |
+
"step": 1633
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 66.26,
|
| 1585 |
+
"learning_rate": 1.9135802469135804e-05,
|
| 1586 |
+
"loss": 0.0929,
|
| 1587 |
+
"step": 1640
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 66.67,
|
| 1591 |
+
"learning_rate": 1.8981481481481482e-05,
|
| 1592 |
+
"loss": 0.0956,
|
| 1593 |
+
"step": 1650
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 66.99,
|
| 1597 |
+
"eval_accuracy": 0.6875,
|
| 1598 |
+
"eval_loss": 1.9748592376708984,
|
| 1599 |
+
"eval_runtime": 0.4206,
|
| 1600 |
+
"eval_samples_per_second": 76.081,
|
| 1601 |
+
"eval_steps_per_second": 4.755,
|
| 1602 |
+
"step": 1658
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 67.07,
|
| 1606 |
+
"learning_rate": 1.8827160493827163e-05,
|
| 1607 |
+
"loss": 0.071,
|
| 1608 |
+
"step": 1660
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 67.47,
|
| 1612 |
+
"learning_rate": 1.867283950617284e-05,
|
| 1613 |
+
"loss": 0.0844,
|
| 1614 |
+
"step": 1670
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 67.88,
|
| 1618 |
+
"learning_rate": 1.851851851851852e-05,
|
| 1619 |
+
"loss": 0.0698,
|
| 1620 |
+
"step": 1680
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 68.0,
|
| 1624 |
+
"eval_accuracy": 0.65625,
|
| 1625 |
+
"eval_loss": 2.0089621543884277,
|
| 1626 |
+
"eval_runtime": 0.4176,
|
| 1627 |
+
"eval_samples_per_second": 76.628,
|
| 1628 |
+
"eval_steps_per_second": 4.789,
|
| 1629 |
+
"step": 1683
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 68.28,
|
| 1633 |
+
"learning_rate": 1.83641975308642e-05,
|
| 1634 |
+
"loss": 0.0837,
|
| 1635 |
+
"step": 1690
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 68.69,
|
| 1639 |
+
"learning_rate": 1.820987654320988e-05,
|
| 1640 |
+
"loss": 0.0635,
|
| 1641 |
+
"step": 1700
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 68.97,
|
| 1645 |
+
"eval_accuracy": 0.625,
|
| 1646 |
+
"eval_loss": 2.160045862197876,
|
| 1647 |
+
"eval_runtime": 0.4306,
|
| 1648 |
+
"eval_samples_per_second": 74.314,
|
| 1649 |
+
"eval_steps_per_second": 4.645,
|
| 1650 |
+
"step": 1707
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 69.09,
|
| 1654 |
+
"learning_rate": 1.8055555555555558e-05,
|
| 1655 |
+
"loss": 0.0719,
|
| 1656 |
+
"step": 1710
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 69.49,
|
| 1660 |
+
"learning_rate": 1.7901234567901236e-05,
|
| 1661 |
+
"loss": 0.0656,
|
| 1662 |
+
"step": 1720
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 69.9,
|
| 1666 |
+
"learning_rate": 1.7746913580246917e-05,
|
| 1667 |
+
"loss": 0.0726,
|
| 1668 |
+
"step": 1730
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 69.98,
|
| 1672 |
+
"eval_accuracy": 0.75,
|
| 1673 |
+
"eval_loss": 1.8476567268371582,
|
| 1674 |
+
"eval_runtime": 0.4076,
|
| 1675 |
+
"eval_samples_per_second": 78.508,
|
| 1676 |
+
"eval_steps_per_second": 4.907,
|
| 1677 |
+
"step": 1732
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 70.3,
|
| 1681 |
+
"learning_rate": 1.7592592592592595e-05,
|
| 1682 |
+
"loss": 0.0651,
|
| 1683 |
+
"step": 1740
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 70.71,
|
| 1687 |
+
"learning_rate": 1.7438271604938272e-05,
|
| 1688 |
+
"loss": 0.0905,
|
| 1689 |
+
"step": 1750
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 70.99,
|
| 1693 |
+
"eval_accuracy": 0.71875,
|
| 1694 |
+
"eval_loss": 1.9970349073410034,
|
| 1695 |
+
"eval_runtime": 0.4801,
|
| 1696 |
+
"eval_samples_per_second": 66.652,
|
| 1697 |
+
"eval_steps_per_second": 4.166,
|
| 1698 |
+
"step": 1757
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 71.11,
|
| 1702 |
+
"learning_rate": 1.728395061728395e-05,
|
| 1703 |
+
"loss": 0.0705,
|
| 1704 |
+
"step": 1760
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 71.52,
|
| 1708 |
+
"learning_rate": 1.712962962962963e-05,
|
| 1709 |
+
"loss": 0.053,
|
| 1710 |
+
"step": 1770
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 71.92,
|
| 1714 |
+
"learning_rate": 1.697530864197531e-05,
|
| 1715 |
+
"loss": 0.0955,
|
| 1716 |
+
"step": 1780
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 72.0,
|
| 1720 |
+
"eval_accuracy": 0.75,
|
| 1721 |
+
"eval_loss": 1.9001073837280273,
|
| 1722 |
+
"eval_runtime": 0.4716,
|
| 1723 |
+
"eval_samples_per_second": 67.852,
|
| 1724 |
+
"eval_steps_per_second": 4.241,
|
| 1725 |
+
"step": 1782
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 72.32,
|
| 1729 |
+
"learning_rate": 1.682098765432099e-05,
|
| 1730 |
+
"loss": 0.0943,
|
| 1731 |
+
"step": 1790
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 72.73,
|
| 1735 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 1736 |
+
"loss": 0.0614,
|
| 1737 |
+
"step": 1800
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 72.97,
|
| 1741 |
+
"eval_accuracy": 0.65625,
|
| 1742 |
+
"eval_loss": 1.9346568584442139,
|
| 1743 |
+
"eval_runtime": 0.4161,
|
| 1744 |
+
"eval_samples_per_second": 76.906,
|
| 1745 |
+
"eval_steps_per_second": 4.807,
|
| 1746 |
+
"step": 1806
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 73.13,
|
| 1750 |
+
"learning_rate": 1.651234567901235e-05,
|
| 1751 |
+
"loss": 0.0652,
|
| 1752 |
+
"step": 1810
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 73.54,
|
| 1756 |
+
"learning_rate": 1.6358024691358026e-05,
|
| 1757 |
+
"loss": 0.0634,
|
| 1758 |
+
"step": 1820
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 73.94,
|
| 1762 |
+
"learning_rate": 1.6203703703703707e-05,
|
| 1763 |
+
"loss": 0.0721,
|
| 1764 |
+
"step": 1830
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 73.98,
|
| 1768 |
+
"eval_accuracy": 0.6875,
|
| 1769 |
+
"eval_loss": 1.900674819946289,
|
| 1770 |
+
"eval_runtime": 0.4776,
|
| 1771 |
+
"eval_samples_per_second": 66.996,
|
| 1772 |
+
"eval_steps_per_second": 4.187,
|
| 1773 |
+
"step": 1831
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 74.34,
|
| 1777 |
+
"learning_rate": 1.6049382716049385e-05,
|
| 1778 |
+
"loss": 0.0633,
|
| 1779 |
+
"step": 1840
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 74.75,
|
| 1783 |
+
"learning_rate": 1.5895061728395063e-05,
|
| 1784 |
+
"loss": 0.0868,
|
| 1785 |
+
"step": 1850
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 74.99,
|
| 1789 |
+
"eval_accuracy": 0.65625,
|
| 1790 |
+
"eval_loss": 2.0204198360443115,
|
| 1791 |
+
"eval_runtime": 0.4036,
|
| 1792 |
+
"eval_samples_per_second": 79.286,
|
| 1793 |
+
"eval_steps_per_second": 4.955,
|
| 1794 |
+
"step": 1856
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 75.15,
|
| 1798 |
+
"learning_rate": 1.5740740740740744e-05,
|
| 1799 |
+
"loss": 0.0798,
|
| 1800 |
+
"step": 1860
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 75.56,
|
| 1804 |
+
"learning_rate": 1.558641975308642e-05,
|
| 1805 |
+
"loss": 0.0783,
|
| 1806 |
+
"step": 1870
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 75.96,
|
| 1810 |
+
"learning_rate": 1.54320987654321e-05,
|
| 1811 |
+
"loss": 0.0817,
|
| 1812 |
+
"step": 1880
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 76.0,
|
| 1816 |
+
"eval_accuracy": 0.71875,
|
| 1817 |
+
"eval_loss": 1.9806559085845947,
|
| 1818 |
+
"eval_runtime": 0.4096,
|
| 1819 |
+
"eval_samples_per_second": 78.123,
|
| 1820 |
+
"eval_steps_per_second": 4.883,
|
| 1821 |
+
"step": 1881
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 76.36,
|
| 1825 |
+
"learning_rate": 1.5277777777777777e-05,
|
| 1826 |
+
"loss": 0.0823,
|
| 1827 |
+
"step": 1890
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 76.77,
|
| 1831 |
+
"learning_rate": 1.5123456790123458e-05,
|
| 1832 |
+
"loss": 0.0533,
|
| 1833 |
+
"step": 1900
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 76.97,
|
| 1837 |
+
"eval_accuracy": 0.75,
|
| 1838 |
+
"eval_loss": 1.978171467781067,
|
| 1839 |
+
"eval_runtime": 0.4191,
|
| 1840 |
+
"eval_samples_per_second": 76.353,
|
| 1841 |
+
"eval_steps_per_second": 4.772,
|
| 1842 |
+
"step": 1905
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 77.17,
|
| 1846 |
+
"learning_rate": 1.4969135802469137e-05,
|
| 1847 |
+
"loss": 0.0866,
|
| 1848 |
+
"step": 1910
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 77.58,
|
| 1852 |
+
"learning_rate": 1.4814814814814815e-05,
|
| 1853 |
+
"loss": 0.0737,
|
| 1854 |
+
"step": 1920
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 77.98,
|
| 1858 |
+
"learning_rate": 1.4660493827160496e-05,
|
| 1859 |
+
"loss": 0.0682,
|
| 1860 |
+
"step": 1930
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 77.98,
|
| 1864 |
+
"eval_accuracy": 0.75,
|
| 1865 |
+
"eval_loss": 1.831963062286377,
|
| 1866 |
+
"eval_runtime": 0.4216,
|
| 1867 |
+
"eval_samples_per_second": 75.902,
|
| 1868 |
+
"eval_steps_per_second": 4.744,
|
| 1869 |
+
"step": 1930
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 78.38,
|
| 1873 |
+
"learning_rate": 1.4506172839506174e-05,
|
| 1874 |
+
"loss": 0.0819,
|
| 1875 |
+
"step": 1940
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 78.79,
|
| 1879 |
+
"learning_rate": 1.4351851851851853e-05,
|
| 1880 |
+
"loss": 0.078,
|
| 1881 |
+
"step": 1950
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 78.99,
|
| 1885 |
+
"eval_accuracy": 0.71875,
|
| 1886 |
+
"eval_loss": 1.835146427154541,
|
| 1887 |
+
"eval_runtime": 0.4216,
|
| 1888 |
+
"eval_samples_per_second": 75.902,
|
| 1889 |
+
"eval_steps_per_second": 4.744,
|
| 1890 |
+
"step": 1955
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 79.19,
|
| 1894 |
+
"learning_rate": 1.4197530864197532e-05,
|
| 1895 |
+
"loss": 0.0574,
|
| 1896 |
+
"step": 1960
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 79.6,
|
| 1900 |
+
"learning_rate": 1.4043209876543212e-05,
|
| 1901 |
+
"loss": 0.0468,
|
| 1902 |
+
"step": 1970
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 80.0,
|
| 1906 |
+
"learning_rate": 1.388888888888889e-05,
|
| 1907 |
+
"loss": 0.0991,
|
| 1908 |
+
"step": 1980
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 80.0,
|
| 1912 |
+
"eval_accuracy": 0.71875,
|
| 1913 |
+
"eval_loss": 1.9694030284881592,
|
| 1914 |
+
"eval_runtime": 0.4798,
|
| 1915 |
+
"eval_samples_per_second": 66.7,
|
| 1916 |
+
"eval_steps_per_second": 4.169,
|
| 1917 |
+
"step": 1980
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 80.4,
|
| 1921 |
+
"learning_rate": 1.373456790123457e-05,
|
| 1922 |
+
"loss": 0.0466,
|
| 1923 |
+
"step": 1990
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 80.81,
|
| 1927 |
+
"learning_rate": 1.3580246913580248e-05,
|
| 1928 |
+
"loss": 0.0601,
|
| 1929 |
+
"step": 2000
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 80.97,
|
| 1933 |
+
"eval_accuracy": 0.71875,
|
| 1934 |
+
"eval_loss": 1.8794611692428589,
|
| 1935 |
+
"eval_runtime": 0.4141,
|
| 1936 |
+
"eval_samples_per_second": 77.276,
|
| 1937 |
+
"eval_steps_per_second": 4.83,
|
| 1938 |
+
"step": 2004
|
| 1939 |
+
},
|
| 1940 |
+
{
|
| 1941 |
+
"epoch": 81.21,
|
| 1942 |
+
"learning_rate": 1.3425925925925926e-05,
|
| 1943 |
+
"loss": 0.0567,
|
| 1944 |
+
"step": 2010
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 81.62,
|
| 1948 |
+
"learning_rate": 1.3271604938271605e-05,
|
| 1949 |
+
"loss": 0.072,
|
| 1950 |
+
"step": 2020
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 81.98,
|
| 1954 |
+
"eval_accuracy": 0.65625,
|
| 1955 |
+
"eval_loss": 2.02938175201416,
|
| 1956 |
+
"eval_runtime": 0.4232,
|
| 1957 |
+
"eval_samples_per_second": 75.623,
|
| 1958 |
+
"eval_steps_per_second": 4.726,
|
| 1959 |
+
"step": 2029
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 82.02,
|
| 1963 |
+
"learning_rate": 1.3117283950617285e-05,
|
| 1964 |
+
"loss": 0.073,
|
| 1965 |
+
"step": 2030
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 82.42,
|
| 1969 |
+
"learning_rate": 1.2962962962962964e-05,
|
| 1970 |
+
"loss": 0.0788,
|
| 1971 |
+
"step": 2040
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 82.83,
|
| 1975 |
+
"learning_rate": 1.2808641975308642e-05,
|
| 1976 |
+
"loss": 0.0746,
|
| 1977 |
+
"step": 2050
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 82.99,
|
| 1981 |
+
"eval_accuracy": 0.71875,
|
| 1982 |
+
"eval_loss": 1.8438613414764404,
|
| 1983 |
+
"eval_runtime": 0.3956,
|
| 1984 |
+
"eval_samples_per_second": 80.89,
|
| 1985 |
+
"eval_steps_per_second": 5.056,
|
| 1986 |
+
"step": 2054
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 83.23,
|
| 1990 |
+
"learning_rate": 1.2654320987654323e-05,
|
| 1991 |
+
"loss": 0.0626,
|
| 1992 |
+
"step": 2060
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 83.64,
|
| 1996 |
+
"learning_rate": 1.25e-05,
|
| 1997 |
+
"loss": 0.0547,
|
| 1998 |
+
"step": 2070
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 84.0,
|
| 2002 |
+
"eval_accuracy": 0.71875,
|
| 2003 |
+
"eval_loss": 1.932082176208496,
|
| 2004 |
+
"eval_runtime": 0.4141,
|
| 2005 |
+
"eval_samples_per_second": 77.273,
|
| 2006 |
+
"eval_steps_per_second": 4.83,
|
| 2007 |
+
"step": 2079
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 84.04,
|
| 2011 |
+
"learning_rate": 1.234567901234568e-05,
|
| 2012 |
+
"loss": 0.0632,
|
| 2013 |
+
"step": 2080
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 84.44,
|
| 2017 |
+
"learning_rate": 1.219135802469136e-05,
|
| 2018 |
+
"loss": 0.0478,
|
| 2019 |
+
"step": 2090
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 84.85,
|
| 2023 |
+
"learning_rate": 1.2037037037037039e-05,
|
| 2024 |
+
"loss": 0.0497,
|
| 2025 |
+
"step": 2100
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 84.97,
|
| 2029 |
+
"eval_accuracy": 0.78125,
|
| 2030 |
+
"eval_loss": 1.8862378597259521,
|
| 2031 |
+
"eval_runtime": 0.4806,
|
| 2032 |
+
"eval_samples_per_second": 66.582,
|
| 2033 |
+
"eval_steps_per_second": 4.161,
|
| 2034 |
+
"step": 2103
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 85.25,
|
| 2038 |
+
"learning_rate": 1.1882716049382716e-05,
|
| 2039 |
+
"loss": 0.0544,
|
| 2040 |
+
"step": 2110
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 85.66,
|
| 2044 |
+
"learning_rate": 1.1728395061728398e-05,
|
| 2045 |
+
"loss": 0.0566,
|
| 2046 |
+
"step": 2120
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 85.98,
|
| 2050 |
+
"eval_accuracy": 0.65625,
|
| 2051 |
+
"eval_loss": 2.0067098140716553,
|
| 2052 |
+
"eval_runtime": 0.5932,
|
| 2053 |
+
"eval_samples_per_second": 53.947,
|
| 2054 |
+
"eval_steps_per_second": 3.372,
|
| 2055 |
+
"step": 2128
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 86.06,
|
| 2059 |
+
"learning_rate": 1.1574074074074075e-05,
|
| 2060 |
+
"loss": 0.0557,
|
| 2061 |
+
"step": 2130
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 86.46,
|
| 2065 |
+
"learning_rate": 1.1419753086419753e-05,
|
| 2066 |
+
"loss": 0.0872,
|
| 2067 |
+
"step": 2140
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 86.87,
|
| 2071 |
+
"learning_rate": 1.1265432098765432e-05,
|
| 2072 |
+
"loss": 0.0353,
|
| 2073 |
+
"step": 2150
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 86.99,
|
| 2077 |
+
"eval_accuracy": 0.71875,
|
| 2078 |
+
"eval_loss": 2.095703363418579,
|
| 2079 |
+
"eval_runtime": 0.4511,
|
| 2080 |
+
"eval_samples_per_second": 70.942,
|
| 2081 |
+
"eval_steps_per_second": 4.434,
|
| 2082 |
+
"step": 2153
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 87.27,
|
| 2086 |
+
"learning_rate": 1.1111111111111113e-05,
|
| 2087 |
+
"loss": 0.0482,
|
| 2088 |
+
"step": 2160
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 87.68,
|
| 2092 |
+
"learning_rate": 1.0956790123456791e-05,
|
| 2093 |
+
"loss": 0.0634,
|
| 2094 |
+
"step": 2170
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 88.0,
|
| 2098 |
+
"eval_accuracy": 0.65625,
|
| 2099 |
+
"eval_loss": 2.15714168548584,
|
| 2100 |
+
"eval_runtime": 0.4471,
|
| 2101 |
+
"eval_samples_per_second": 71.57,
|
| 2102 |
+
"eval_steps_per_second": 4.473,
|
| 2103 |
+
"step": 2178
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 88.08,
|
| 2107 |
+
"learning_rate": 1.0802469135802469e-05,
|
| 2108 |
+
"loss": 0.0632,
|
| 2109 |
+
"step": 2180
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 88.48,
|
| 2113 |
+
"learning_rate": 1.064814814814815e-05,
|
| 2114 |
+
"loss": 0.0504,
|
| 2115 |
+
"step": 2190
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 88.89,
|
| 2119 |
+
"learning_rate": 1.0493827160493827e-05,
|
| 2120 |
+
"loss": 0.0477,
|
| 2121 |
+
"step": 2200
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 88.97,
|
| 2125 |
+
"eval_accuracy": 0.6875,
|
| 2126 |
+
"eval_loss": 2.0384438037872314,
|
| 2127 |
+
"eval_runtime": 0.4271,
|
| 2128 |
+
"eval_samples_per_second": 74.924,
|
| 2129 |
+
"eval_steps_per_second": 4.683,
|
| 2130 |
+
"step": 2202
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 89.29,
|
| 2134 |
+
"learning_rate": 1.0339506172839507e-05,
|
| 2135 |
+
"loss": 0.0446,
|
| 2136 |
+
"step": 2210
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 89.7,
|
| 2140 |
+
"learning_rate": 1.0185185185185186e-05,
|
| 2141 |
+
"loss": 0.0513,
|
| 2142 |
+
"step": 2220
|
| 2143 |
+
},
|
| 2144 |
+
{
|
| 2145 |
+
"epoch": 89.98,
|
| 2146 |
+
"eval_accuracy": 0.75,
|
| 2147 |
+
"eval_loss": 1.9145632982254028,
|
| 2148 |
+
"eval_runtime": 0.4711,
|
| 2149 |
+
"eval_samples_per_second": 67.923,
|
| 2150 |
+
"eval_steps_per_second": 4.245,
|
| 2151 |
+
"step": 2227
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 90.1,
|
| 2155 |
+
"learning_rate": 1.0030864197530866e-05,
|
| 2156 |
+
"loss": 0.0626,
|
| 2157 |
+
"step": 2230
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 90.51,
|
| 2161 |
+
"learning_rate": 9.876543209876543e-06,
|
| 2162 |
+
"loss": 0.0354,
|
| 2163 |
+
"step": 2240
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 90.91,
|
| 2167 |
+
"learning_rate": 9.722222222222223e-06,
|
| 2168 |
+
"loss": 0.0717,
|
| 2169 |
+
"step": 2250
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 90.99,
|
| 2173 |
+
"eval_accuracy": 0.71875,
|
| 2174 |
+
"eval_loss": 1.8837898969650269,
|
| 2175 |
+
"eval_runtime": 0.4031,
|
| 2176 |
+
"eval_samples_per_second": 79.378,
|
| 2177 |
+
"eval_steps_per_second": 4.961,
|
| 2178 |
+
"step": 2252
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 91.31,
|
| 2182 |
+
"learning_rate": 9.567901234567902e-06,
|
| 2183 |
+
"loss": 0.0486,
|
| 2184 |
+
"step": 2260
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 91.72,
|
| 2188 |
+
"learning_rate": 9.413580246913581e-06,
|
| 2189 |
+
"loss": 0.0644,
|
| 2190 |
+
"step": 2270
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 92.0,
|
| 2194 |
+
"eval_accuracy": 0.6875,
|
| 2195 |
+
"eval_loss": 1.9186292886734009,
|
| 2196 |
+
"eval_runtime": 0.4621,
|
| 2197 |
+
"eval_samples_per_second": 69.247,
|
| 2198 |
+
"eval_steps_per_second": 4.328,
|
| 2199 |
+
"step": 2277
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 92.12,
|
| 2203 |
+
"learning_rate": 9.25925925925926e-06,
|
| 2204 |
+
"loss": 0.048,
|
| 2205 |
+
"step": 2280
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 92.53,
|
| 2209 |
+
"learning_rate": 9.10493827160494e-06,
|
| 2210 |
+
"loss": 0.0711,
|
| 2211 |
+
"step": 2290
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 92.93,
|
| 2215 |
+
"learning_rate": 8.950617283950618e-06,
|
| 2216 |
+
"loss": 0.0848,
|
| 2217 |
+
"step": 2300
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 92.97,
|
| 2221 |
+
"eval_accuracy": 0.71875,
|
| 2222 |
+
"eval_loss": 1.882826805114746,
|
| 2223 |
+
"eval_runtime": 0.4741,
|
| 2224 |
+
"eval_samples_per_second": 67.495,
|
| 2225 |
+
"eval_steps_per_second": 4.218,
|
| 2226 |
+
"step": 2301
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 93.33,
|
| 2230 |
+
"learning_rate": 8.796296296296297e-06,
|
| 2231 |
+
"loss": 0.038,
|
| 2232 |
+
"step": 2310
|
| 2233 |
+
},
|
| 2234 |
+
{
|
| 2235 |
+
"epoch": 93.74,
|
| 2236 |
+
"learning_rate": 8.641975308641975e-06,
|
| 2237 |
+
"loss": 0.0393,
|
| 2238 |
+
"step": 2320
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 93.98,
|
| 2242 |
+
"eval_accuracy": 0.71875,
|
| 2243 |
+
"eval_loss": 1.9442168474197388,
|
| 2244 |
+
"eval_runtime": 0.4091,
|
| 2245 |
+
"eval_samples_per_second": 78.217,
|
| 2246 |
+
"eval_steps_per_second": 4.889,
|
| 2247 |
+
"step": 2326
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 94.14,
|
| 2251 |
+
"learning_rate": 8.487654320987654e-06,
|
| 2252 |
+
"loss": 0.038,
|
| 2253 |
+
"step": 2330
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 94.55,
|
| 2257 |
+
"learning_rate": 8.333333333333334e-06,
|
| 2258 |
+
"loss": 0.0703,
|
| 2259 |
+
"step": 2340
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 94.95,
|
| 2263 |
+
"learning_rate": 8.179012345679013e-06,
|
| 2264 |
+
"loss": 0.046,
|
| 2265 |
+
"step": 2350
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 94.99,
|
| 2269 |
+
"eval_accuracy": 0.71875,
|
| 2270 |
+
"eval_loss": 1.8865550756454468,
|
| 2271 |
+
"eval_runtime": 0.3931,
|
| 2272 |
+
"eval_samples_per_second": 81.403,
|
| 2273 |
+
"eval_steps_per_second": 5.088,
|
| 2274 |
+
"step": 2351
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 95.35,
|
| 2278 |
+
"learning_rate": 8.024691358024692e-06,
|
| 2279 |
+
"loss": 0.0762,
|
| 2280 |
+
"step": 2360
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 95.76,
|
| 2284 |
+
"learning_rate": 7.870370370370372e-06,
|
| 2285 |
+
"loss": 0.0487,
|
| 2286 |
+
"step": 2370
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 96.0,
|
| 2290 |
+
"eval_accuracy": 0.6875,
|
| 2291 |
+
"eval_loss": 1.9787421226501465,
|
| 2292 |
+
"eval_runtime": 0.4221,
|
| 2293 |
+
"eval_samples_per_second": 75.807,
|
| 2294 |
+
"eval_steps_per_second": 4.738,
|
| 2295 |
+
"step": 2376
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 96.16,
|
| 2299 |
+
"learning_rate": 7.71604938271605e-06,
|
| 2300 |
+
"loss": 0.0724,
|
| 2301 |
+
"step": 2380
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 96.57,
|
| 2305 |
+
"learning_rate": 7.561728395061729e-06,
|
| 2306 |
+
"loss": 0.0563,
|
| 2307 |
+
"step": 2390
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 96.97,
|
| 2311 |
+
"learning_rate": 7.4074074074074075e-06,
|
| 2312 |
+
"loss": 0.074,
|
| 2313 |
+
"step": 2400
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 96.97,
|
| 2317 |
+
"eval_accuracy": 0.6875,
|
| 2318 |
+
"eval_loss": 2.0081026554107666,
|
| 2319 |
+
"eval_runtime": 0.6192,
|
| 2320 |
+
"eval_samples_per_second": 51.682,
|
| 2321 |
+
"eval_steps_per_second": 3.23,
|
| 2322 |
+
"step": 2400
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 97.37,
|
| 2326 |
+
"learning_rate": 7.253086419753087e-06,
|
| 2327 |
+
"loss": 0.0624,
|
| 2328 |
+
"step": 2410
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 97.78,
|
| 2332 |
+
"learning_rate": 7.098765432098766e-06,
|
| 2333 |
+
"loss": 0.0435,
|
| 2334 |
+
"step": 2420
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 97.98,
|
| 2338 |
+
"eval_accuracy": 0.75,
|
| 2339 |
+
"eval_loss": 1.8838809728622437,
|
| 2340 |
+
"eval_runtime": 0.4211,
|
| 2341 |
+
"eval_samples_per_second": 75.988,
|
| 2342 |
+
"eval_steps_per_second": 4.749,
|
| 2343 |
+
"step": 2425
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 98.18,
|
| 2347 |
+
"learning_rate": 6.944444444444445e-06,
|
| 2348 |
+
"loss": 0.0688,
|
| 2349 |
+
"step": 2430
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 98.59,
|
| 2353 |
+
"learning_rate": 6.790123456790124e-06,
|
| 2354 |
+
"loss": 0.0441,
|
| 2355 |
+
"step": 2440
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 98.99,
|
| 2359 |
+
"learning_rate": 6.635802469135803e-06,
|
| 2360 |
+
"loss": 0.0509,
|
| 2361 |
+
"step": 2450
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 98.99,
|
| 2365 |
+
"eval_accuracy": 0.71875,
|
| 2366 |
+
"eval_loss": 1.9207700490951538,
|
| 2367 |
+
"eval_runtime": 0.4015,
|
| 2368 |
+
"eval_samples_per_second": 79.693,
|
| 2369 |
+
"eval_steps_per_second": 4.981,
|
| 2370 |
+
"step": 2450
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 99.39,
|
| 2374 |
+
"learning_rate": 6.481481481481482e-06,
|
| 2375 |
+
"loss": 0.0557,
|
| 2376 |
+
"step": 2460
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 99.8,
|
| 2380 |
+
"learning_rate": 6.3271604938271615e-06,
|
| 2381 |
+
"loss": 0.0571,
|
| 2382 |
+
"step": 2470
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 100.0,
|
| 2386 |
+
"eval_accuracy": 0.71875,
|
| 2387 |
+
"eval_loss": 1.9770445823669434,
|
| 2388 |
+
"eval_runtime": 0.4191,
|
| 2389 |
+
"eval_samples_per_second": 76.363,
|
| 2390 |
+
"eval_steps_per_second": 4.773,
|
| 2391 |
+
"step": 2475
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 100.2,
|
| 2395 |
+
"learning_rate": 6.17283950617284e-06,
|
| 2396 |
+
"loss": 0.0505,
|
| 2397 |
+
"step": 2480
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 100.61,
|
| 2401 |
+
"learning_rate": 6.018518518518519e-06,
|
| 2402 |
+
"loss": 0.0327,
|
| 2403 |
+
"step": 2490
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 100.97,
|
| 2407 |
+
"eval_accuracy": 0.71875,
|
| 2408 |
+
"eval_loss": 1.969954013824463,
|
| 2409 |
+
"eval_runtime": 0.5071,
|
| 2410 |
+
"eval_samples_per_second": 63.101,
|
| 2411 |
+
"eval_steps_per_second": 3.944,
|
| 2412 |
+
"step": 2499
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 101.01,
|
| 2416 |
+
"learning_rate": 5.864197530864199e-06,
|
| 2417 |
+
"loss": 0.0472,
|
| 2418 |
+
"step": 2500
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 101.41,
|
| 2422 |
+
"learning_rate": 5.7098765432098764e-06,
|
| 2423 |
+
"loss": 0.0512,
|
| 2424 |
+
"step": 2510
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 101.82,
|
| 2428 |
+
"learning_rate": 5.555555555555557e-06,
|
| 2429 |
+
"loss": 0.0387,
|
| 2430 |
+
"step": 2520
|
| 2431 |
+
},
|
| 2432 |
+
{
|
| 2433 |
+
"epoch": 101.98,
|
| 2434 |
+
"eval_accuracy": 0.75,
|
| 2435 |
+
"eval_loss": 1.9250848293304443,
|
| 2436 |
+
"eval_runtime": 0.4241,
|
| 2437 |
+
"eval_samples_per_second": 75.45,
|
| 2438 |
+
"eval_steps_per_second": 4.716,
|
| 2439 |
+
"step": 2524
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 102.22,
|
| 2443 |
+
"learning_rate": 5.401234567901234e-06,
|
| 2444 |
+
"loss": 0.0614,
|
| 2445 |
+
"step": 2530
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 102.63,
|
| 2449 |
+
"learning_rate": 5.246913580246914e-06,
|
| 2450 |
+
"loss": 0.029,
|
| 2451 |
+
"step": 2540
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 102.99,
|
| 2455 |
+
"eval_accuracy": 0.71875,
|
| 2456 |
+
"eval_loss": 1.9489786624908447,
|
| 2457 |
+
"eval_runtime": 0.4031,
|
| 2458 |
+
"eval_samples_per_second": 79.385,
|
| 2459 |
+
"eval_steps_per_second": 4.962,
|
| 2460 |
+
"step": 2549
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 103.03,
|
| 2464 |
+
"learning_rate": 5.092592592592593e-06,
|
| 2465 |
+
"loss": 0.052,
|
| 2466 |
+
"step": 2550
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 103.43,
|
| 2470 |
+
"learning_rate": 4.938271604938272e-06,
|
| 2471 |
+
"loss": 0.0696,
|
| 2472 |
+
"step": 2560
|
| 2473 |
+
},
|
| 2474 |
+
{
|
| 2475 |
+
"epoch": 103.84,
|
| 2476 |
+
"learning_rate": 4.783950617283951e-06,
|
| 2477 |
+
"loss": 0.0478,
|
| 2478 |
+
"step": 2570
|
| 2479 |
+
},
|
| 2480 |
+
{
|
| 2481 |
+
"epoch": 104.0,
|
| 2482 |
+
"eval_accuracy": 0.71875,
|
| 2483 |
+
"eval_loss": 1.9358038902282715,
|
| 2484 |
+
"eval_runtime": 0.4036,
|
| 2485 |
+
"eval_samples_per_second": 79.283,
|
| 2486 |
+
"eval_steps_per_second": 4.955,
|
| 2487 |
+
"step": 2574
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 104.24,
|
| 2491 |
+
"learning_rate": 4.62962962962963e-06,
|
| 2492 |
+
"loss": 0.0358,
|
| 2493 |
+
"step": 2580
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 104.65,
|
| 2497 |
+
"learning_rate": 4.475308641975309e-06,
|
| 2498 |
+
"loss": 0.0587,
|
| 2499 |
+
"step": 2590
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 104.97,
|
| 2503 |
+
"eval_accuracy": 0.75,
|
| 2504 |
+
"eval_loss": 1.9197365045547485,
|
| 2505 |
+
"eval_runtime": 0.4296,
|
| 2506 |
+
"eval_samples_per_second": 74.488,
|
| 2507 |
+
"eval_steps_per_second": 4.655,
|
| 2508 |
+
"step": 2598
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 105.05,
|
| 2512 |
+
"learning_rate": 4.3209876543209875e-06,
|
| 2513 |
+
"loss": 0.0417,
|
| 2514 |
+
"step": 2600
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 105.45,
|
| 2518 |
+
"learning_rate": 4.166666666666667e-06,
|
| 2519 |
+
"loss": 0.0644,
|
| 2520 |
+
"step": 2610
|
| 2521 |
+
},
|
| 2522 |
+
{
|
| 2523 |
+
"epoch": 105.86,
|
| 2524 |
+
"learning_rate": 4.012345679012346e-06,
|
| 2525 |
+
"loss": 0.0523,
|
| 2526 |
+
"step": 2620
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 105.98,
|
| 2530 |
+
"eval_accuracy": 0.71875,
|
| 2531 |
+
"eval_loss": 1.9309440851211548,
|
| 2532 |
+
"eval_runtime": 0.4621,
|
| 2533 |
+
"eval_samples_per_second": 69.248,
|
| 2534 |
+
"eval_steps_per_second": 4.328,
|
| 2535 |
+
"step": 2623
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 106.26,
|
| 2539 |
+
"learning_rate": 3.858024691358025e-06,
|
| 2540 |
+
"loss": 0.0365,
|
| 2541 |
+
"step": 2630
|
| 2542 |
+
},
|
| 2543 |
+
{
|
| 2544 |
+
"epoch": 106.67,
|
| 2545 |
+
"learning_rate": 3.7037037037037037e-06,
|
| 2546 |
+
"loss": 0.0581,
|
| 2547 |
+
"step": 2640
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 106.99,
|
| 2551 |
+
"eval_accuracy": 0.71875,
|
| 2552 |
+
"eval_loss": 1.9828894138336182,
|
| 2553 |
+
"eval_runtime": 0.4227,
|
| 2554 |
+
"eval_samples_per_second": 75.712,
|
| 2555 |
+
"eval_steps_per_second": 4.732,
|
| 2556 |
+
"step": 2648
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 107.07,
|
| 2560 |
+
"learning_rate": 3.549382716049383e-06,
|
| 2561 |
+
"loss": 0.0459,
|
| 2562 |
+
"step": 2650
|
| 2563 |
+
},
|
| 2564 |
+
{
|
| 2565 |
+
"epoch": 107.47,
|
| 2566 |
+
"learning_rate": 3.395061728395062e-06,
|
| 2567 |
+
"loss": 0.031,
|
| 2568 |
+
"step": 2660
|
| 2569 |
+
},
|
| 2570 |
+
{
|
| 2571 |
+
"epoch": 107.88,
|
| 2572 |
+
"learning_rate": 3.240740740740741e-06,
|
| 2573 |
+
"loss": 0.0352,
|
| 2574 |
+
"step": 2670
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 108.0,
|
| 2578 |
+
"eval_accuracy": 0.6875,
|
| 2579 |
+
"eval_loss": 2.00472354888916,
|
| 2580 |
+
"eval_runtime": 0.4442,
|
| 2581 |
+
"eval_samples_per_second": 72.042,
|
| 2582 |
+
"eval_steps_per_second": 4.503,
|
| 2583 |
+
"step": 2673
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 108.28,
|
| 2587 |
+
"learning_rate": 3.08641975308642e-06,
|
| 2588 |
+
"loss": 0.0553,
|
| 2589 |
+
"step": 2680
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 108.69,
|
| 2593 |
+
"learning_rate": 2.9320987654320994e-06,
|
| 2594 |
+
"loss": 0.0373,
|
| 2595 |
+
"step": 2690
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 108.97,
|
| 2599 |
+
"eval_accuracy": 0.71875,
|
| 2600 |
+
"eval_loss": 1.9897255897521973,
|
| 2601 |
+
"eval_runtime": 0.4256,
|
| 2602 |
+
"eval_samples_per_second": 75.187,
|
| 2603 |
+
"eval_steps_per_second": 4.699,
|
| 2604 |
+
"step": 2697
|
| 2605 |
+
},
|
| 2606 |
+
{
|
| 2607 |
+
"epoch": 109.09,
|
| 2608 |
+
"learning_rate": 2.7777777777777783e-06,
|
| 2609 |
+
"loss": 0.049,
|
| 2610 |
+
"step": 2700
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 109.49,
|
| 2614 |
+
"learning_rate": 2.623456790123457e-06,
|
| 2615 |
+
"loss": 0.0294,
|
| 2616 |
+
"step": 2710
|
| 2617 |
+
},
|
| 2618 |
+
{
|
| 2619 |
+
"epoch": 109.9,
|
| 2620 |
+
"learning_rate": 2.469135802469136e-06,
|
| 2621 |
+
"loss": 0.0258,
|
| 2622 |
+
"step": 2720
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 109.98,
|
| 2626 |
+
"eval_accuracy": 0.71875,
|
| 2627 |
+
"eval_loss": 1.938352346420288,
|
| 2628 |
+
"eval_runtime": 0.4216,
|
| 2629 |
+
"eval_samples_per_second": 75.898,
|
| 2630 |
+
"eval_steps_per_second": 4.744,
|
| 2631 |
+
"step": 2722
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 110.3,
|
| 2635 |
+
"learning_rate": 2.314814814814815e-06,
|
| 2636 |
+
"loss": 0.0484,
|
| 2637 |
+
"step": 2730
|
| 2638 |
+
},
|
| 2639 |
+
{
|
| 2640 |
+
"epoch": 110.71,
|
| 2641 |
+
"learning_rate": 2.1604938271604937e-06,
|
| 2642 |
+
"loss": 0.039,
|
| 2643 |
+
"step": 2740
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 110.99,
|
| 2647 |
+
"eval_accuracy": 0.6875,
|
| 2648 |
+
"eval_loss": 1.9356164932250977,
|
| 2649 |
+
"eval_runtime": 0.4076,
|
| 2650 |
+
"eval_samples_per_second": 78.509,
|
| 2651 |
+
"eval_steps_per_second": 4.907,
|
| 2652 |
+
"step": 2747
|
| 2653 |
+
},
|
| 2654 |
+
{
|
| 2655 |
+
"epoch": 111.11,
|
| 2656 |
+
"learning_rate": 2.006172839506173e-06,
|
| 2657 |
+
"loss": 0.0326,
|
| 2658 |
+
"step": 2750
|
| 2659 |
+
},
|
| 2660 |
+
{
|
| 2661 |
+
"epoch": 111.52,
|
| 2662 |
+
"learning_rate": 1.8518518518518519e-06,
|
| 2663 |
+
"loss": 0.066,
|
| 2664 |
+
"step": 2760
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 111.92,
|
| 2668 |
+
"learning_rate": 1.697530864197531e-06,
|
| 2669 |
+
"loss": 0.0333,
|
| 2670 |
+
"step": 2770
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 112.0,
|
| 2674 |
+
"eval_accuracy": 0.71875,
|
| 2675 |
+
"eval_loss": 1.980454683303833,
|
| 2676 |
+
"eval_runtime": 0.4112,
|
| 2677 |
+
"eval_samples_per_second": 77.828,
|
| 2678 |
+
"eval_steps_per_second": 4.864,
|
| 2679 |
+
"step": 2772
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 112.32,
|
| 2683 |
+
"learning_rate": 1.54320987654321e-06,
|
| 2684 |
+
"loss": 0.0524,
|
| 2685 |
+
"step": 2780
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 112.73,
|
| 2689 |
+
"learning_rate": 1.3888888888888892e-06,
|
| 2690 |
+
"loss": 0.0641,
|
| 2691 |
+
"step": 2790
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 112.97,
|
| 2695 |
+
"eval_accuracy": 0.6875,
|
| 2696 |
+
"eval_loss": 1.9813566207885742,
|
| 2697 |
+
"eval_runtime": 0.5541,
|
| 2698 |
+
"eval_samples_per_second": 57.748,
|
| 2699 |
+
"eval_steps_per_second": 3.609,
|
| 2700 |
+
"step": 2796
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 113.13,
|
| 2704 |
+
"learning_rate": 1.234567901234568e-06,
|
| 2705 |
+
"loss": 0.0505,
|
| 2706 |
+
"step": 2800
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 113.54,
|
| 2710 |
+
"learning_rate": 1.0802469135802469e-06,
|
| 2711 |
+
"loss": 0.0222,
|
| 2712 |
+
"step": 2810
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 113.94,
|
| 2716 |
+
"learning_rate": 9.259259259259259e-07,
|
| 2717 |
+
"loss": 0.0649,
|
| 2718 |
+
"step": 2820
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 113.98,
|
| 2722 |
+
"eval_accuracy": 0.6875,
|
| 2723 |
+
"eval_loss": 1.972588062286377,
|
| 2724 |
+
"eval_runtime": 0.4431,
|
| 2725 |
+
"eval_samples_per_second": 72.211,
|
| 2726 |
+
"eval_steps_per_second": 4.513,
|
| 2727 |
+
"step": 2821
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 114.34,
|
| 2731 |
+
"learning_rate": 7.71604938271605e-07,
|
| 2732 |
+
"loss": 0.0354,
|
| 2733 |
+
"step": 2830
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 114.75,
|
| 2737 |
+
"learning_rate": 6.17283950617284e-07,
|
| 2738 |
+
"loss": 0.0241,
|
| 2739 |
+
"step": 2840
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 114.99,
|
| 2743 |
+
"eval_accuracy": 0.6875,
|
| 2744 |
+
"eval_loss": 1.9736533164978027,
|
| 2745 |
+
"eval_runtime": 0.4011,
|
| 2746 |
+
"eval_samples_per_second": 79.777,
|
| 2747 |
+
"eval_steps_per_second": 4.986,
|
| 2748 |
+
"step": 2846
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 115.15,
|
| 2752 |
+
"learning_rate": 4.6296296296296297e-07,
|
| 2753 |
+
"loss": 0.0418,
|
| 2754 |
+
"step": 2850
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 115.56,
|
| 2758 |
+
"learning_rate": 3.08641975308642e-07,
|
| 2759 |
+
"loss": 0.0382,
|
| 2760 |
+
"step": 2860
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 115.96,
|
| 2764 |
+
"learning_rate": 1.54320987654321e-07,
|
| 2765 |
+
"loss": 0.0356,
|
| 2766 |
+
"step": 2870
|
| 2767 |
+
},
|
| 2768 |
+
{
|
| 2769 |
+
"epoch": 116.0,
|
| 2770 |
+
"eval_accuracy": 0.6875,
|
| 2771 |
+
"eval_loss": 1.985576868057251,
|
| 2772 |
+
"eval_runtime": 0.4721,
|
| 2773 |
+
"eval_samples_per_second": 67.78,
|
| 2774 |
+
"eval_steps_per_second": 4.236,
|
| 2775 |
+
"step": 2871
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 116.36,
|
| 2779 |
+
"learning_rate": 0.0,
|
| 2780 |
+
"loss": 0.0601,
|
| 2781 |
+
"step": 2880
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 116.36,
|
| 2785 |
+
"eval_accuracy": 0.6875,
|
| 2786 |
+
"eval_loss": 1.9852969646453857,
|
| 2787 |
+
"eval_runtime": 0.4181,
|
| 2788 |
+
"eval_samples_per_second": 76.535,
|
| 2789 |
+
"eval_steps_per_second": 4.783,
|
| 2790 |
+
"step": 2880
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 116.36,
|
| 2794 |
+
"step": 2880,
|
| 2795 |
+
"total_flos": 5.970642494234296e+18,
|
| 2796 |
+
"train_loss": 0.4532645735475752,
|
| 2797 |
+
"train_runtime": 3508.1723,
|
| 2798 |
+
"train_samples_per_second": 53.943,
|
| 2799 |
+
"train_steps_per_second": 0.821
|
| 2800 |
+
}
|
| 2801 |
+
],
|
| 2802 |
+
"logging_steps": 10,
|
| 2803 |
+
"max_steps": 2880,
|
| 2804 |
+
"num_input_tokens_seen": 0,
|
| 2805 |
+
"num_train_epochs": 120,
|
| 2806 |
+
"save_steps": 500,
|
| 2807 |
+
"total_flos": 5.970642494234296e+18,
|
| 2808 |
+
"train_batch_size": 16,
|
| 2809 |
+
"trial_name": null,
|
| 2810 |
+
"trial_params": null
|
| 2811 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f785aad88b21acf0d08fe9f2cc7cc5ba340f9f6c87d81a6c308e2c76468704e7
|
| 3 |
+
size 4728
|