End of training
Browse files- README.md +195 -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_18-02-31_DESKTOP-SKBE9FB/events.out.tfevents.1733443352.DESKTOP-SKBE9FB.17596.0 +3 -0
- runs/Dec05_18-02-31_DESKTOP-SKBE9FB/events.out.tfevents.1733446706.DESKTOP-SKBE9FB.17596.1 +3 -0
- train_results.json +7 -0
- trainer_state.json +2676 -0
- training_args.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,195 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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-DA3-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: 0.782608695652174
|
| 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-DA3-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: 1.6604
|
| 36 |
+
- Accuracy: 0.7826
|
| 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: 5e-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 |
+
| 1.4177 | 0.98 | 22 | 1.3455 | 0.4348 |
|
| 71 |
+
| 1.4043 | 2.0 | 45 | 1.3475 | 0.4565 |
|
| 72 |
+
| 1.3628 | 2.98 | 67 | 1.3397 | 0.4565 |
|
| 73 |
+
| 1.21 | 4.0 | 90 | 1.2593 | 0.4565 |
|
| 74 |
+
| 1.0504 | 4.98 | 112 | 1.1194 | 0.4130 |
|
| 75 |
+
| 0.9129 | 6.0 | 135 | 1.0522 | 0.4130 |
|
| 76 |
+
| 0.7811 | 6.98 | 157 | 1.1184 | 0.4348 |
|
| 77 |
+
| 0.6572 | 8.0 | 180 | 0.9951 | 0.5870 |
|
| 78 |
+
| 0.5207 | 8.98 | 202 | 0.9055 | 0.6522 |
|
| 79 |
+
| 0.6234 | 10.0 | 225 | 0.9277 | 0.6087 |
|
| 80 |
+
| 0.4721 | 10.98 | 247 | 0.8458 | 0.6739 |
|
| 81 |
+
| 0.3944 | 12.0 | 270 | 0.8837 | 0.6522 |
|
| 82 |
+
| 0.3572 | 12.98 | 292 | 0.8719 | 0.7174 |
|
| 83 |
+
| 0.2911 | 14.0 | 315 | 1.0367 | 0.6304 |
|
| 84 |
+
| 0.3224 | 14.98 | 337 | 0.9641 | 0.6087 |
|
| 85 |
+
| 0.2663 | 16.0 | 360 | 1.3670 | 0.5870 |
|
| 86 |
+
| 0.2132 | 16.98 | 382 | 1.3090 | 0.6304 |
|
| 87 |
+
| 0.266 | 18.0 | 405 | 1.1247 | 0.7174 |
|
| 88 |
+
| 0.1929 | 18.98 | 427 | 1.1458 | 0.6087 |
|
| 89 |
+
| 0.1831 | 20.0 | 450 | 1.0267 | 0.7391 |
|
| 90 |
+
| 0.2298 | 20.98 | 472 | 1.1863 | 0.6304 |
|
| 91 |
+
| 0.1825 | 22.0 | 495 | 1.0458 | 0.6957 |
|
| 92 |
+
| 0.1701 | 22.98 | 517 | 1.3520 | 0.6087 |
|
| 93 |
+
| 0.1964 | 24.0 | 540 | 1.3927 | 0.6522 |
|
| 94 |
+
| 0.1731 | 24.98 | 562 | 1.4361 | 0.6522 |
|
| 95 |
+
| 0.1565 | 26.0 | 585 | 1.0449 | 0.6957 |
|
| 96 |
+
| 0.1844 | 26.98 | 607 | 1.3166 | 0.6087 |
|
| 97 |
+
| 0.1187 | 28.0 | 630 | 1.7950 | 0.6304 |
|
| 98 |
+
| 0.129 | 28.98 | 652 | 1.2753 | 0.6957 |
|
| 99 |
+
| 0.1269 | 30.0 | 675 | 1.4244 | 0.6739 |
|
| 100 |
+
| 0.1522 | 30.98 | 697 | 1.4873 | 0.6522 |
|
| 101 |
+
| 0.1497 | 32.0 | 720 | 1.3693 | 0.6739 |
|
| 102 |
+
| 0.1215 | 32.98 | 742 | 1.8168 | 0.6739 |
|
| 103 |
+
| 0.1049 | 34.0 | 765 | 1.2749 | 0.7609 |
|
| 104 |
+
| 0.1013 | 34.98 | 787 | 1.5098 | 0.7609 |
|
| 105 |
+
| 0.1499 | 36.0 | 810 | 1.6464 | 0.6304 |
|
| 106 |
+
| 0.0823 | 36.98 | 832 | 1.7892 | 0.6957 |
|
| 107 |
+
| 0.092 | 38.0 | 855 | 1.6448 | 0.6739 |
|
| 108 |
+
| 0.1076 | 38.98 | 877 | 1.6955 | 0.6304 |
|
| 109 |
+
| 0.1163 | 40.0 | 900 | 1.5780 | 0.6739 |
|
| 110 |
+
| 0.0952 | 40.98 | 922 | 1.8121 | 0.6522 |
|
| 111 |
+
| 0.0833 | 42.0 | 945 | 1.4459 | 0.7174 |
|
| 112 |
+
| 0.1045 | 42.98 | 967 | 1.7307 | 0.6739 |
|
| 113 |
+
| 0.094 | 44.0 | 990 | 1.4970 | 0.7391 |
|
| 114 |
+
| 0.092 | 44.98 | 1012 | 1.5766 | 0.7174 |
|
| 115 |
+
| 0.0863 | 46.0 | 1035 | 1.7600 | 0.6522 |
|
| 116 |
+
| 0.101 | 46.98 | 1057 | 1.4763 | 0.6957 |
|
| 117 |
+
| 0.0995 | 48.0 | 1080 | 2.0018 | 0.6522 |
|
| 118 |
+
| 0.0893 | 48.98 | 1102 | 1.4872 | 0.7391 |
|
| 119 |
+
| 0.0965 | 50.0 | 1125 | 1.6165 | 0.7391 |
|
| 120 |
+
| 0.0595 | 50.98 | 1147 | 1.6608 | 0.7391 |
|
| 121 |
+
| 0.0606 | 52.0 | 1170 | 1.6604 | 0.7826 |
|
| 122 |
+
| 0.0794 | 52.98 | 1192 | 1.9967 | 0.7174 |
|
| 123 |
+
| 0.0919 | 54.0 | 1215 | 1.7728 | 0.6957 |
|
| 124 |
+
| 0.0666 | 54.98 | 1237 | 1.7364 | 0.7391 |
|
| 125 |
+
| 0.0842 | 56.0 | 1260 | 1.6661 | 0.7609 |
|
| 126 |
+
| 0.0781 | 56.98 | 1282 | 1.9340 | 0.7174 |
|
| 127 |
+
| 0.0565 | 58.0 | 1305 | 1.7399 | 0.7391 |
|
| 128 |
+
| 0.0939 | 58.98 | 1327 | 1.6644 | 0.7609 |
|
| 129 |
+
| 0.0666 | 60.0 | 1350 | 1.6804 | 0.6957 |
|
| 130 |
+
| 0.0577 | 60.98 | 1372 | 1.8968 | 0.6957 |
|
| 131 |
+
| 0.0534 | 62.0 | 1395 | 1.8967 | 0.7391 |
|
| 132 |
+
| 0.0592 | 62.98 | 1417 | 2.0113 | 0.7174 |
|
| 133 |
+
| 0.0732 | 64.0 | 1440 | 1.9213 | 0.6522 |
|
| 134 |
+
| 0.0768 | 64.98 | 1462 | 1.8912 | 0.7391 |
|
| 135 |
+
| 0.0415 | 66.0 | 1485 | 1.7955 | 0.7391 |
|
| 136 |
+
| 0.036 | 66.98 | 1507 | 1.6584 | 0.7391 |
|
| 137 |
+
| 0.0617 | 68.0 | 1530 | 1.9461 | 0.7391 |
|
| 138 |
+
| 0.0622 | 68.98 | 1552 | 1.7302 | 0.7826 |
|
| 139 |
+
| 0.0362 | 70.0 | 1575 | 1.7996 | 0.7609 |
|
| 140 |
+
| 0.0526 | 70.98 | 1597 | 1.6479 | 0.7391 |
|
| 141 |
+
| 0.0493 | 72.0 | 1620 | 1.7251 | 0.6739 |
|
| 142 |
+
| 0.0703 | 72.98 | 1642 | 2.0378 | 0.7391 |
|
| 143 |
+
| 0.0692 | 74.0 | 1665 | 2.0999 | 0.6522 |
|
| 144 |
+
| 0.0396 | 74.98 | 1687 | 2.0074 | 0.6739 |
|
| 145 |
+
| 0.0505 | 76.0 | 1710 | 1.7463 | 0.7174 |
|
| 146 |
+
| 0.0512 | 76.98 | 1732 | 1.6401 | 0.7609 |
|
| 147 |
+
| 0.0653 | 78.0 | 1755 | 1.7836 | 0.6957 |
|
| 148 |
+
| 0.0764 | 78.98 | 1777 | 1.7904 | 0.7609 |
|
| 149 |
+
| 0.0598 | 80.0 | 1800 | 1.8720 | 0.7609 |
|
| 150 |
+
| 0.0523 | 80.98 | 1822 | 1.7356 | 0.7174 |
|
| 151 |
+
| 0.0499 | 82.0 | 1845 | 1.9223 | 0.7609 |
|
| 152 |
+
| 0.0841 | 82.98 | 1867 | 1.9060 | 0.7609 |
|
| 153 |
+
| 0.0597 | 84.0 | 1890 | 1.9092 | 0.7174 |
|
| 154 |
+
| 0.0696 | 84.98 | 1912 | 1.9534 | 0.6957 |
|
| 155 |
+
| 0.0566 | 86.0 | 1935 | 1.9356 | 0.7609 |
|
| 156 |
+
| 0.0435 | 86.98 | 1957 | 2.0566 | 0.7391 |
|
| 157 |
+
| 0.024 | 88.0 | 1980 | 1.8717 | 0.7174 |
|
| 158 |
+
| 0.0137 | 88.98 | 2002 | 2.0280 | 0.7174 |
|
| 159 |
+
| 0.0663 | 90.0 | 2025 | 1.8829 | 0.7174 |
|
| 160 |
+
| 0.035 | 90.98 | 2047 | 1.9688 | 0.7174 |
|
| 161 |
+
| 0.0504 | 92.0 | 2070 | 2.0456 | 0.7391 |
|
| 162 |
+
| 0.025 | 92.98 | 2092 | 2.1568 | 0.6304 |
|
| 163 |
+
| 0.0405 | 94.0 | 2115 | 2.0628 | 0.6304 |
|
| 164 |
+
| 0.0247 | 94.98 | 2137 | 2.0686 | 0.6957 |
|
| 165 |
+
| 0.0429 | 96.0 | 2160 | 2.1125 | 0.7174 |
|
| 166 |
+
| 0.0408 | 96.98 | 2182 | 2.1003 | 0.7391 |
|
| 167 |
+
| 0.0385 | 98.0 | 2205 | 2.2997 | 0.6739 |
|
| 168 |
+
| 0.0364 | 98.98 | 2227 | 2.1442 | 0.7174 |
|
| 169 |
+
| 0.0415 | 100.0 | 2250 | 2.1043 | 0.7174 |
|
| 170 |
+
| 0.0175 | 100.98 | 2272 | 2.1847 | 0.6739 |
|
| 171 |
+
| 0.0281 | 102.0 | 2295 | 2.3262 | 0.6522 |
|
| 172 |
+
| 0.0268 | 102.98 | 2317 | 2.2843 | 0.6957 |
|
| 173 |
+
| 0.022 | 104.0 | 2340 | 2.3522 | 0.6957 |
|
| 174 |
+
| 0.0279 | 104.98 | 2362 | 2.4827 | 0.6739 |
|
| 175 |
+
| 0.0188 | 106.0 | 2385 | 2.5688 | 0.6522 |
|
| 176 |
+
| 0.0303 | 106.98 | 2407 | 2.4357 | 0.6522 |
|
| 177 |
+
| 0.0439 | 108.0 | 2430 | 2.4359 | 0.6522 |
|
| 178 |
+
| 0.0422 | 108.98 | 2452 | 2.4725 | 0.6739 |
|
| 179 |
+
| 0.032 | 110.0 | 2475 | 2.2899 | 0.6522 |
|
| 180 |
+
| 0.0414 | 110.98 | 2497 | 2.2685 | 0.6957 |
|
| 181 |
+
| 0.03 | 112.0 | 2520 | 2.3063 | 0.6739 |
|
| 182 |
+
| 0.0293 | 112.98 | 2542 | 2.3524 | 0.6739 |
|
| 183 |
+
| 0.0514 | 114.0 | 2565 | 2.3612 | 0.6739 |
|
| 184 |
+
| 0.0234 | 114.98 | 2587 | 2.3711 | 0.6522 |
|
| 185 |
+
| 0.0476 | 116.0 | 2610 | 2.3548 | 0.6739 |
|
| 186 |
+
| 0.0307 | 116.98 | 2632 | 2.3432 | 0.6739 |
|
| 187 |
+
| 0.028 | 117.33 | 2640 | 2.3432 | 0.6739 |
|
| 188 |
+
|
| 189 |
+
|
| 190 |
+
### Framework versions
|
| 191 |
+
|
| 192 |
+
- Transformers 4.36.2
|
| 193 |
+
- Pytorch 2.1.2+cu118
|
| 194 |
+
- Datasets 2.16.1
|
| 195 |
+
- Tokenizers 0.15.0
|
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 117.33,
|
| 3 |
+
"eval_accuracy": 0.782608695652174,
|
| 4 |
+
"eval_loss": 1.6604318618774414,
|
| 5 |
+
"eval_runtime": 0.7062,
|
| 6 |
+
"eval_samples_per_second": 65.138,
|
| 7 |
+
"eval_steps_per_second": 4.248,
|
| 8 |
+
"train_loss": 0.1724454693552671,
|
| 9 |
+
"train_runtime": 3352.4175,
|
| 10 |
+
"train_samples_per_second": 51.259,
|
| 11 |
+
"train_steps_per_second": 0.787
|
| 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": 117.33,
|
| 3 |
+
"eval_accuracy": 0.782608695652174,
|
| 4 |
+
"eval_loss": 1.6604318618774414,
|
| 5 |
+
"eval_runtime": 0.7062,
|
| 6 |
+
"eval_samples_per_second": 65.138,
|
| 7 |
+
"eval_steps_per_second": 4.248
|
| 8 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da0954f20223f2b4e5bd2b57ba6f1859d376bca75f53c8e3216c9272f6a29979
|
| 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_18-02-31_DESKTOP-SKBE9FB/events.out.tfevents.1733443352.DESKTOP-SKBE9FB.17596.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eb8b6204e129938508ef1c22b94adb0988bc25a09f54f1cc05e923f6512efba5
|
| 3 |
+
size 84601
|
runs/Dec05_18-02-31_DESKTOP-SKBE9FB/events.out.tfevents.1733446706.DESKTOP-SKBE9FB.17596.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:df1f65dd5d22a0d169d9d5e0286212bf0958069959b887f8136d8764f06e275d
|
| 3 |
+
size 411
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 117.33,
|
| 3 |
+
"train_loss": 0.1724454693552671,
|
| 4 |
+
"train_runtime": 3352.4175,
|
| 5 |
+
"train_samples_per_second": 51.259,
|
| 6 |
+
"train_steps_per_second": 0.787
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2676 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.782608695652174,
|
| 3 |
+
"best_model_checkpoint": "Swin-dmae-DA3-N-Colab\\checkpoint-1170",
|
| 4 |
+
"epoch": 117.33333333333333,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2640,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.44,
|
| 13 |
+
"learning_rate": 1.8939393939393941e-06,
|
| 14 |
+
"loss": 1.4044,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.89,
|
| 19 |
+
"learning_rate": 3.7878787878787882e-06,
|
| 20 |
+
"loss": 1.4177,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.98,
|
| 25 |
+
"eval_accuracy": 0.43478260869565216,
|
| 26 |
+
"eval_loss": 1.3454896211624146,
|
| 27 |
+
"eval_runtime": 0.8077,
|
| 28 |
+
"eval_samples_per_second": 56.952,
|
| 29 |
+
"eval_steps_per_second": 3.714,
|
| 30 |
+
"step": 22
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 1.33,
|
| 34 |
+
"learning_rate": 5.681818181818182e-06,
|
| 35 |
+
"loss": 1.4099,
|
| 36 |
+
"step": 30
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 1.78,
|
| 40 |
+
"learning_rate": 7.5757575757575764e-06,
|
| 41 |
+
"loss": 1.4043,
|
| 42 |
+
"step": 40
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 2.0,
|
| 46 |
+
"eval_accuracy": 0.45652173913043476,
|
| 47 |
+
"eval_loss": 1.3475043773651123,
|
| 48 |
+
"eval_runtime": 0.5832,
|
| 49 |
+
"eval_samples_per_second": 78.881,
|
| 50 |
+
"eval_steps_per_second": 5.144,
|
| 51 |
+
"step": 45
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 2.22,
|
| 55 |
+
"learning_rate": 9.46969696969697e-06,
|
| 56 |
+
"loss": 1.3826,
|
| 57 |
+
"step": 50
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 2.67,
|
| 61 |
+
"learning_rate": 1.1363636363636365e-05,
|
| 62 |
+
"loss": 1.3628,
|
| 63 |
+
"step": 60
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.98,
|
| 67 |
+
"eval_accuracy": 0.45652173913043476,
|
| 68 |
+
"eval_loss": 1.339664340019226,
|
| 69 |
+
"eval_runtime": 0.6431,
|
| 70 |
+
"eval_samples_per_second": 71.523,
|
| 71 |
+
"eval_steps_per_second": 4.665,
|
| 72 |
+
"step": 67
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 3.11,
|
| 76 |
+
"learning_rate": 1.3257575757575758e-05,
|
| 77 |
+
"loss": 1.3198,
|
| 78 |
+
"step": 70
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 3.56,
|
| 82 |
+
"learning_rate": 1.5151515151515153e-05,
|
| 83 |
+
"loss": 1.2586,
|
| 84 |
+
"step": 80
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 4.0,
|
| 88 |
+
"learning_rate": 1.7045454545454546e-05,
|
| 89 |
+
"loss": 1.21,
|
| 90 |
+
"step": 90
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 4.0,
|
| 94 |
+
"eval_accuracy": 0.45652173913043476,
|
| 95 |
+
"eval_loss": 1.259264588356018,
|
| 96 |
+
"eval_runtime": 0.5817,
|
| 97 |
+
"eval_samples_per_second": 79.084,
|
| 98 |
+
"eval_steps_per_second": 5.158,
|
| 99 |
+
"step": 90
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 4.44,
|
| 103 |
+
"learning_rate": 1.893939393939394e-05,
|
| 104 |
+
"loss": 1.1047,
|
| 105 |
+
"step": 100
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 4.89,
|
| 109 |
+
"learning_rate": 2.0833333333333336e-05,
|
| 110 |
+
"loss": 1.0504,
|
| 111 |
+
"step": 110
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 4.98,
|
| 115 |
+
"eval_accuracy": 0.41304347826086957,
|
| 116 |
+
"eval_loss": 1.1193994283676147,
|
| 117 |
+
"eval_runtime": 0.5866,
|
| 118 |
+
"eval_samples_per_second": 78.413,
|
| 119 |
+
"eval_steps_per_second": 5.114,
|
| 120 |
+
"step": 112
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 5.33,
|
| 124 |
+
"learning_rate": 2.272727272727273e-05,
|
| 125 |
+
"loss": 0.975,
|
| 126 |
+
"step": 120
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 5.78,
|
| 130 |
+
"learning_rate": 2.4621212121212123e-05,
|
| 131 |
+
"loss": 0.9129,
|
| 132 |
+
"step": 130
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 6.0,
|
| 136 |
+
"eval_accuracy": 0.41304347826086957,
|
| 137 |
+
"eval_loss": 1.0521512031555176,
|
| 138 |
+
"eval_runtime": 0.5861,
|
| 139 |
+
"eval_samples_per_second": 78.48,
|
| 140 |
+
"eval_steps_per_second": 5.118,
|
| 141 |
+
"step": 135
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 6.22,
|
| 145 |
+
"learning_rate": 2.6515151515151516e-05,
|
| 146 |
+
"loss": 0.8413,
|
| 147 |
+
"step": 140
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 6.67,
|
| 151 |
+
"learning_rate": 2.8409090909090912e-05,
|
| 152 |
+
"loss": 0.7811,
|
| 153 |
+
"step": 150
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 6.98,
|
| 157 |
+
"eval_accuracy": 0.43478260869565216,
|
| 158 |
+
"eval_loss": 1.1183944940567017,
|
| 159 |
+
"eval_runtime": 0.6136,
|
| 160 |
+
"eval_samples_per_second": 74.962,
|
| 161 |
+
"eval_steps_per_second": 4.889,
|
| 162 |
+
"step": 157
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 7.11,
|
| 166 |
+
"learning_rate": 3.0303030303030306e-05,
|
| 167 |
+
"loss": 0.7098,
|
| 168 |
+
"step": 160
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 7.56,
|
| 172 |
+
"learning_rate": 3.2196969696969696e-05,
|
| 173 |
+
"loss": 0.629,
|
| 174 |
+
"step": 170
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 8.0,
|
| 178 |
+
"learning_rate": 3.409090909090909e-05,
|
| 179 |
+
"loss": 0.6572,
|
| 180 |
+
"step": 180
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 8.0,
|
| 184 |
+
"eval_accuracy": 0.5869565217391305,
|
| 185 |
+
"eval_loss": 0.995121419429779,
|
| 186 |
+
"eval_runtime": 0.6492,
|
| 187 |
+
"eval_samples_per_second": 70.86,
|
| 188 |
+
"eval_steps_per_second": 4.621,
|
| 189 |
+
"step": 180
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 8.44,
|
| 193 |
+
"learning_rate": 3.598484848484849e-05,
|
| 194 |
+
"loss": 0.6062,
|
| 195 |
+
"step": 190
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 8.89,
|
| 199 |
+
"learning_rate": 3.787878787878788e-05,
|
| 200 |
+
"loss": 0.5207,
|
| 201 |
+
"step": 200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 8.98,
|
| 205 |
+
"eval_accuracy": 0.6521739130434783,
|
| 206 |
+
"eval_loss": 0.9054760336875916,
|
| 207 |
+
"eval_runtime": 0.6757,
|
| 208 |
+
"eval_samples_per_second": 68.08,
|
| 209 |
+
"eval_steps_per_second": 4.44,
|
| 210 |
+
"step": 202
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 9.33,
|
| 214 |
+
"learning_rate": 3.9772727272727275e-05,
|
| 215 |
+
"loss": 0.5179,
|
| 216 |
+
"step": 210
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 9.78,
|
| 220 |
+
"learning_rate": 4.166666666666667e-05,
|
| 221 |
+
"loss": 0.6234,
|
| 222 |
+
"step": 220
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 10.0,
|
| 226 |
+
"eval_accuracy": 0.6086956521739131,
|
| 227 |
+
"eval_loss": 0.9276577234268188,
|
| 228 |
+
"eval_runtime": 0.8717,
|
| 229 |
+
"eval_samples_per_second": 52.769,
|
| 230 |
+
"eval_steps_per_second": 3.441,
|
| 231 |
+
"step": 225
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 10.22,
|
| 235 |
+
"learning_rate": 4.356060606060606e-05,
|
| 236 |
+
"loss": 0.5149,
|
| 237 |
+
"step": 230
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 10.67,
|
| 241 |
+
"learning_rate": 4.545454545454546e-05,
|
| 242 |
+
"loss": 0.4721,
|
| 243 |
+
"step": 240
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 10.98,
|
| 247 |
+
"eval_accuracy": 0.6739130434782609,
|
| 248 |
+
"eval_loss": 0.8458214402198792,
|
| 249 |
+
"eval_runtime": 0.6357,
|
| 250 |
+
"eval_samples_per_second": 72.366,
|
| 251 |
+
"eval_steps_per_second": 4.72,
|
| 252 |
+
"step": 247
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 11.11,
|
| 256 |
+
"learning_rate": 4.7348484848484855e-05,
|
| 257 |
+
"loss": 0.4504,
|
| 258 |
+
"step": 250
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 11.56,
|
| 262 |
+
"learning_rate": 4.9242424242424245e-05,
|
| 263 |
+
"loss": 0.4534,
|
| 264 |
+
"step": 260
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 12.0,
|
| 268 |
+
"learning_rate": 4.9873737373737375e-05,
|
| 269 |
+
"loss": 0.3944,
|
| 270 |
+
"step": 270
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 12.0,
|
| 274 |
+
"eval_accuracy": 0.6521739130434783,
|
| 275 |
+
"eval_loss": 0.8836989998817444,
|
| 276 |
+
"eval_runtime": 0.6522,
|
| 277 |
+
"eval_samples_per_second": 70.533,
|
| 278 |
+
"eval_steps_per_second": 4.6,
|
| 279 |
+
"step": 270
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 12.44,
|
| 283 |
+
"learning_rate": 4.966329966329967e-05,
|
| 284 |
+
"loss": 0.3923,
|
| 285 |
+
"step": 280
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 12.89,
|
| 289 |
+
"learning_rate": 4.945286195286196e-05,
|
| 290 |
+
"loss": 0.3572,
|
| 291 |
+
"step": 290
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 12.98,
|
| 295 |
+
"eval_accuracy": 0.717391304347826,
|
| 296 |
+
"eval_loss": 0.8719282746315002,
|
| 297 |
+
"eval_runtime": 0.6304,
|
| 298 |
+
"eval_samples_per_second": 72.972,
|
| 299 |
+
"eval_steps_per_second": 4.759,
|
| 300 |
+
"step": 292
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 13.33,
|
| 304 |
+
"learning_rate": 4.9242424242424245e-05,
|
| 305 |
+
"loss": 0.3524,
|
| 306 |
+
"step": 300
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 13.78,
|
| 310 |
+
"learning_rate": 4.903198653198653e-05,
|
| 311 |
+
"loss": 0.2911,
|
| 312 |
+
"step": 310
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 14.0,
|
| 316 |
+
"eval_accuracy": 0.6304347826086957,
|
| 317 |
+
"eval_loss": 1.036683201789856,
|
| 318 |
+
"eval_runtime": 0.6022,
|
| 319 |
+
"eval_samples_per_second": 76.392,
|
| 320 |
+
"eval_steps_per_second": 4.982,
|
| 321 |
+
"step": 315
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 14.22,
|
| 325 |
+
"learning_rate": 4.882154882154882e-05,
|
| 326 |
+
"loss": 0.3364,
|
| 327 |
+
"step": 320
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 14.67,
|
| 331 |
+
"learning_rate": 4.8611111111111115e-05,
|
| 332 |
+
"loss": 0.3224,
|
| 333 |
+
"step": 330
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 14.98,
|
| 337 |
+
"eval_accuracy": 0.6086956521739131,
|
| 338 |
+
"eval_loss": 0.9640800356864929,
|
| 339 |
+
"eval_runtime": 0.5942,
|
| 340 |
+
"eval_samples_per_second": 77.42,
|
| 341 |
+
"eval_steps_per_second": 5.049,
|
| 342 |
+
"step": 337
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 15.11,
|
| 346 |
+
"learning_rate": 4.84006734006734e-05,
|
| 347 |
+
"loss": 0.2519,
|
| 348 |
+
"step": 340
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 15.56,
|
| 352 |
+
"learning_rate": 4.819023569023569e-05,
|
| 353 |
+
"loss": 0.266,
|
| 354 |
+
"step": 350
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 16.0,
|
| 358 |
+
"learning_rate": 4.797979797979798e-05,
|
| 359 |
+
"loss": 0.2663,
|
| 360 |
+
"step": 360
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 16.0,
|
| 364 |
+
"eval_accuracy": 0.5869565217391305,
|
| 365 |
+
"eval_loss": 1.3670061826705933,
|
| 366 |
+
"eval_runtime": 0.6168,
|
| 367 |
+
"eval_samples_per_second": 74.573,
|
| 368 |
+
"eval_steps_per_second": 4.863,
|
| 369 |
+
"step": 360
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 16.44,
|
| 373 |
+
"learning_rate": 4.7769360269360266e-05,
|
| 374 |
+
"loss": 0.3245,
|
| 375 |
+
"step": 370
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 16.89,
|
| 379 |
+
"learning_rate": 4.755892255892256e-05,
|
| 380 |
+
"loss": 0.2132,
|
| 381 |
+
"step": 380
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 16.98,
|
| 385 |
+
"eval_accuracy": 0.6304347826086957,
|
| 386 |
+
"eval_loss": 1.309033751487732,
|
| 387 |
+
"eval_runtime": 0.6001,
|
| 388 |
+
"eval_samples_per_second": 76.649,
|
| 389 |
+
"eval_steps_per_second": 4.999,
|
| 390 |
+
"step": 382
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 17.33,
|
| 394 |
+
"learning_rate": 4.7348484848484855e-05,
|
| 395 |
+
"loss": 0.2687,
|
| 396 |
+
"step": 390
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 17.78,
|
| 400 |
+
"learning_rate": 4.713804713804714e-05,
|
| 401 |
+
"loss": 0.266,
|
| 402 |
+
"step": 400
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 18.0,
|
| 406 |
+
"eval_accuracy": 0.717391304347826,
|
| 407 |
+
"eval_loss": 1.1246769428253174,
|
| 408 |
+
"eval_runtime": 0.6912,
|
| 409 |
+
"eval_samples_per_second": 66.554,
|
| 410 |
+
"eval_steps_per_second": 4.34,
|
| 411 |
+
"step": 405
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 18.22,
|
| 415 |
+
"learning_rate": 4.692760942760943e-05,
|
| 416 |
+
"loss": 0.2472,
|
| 417 |
+
"step": 410
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 18.67,
|
| 421 |
+
"learning_rate": 4.671717171717172e-05,
|
| 422 |
+
"loss": 0.1929,
|
| 423 |
+
"step": 420
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 18.98,
|
| 427 |
+
"eval_accuracy": 0.6086956521739131,
|
| 428 |
+
"eval_loss": 1.1458401679992676,
|
| 429 |
+
"eval_runtime": 0.6044,
|
| 430 |
+
"eval_samples_per_second": 76.11,
|
| 431 |
+
"eval_steps_per_second": 4.964,
|
| 432 |
+
"step": 427
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 19.11,
|
| 436 |
+
"learning_rate": 4.6506734006734006e-05,
|
| 437 |
+
"loss": 0.2374,
|
| 438 |
+
"step": 430
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 19.56,
|
| 442 |
+
"learning_rate": 4.62962962962963e-05,
|
| 443 |
+
"loss": 0.2191,
|
| 444 |
+
"step": 440
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 20.0,
|
| 448 |
+
"learning_rate": 4.608585858585859e-05,
|
| 449 |
+
"loss": 0.1831,
|
| 450 |
+
"step": 450
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 20.0,
|
| 454 |
+
"eval_accuracy": 0.7391304347826086,
|
| 455 |
+
"eval_loss": 1.026720404624939,
|
| 456 |
+
"eval_runtime": 0.6282,
|
| 457 |
+
"eval_samples_per_second": 73.23,
|
| 458 |
+
"eval_steps_per_second": 4.776,
|
| 459 |
+
"step": 450
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 20.44,
|
| 463 |
+
"learning_rate": 4.5875420875420876e-05,
|
| 464 |
+
"loss": 0.196,
|
| 465 |
+
"step": 460
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 20.89,
|
| 469 |
+
"learning_rate": 4.5664983164983164e-05,
|
| 470 |
+
"loss": 0.2298,
|
| 471 |
+
"step": 470
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 20.98,
|
| 475 |
+
"eval_accuracy": 0.6304347826086957,
|
| 476 |
+
"eval_loss": 1.1862982511520386,
|
| 477 |
+
"eval_runtime": 0.6377,
|
| 478 |
+
"eval_samples_per_second": 72.135,
|
| 479 |
+
"eval_steps_per_second": 4.704,
|
| 480 |
+
"step": 472
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 21.33,
|
| 484 |
+
"learning_rate": 4.545454545454546e-05,
|
| 485 |
+
"loss": 0.1991,
|
| 486 |
+
"step": 480
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 21.78,
|
| 490 |
+
"learning_rate": 4.5244107744107746e-05,
|
| 491 |
+
"loss": 0.1825,
|
| 492 |
+
"step": 490
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 22.0,
|
| 496 |
+
"eval_accuracy": 0.6956521739130435,
|
| 497 |
+
"eval_loss": 1.0457817316055298,
|
| 498 |
+
"eval_runtime": 0.6892,
|
| 499 |
+
"eval_samples_per_second": 66.748,
|
| 500 |
+
"eval_steps_per_second": 4.353,
|
| 501 |
+
"step": 495
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 22.22,
|
| 505 |
+
"learning_rate": 4.5033670033670034e-05,
|
| 506 |
+
"loss": 0.1844,
|
| 507 |
+
"step": 500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 22.67,
|
| 511 |
+
"learning_rate": 4.482323232323233e-05,
|
| 512 |
+
"loss": 0.1701,
|
| 513 |
+
"step": 510
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 22.98,
|
| 517 |
+
"eval_accuracy": 0.6086956521739131,
|
| 518 |
+
"eval_loss": 1.3519866466522217,
|
| 519 |
+
"eval_runtime": 0.6687,
|
| 520 |
+
"eval_samples_per_second": 68.792,
|
| 521 |
+
"eval_steps_per_second": 4.486,
|
| 522 |
+
"step": 517
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 23.11,
|
| 526 |
+
"learning_rate": 4.4612794612794616e-05,
|
| 527 |
+
"loss": 0.1672,
|
| 528 |
+
"step": 520
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 23.56,
|
| 532 |
+
"learning_rate": 4.4402356902356904e-05,
|
| 533 |
+
"loss": 0.1697,
|
| 534 |
+
"step": 530
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 24.0,
|
| 538 |
+
"learning_rate": 4.41919191919192e-05,
|
| 539 |
+
"loss": 0.1964,
|
| 540 |
+
"step": 540
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 24.0,
|
| 544 |
+
"eval_accuracy": 0.6521739130434783,
|
| 545 |
+
"eval_loss": 1.3927210569381714,
|
| 546 |
+
"eval_runtime": 0.6387,
|
| 547 |
+
"eval_samples_per_second": 72.025,
|
| 548 |
+
"eval_steps_per_second": 4.697,
|
| 549 |
+
"step": 540
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 24.44,
|
| 553 |
+
"learning_rate": 4.3981481481481486e-05,
|
| 554 |
+
"loss": 0.1503,
|
| 555 |
+
"step": 550
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 24.89,
|
| 559 |
+
"learning_rate": 4.3771043771043774e-05,
|
| 560 |
+
"loss": 0.1731,
|
| 561 |
+
"step": 560
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 24.98,
|
| 565 |
+
"eval_accuracy": 0.6521739130434783,
|
| 566 |
+
"eval_loss": 1.4360977411270142,
|
| 567 |
+
"eval_runtime": 0.7162,
|
| 568 |
+
"eval_samples_per_second": 64.231,
|
| 569 |
+
"eval_steps_per_second": 4.189,
|
| 570 |
+
"step": 562
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 25.33,
|
| 574 |
+
"learning_rate": 4.356060606060606e-05,
|
| 575 |
+
"loss": 0.1654,
|
| 576 |
+
"step": 570
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 25.78,
|
| 580 |
+
"learning_rate": 4.335016835016835e-05,
|
| 581 |
+
"loss": 0.1565,
|
| 582 |
+
"step": 580
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 26.0,
|
| 586 |
+
"eval_accuracy": 0.6956521739130435,
|
| 587 |
+
"eval_loss": 1.0449482202529907,
|
| 588 |
+
"eval_runtime": 0.6361,
|
| 589 |
+
"eval_samples_per_second": 72.311,
|
| 590 |
+
"eval_steps_per_second": 4.716,
|
| 591 |
+
"step": 585
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 26.22,
|
| 595 |
+
"learning_rate": 4.3139730639730644e-05,
|
| 596 |
+
"loss": 0.1639,
|
| 597 |
+
"step": 590
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 26.67,
|
| 601 |
+
"learning_rate": 4.292929292929293e-05,
|
| 602 |
+
"loss": 0.1844,
|
| 603 |
+
"step": 600
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 26.98,
|
| 607 |
+
"eval_accuracy": 0.6086956521739131,
|
| 608 |
+
"eval_loss": 1.3166261911392212,
|
| 609 |
+
"eval_runtime": 0.6943,
|
| 610 |
+
"eval_samples_per_second": 66.252,
|
| 611 |
+
"eval_steps_per_second": 4.321,
|
| 612 |
+
"step": 607
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 27.11,
|
| 616 |
+
"learning_rate": 4.271885521885522e-05,
|
| 617 |
+
"loss": 0.1478,
|
| 618 |
+
"step": 610
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 27.56,
|
| 622 |
+
"learning_rate": 4.250841750841751e-05,
|
| 623 |
+
"loss": 0.1709,
|
| 624 |
+
"step": 620
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 28.0,
|
| 628 |
+
"learning_rate": 4.2297979797979795e-05,
|
| 629 |
+
"loss": 0.1187,
|
| 630 |
+
"step": 630
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 28.0,
|
| 634 |
+
"eval_accuracy": 0.6304347826086957,
|
| 635 |
+
"eval_loss": 1.7950323820114136,
|
| 636 |
+
"eval_runtime": 0.7205,
|
| 637 |
+
"eval_samples_per_second": 63.842,
|
| 638 |
+
"eval_steps_per_second": 4.164,
|
| 639 |
+
"step": 630
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 28.44,
|
| 643 |
+
"learning_rate": 4.208754208754209e-05,
|
| 644 |
+
"loss": 0.1371,
|
| 645 |
+
"step": 640
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 28.89,
|
| 649 |
+
"learning_rate": 4.1877104377104384e-05,
|
| 650 |
+
"loss": 0.129,
|
| 651 |
+
"step": 650
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 28.98,
|
| 655 |
+
"eval_accuracy": 0.6956521739130435,
|
| 656 |
+
"eval_loss": 1.2752825021743774,
|
| 657 |
+
"eval_runtime": 0.6192,
|
| 658 |
+
"eval_samples_per_second": 74.293,
|
| 659 |
+
"eval_steps_per_second": 4.845,
|
| 660 |
+
"step": 652
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 29.33,
|
| 664 |
+
"learning_rate": 4.166666666666667e-05,
|
| 665 |
+
"loss": 0.1172,
|
| 666 |
+
"step": 660
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 29.78,
|
| 670 |
+
"learning_rate": 4.145622895622896e-05,
|
| 671 |
+
"loss": 0.1269,
|
| 672 |
+
"step": 670
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 30.0,
|
| 676 |
+
"eval_accuracy": 0.6739130434782609,
|
| 677 |
+
"eval_loss": 1.4243661165237427,
|
| 678 |
+
"eval_runtime": 0.6637,
|
| 679 |
+
"eval_samples_per_second": 69.309,
|
| 680 |
+
"eval_steps_per_second": 4.52,
|
| 681 |
+
"step": 675
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 30.22,
|
| 685 |
+
"learning_rate": 4.124579124579125e-05,
|
| 686 |
+
"loss": 0.1689,
|
| 687 |
+
"step": 680
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 30.67,
|
| 691 |
+
"learning_rate": 4.1035353535353535e-05,
|
| 692 |
+
"loss": 0.1522,
|
| 693 |
+
"step": 690
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 30.98,
|
| 697 |
+
"eval_accuracy": 0.6521739130434783,
|
| 698 |
+
"eval_loss": 1.4873253107070923,
|
| 699 |
+
"eval_runtime": 0.6497,
|
| 700 |
+
"eval_samples_per_second": 70.807,
|
| 701 |
+
"eval_steps_per_second": 4.618,
|
| 702 |
+
"step": 697
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 31.11,
|
| 706 |
+
"learning_rate": 4.082491582491583e-05,
|
| 707 |
+
"loss": 0.1308,
|
| 708 |
+
"step": 700
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 31.56,
|
| 712 |
+
"learning_rate": 4.061447811447812e-05,
|
| 713 |
+
"loss": 0.1535,
|
| 714 |
+
"step": 710
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 32.0,
|
| 718 |
+
"learning_rate": 4.0404040404040405e-05,
|
| 719 |
+
"loss": 0.1497,
|
| 720 |
+
"step": 720
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 32.0,
|
| 724 |
+
"eval_accuracy": 0.6739130434782609,
|
| 725 |
+
"eval_loss": 1.3692686557769775,
|
| 726 |
+
"eval_runtime": 0.8302,
|
| 727 |
+
"eval_samples_per_second": 55.406,
|
| 728 |
+
"eval_steps_per_second": 3.613,
|
| 729 |
+
"step": 720
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 32.44,
|
| 733 |
+
"learning_rate": 4.019360269360269e-05,
|
| 734 |
+
"loss": 0.122,
|
| 735 |
+
"step": 730
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 32.89,
|
| 739 |
+
"learning_rate": 3.998316498316498e-05,
|
| 740 |
+
"loss": 0.1215,
|
| 741 |
+
"step": 740
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 32.98,
|
| 745 |
+
"eval_accuracy": 0.6739130434782609,
|
| 746 |
+
"eval_loss": 1.8168487548828125,
|
| 747 |
+
"eval_runtime": 0.6432,
|
| 748 |
+
"eval_samples_per_second": 71.52,
|
| 749 |
+
"eval_steps_per_second": 4.664,
|
| 750 |
+
"step": 742
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 33.33,
|
| 754 |
+
"learning_rate": 3.9772727272727275e-05,
|
| 755 |
+
"loss": 0.1429,
|
| 756 |
+
"step": 750
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 33.78,
|
| 760 |
+
"learning_rate": 3.956228956228956e-05,
|
| 761 |
+
"loss": 0.1049,
|
| 762 |
+
"step": 760
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 34.0,
|
| 766 |
+
"eval_accuracy": 0.7608695652173914,
|
| 767 |
+
"eval_loss": 1.2748662233352661,
|
| 768 |
+
"eval_runtime": 0.6487,
|
| 769 |
+
"eval_samples_per_second": 70.915,
|
| 770 |
+
"eval_steps_per_second": 4.625,
|
| 771 |
+
"step": 765
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 34.22,
|
| 775 |
+
"learning_rate": 3.935185185185186e-05,
|
| 776 |
+
"loss": 0.0886,
|
| 777 |
+
"step": 770
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 34.67,
|
| 781 |
+
"learning_rate": 3.9141414141414145e-05,
|
| 782 |
+
"loss": 0.1013,
|
| 783 |
+
"step": 780
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 34.98,
|
| 787 |
+
"eval_accuracy": 0.7608695652173914,
|
| 788 |
+
"eval_loss": 1.5097851753234863,
|
| 789 |
+
"eval_runtime": 0.6176,
|
| 790 |
+
"eval_samples_per_second": 74.477,
|
| 791 |
+
"eval_steps_per_second": 4.857,
|
| 792 |
+
"step": 787
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 35.11,
|
| 796 |
+
"learning_rate": 3.893097643097643e-05,
|
| 797 |
+
"loss": 0.0993,
|
| 798 |
+
"step": 790
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 35.56,
|
| 802 |
+
"learning_rate": 3.872053872053872e-05,
|
| 803 |
+
"loss": 0.1178,
|
| 804 |
+
"step": 800
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 36.0,
|
| 808 |
+
"learning_rate": 3.8510101010101015e-05,
|
| 809 |
+
"loss": 0.1499,
|
| 810 |
+
"step": 810
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 36.0,
|
| 814 |
+
"eval_accuracy": 0.6304347826086957,
|
| 815 |
+
"eval_loss": 1.6463576555252075,
|
| 816 |
+
"eval_runtime": 0.9393,
|
| 817 |
+
"eval_samples_per_second": 48.974,
|
| 818 |
+
"eval_steps_per_second": 3.194,
|
| 819 |
+
"step": 810
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 36.44,
|
| 823 |
+
"learning_rate": 3.82996632996633e-05,
|
| 824 |
+
"loss": 0.132,
|
| 825 |
+
"step": 820
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 36.89,
|
| 829 |
+
"learning_rate": 3.808922558922559e-05,
|
| 830 |
+
"loss": 0.0823,
|
| 831 |
+
"step": 830
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 36.98,
|
| 835 |
+
"eval_accuracy": 0.6956521739130435,
|
| 836 |
+
"eval_loss": 1.7892482280731201,
|
| 837 |
+
"eval_runtime": 0.6542,
|
| 838 |
+
"eval_samples_per_second": 70.32,
|
| 839 |
+
"eval_steps_per_second": 4.586,
|
| 840 |
+
"step": 832
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 37.33,
|
| 844 |
+
"learning_rate": 3.787878787878788e-05,
|
| 845 |
+
"loss": 0.1058,
|
| 846 |
+
"step": 840
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 37.78,
|
| 850 |
+
"learning_rate": 3.7668350168350167e-05,
|
| 851 |
+
"loss": 0.092,
|
| 852 |
+
"step": 850
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 38.0,
|
| 856 |
+
"eval_accuracy": 0.6739130434782609,
|
| 857 |
+
"eval_loss": 1.6448333263397217,
|
| 858 |
+
"eval_runtime": 0.7064,
|
| 859 |
+
"eval_samples_per_second": 65.122,
|
| 860 |
+
"eval_steps_per_second": 4.247,
|
| 861 |
+
"step": 855
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 38.22,
|
| 865 |
+
"learning_rate": 3.745791245791246e-05,
|
| 866 |
+
"loss": 0.0832,
|
| 867 |
+
"step": 860
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 38.67,
|
| 871 |
+
"learning_rate": 3.724747474747475e-05,
|
| 872 |
+
"loss": 0.1076,
|
| 873 |
+
"step": 870
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 38.98,
|
| 877 |
+
"eval_accuracy": 0.6304347826086957,
|
| 878 |
+
"eval_loss": 1.6955077648162842,
|
| 879 |
+
"eval_runtime": 0.6982,
|
| 880 |
+
"eval_samples_per_second": 65.886,
|
| 881 |
+
"eval_steps_per_second": 4.297,
|
| 882 |
+
"step": 877
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 39.11,
|
| 886 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 887 |
+
"loss": 0.1159,
|
| 888 |
+
"step": 880
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 39.56,
|
| 892 |
+
"learning_rate": 3.6826599326599324e-05,
|
| 893 |
+
"loss": 0.1204,
|
| 894 |
+
"step": 890
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 40.0,
|
| 898 |
+
"learning_rate": 3.661616161616162e-05,
|
| 899 |
+
"loss": 0.1163,
|
| 900 |
+
"step": 900
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 40.0,
|
| 904 |
+
"eval_accuracy": 0.6739130434782609,
|
| 905 |
+
"eval_loss": 1.5779627561569214,
|
| 906 |
+
"eval_runtime": 0.6312,
|
| 907 |
+
"eval_samples_per_second": 72.882,
|
| 908 |
+
"eval_steps_per_second": 4.753,
|
| 909 |
+
"step": 900
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 40.44,
|
| 913 |
+
"learning_rate": 3.6405723905723907e-05,
|
| 914 |
+
"loss": 0.1122,
|
| 915 |
+
"step": 910
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 40.89,
|
| 919 |
+
"learning_rate": 3.61952861952862e-05,
|
| 920 |
+
"loss": 0.0952,
|
| 921 |
+
"step": 920
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 40.98,
|
| 925 |
+
"eval_accuracy": 0.6521739130434783,
|
| 926 |
+
"eval_loss": 1.8121063709259033,
|
| 927 |
+
"eval_runtime": 0.7017,
|
| 928 |
+
"eval_samples_per_second": 65.557,
|
| 929 |
+
"eval_steps_per_second": 4.275,
|
| 930 |
+
"step": 922
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 41.33,
|
| 934 |
+
"learning_rate": 3.598484848484849e-05,
|
| 935 |
+
"loss": 0.0946,
|
| 936 |
+
"step": 930
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 41.78,
|
| 940 |
+
"learning_rate": 3.577441077441078e-05,
|
| 941 |
+
"loss": 0.0833,
|
| 942 |
+
"step": 940
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 42.0,
|
| 946 |
+
"eval_accuracy": 0.717391304347826,
|
| 947 |
+
"eval_loss": 1.4458924531936646,
|
| 948 |
+
"eval_runtime": 0.6203,
|
| 949 |
+
"eval_samples_per_second": 74.161,
|
| 950 |
+
"eval_steps_per_second": 4.837,
|
| 951 |
+
"step": 945
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 42.22,
|
| 955 |
+
"learning_rate": 3.5563973063973064e-05,
|
| 956 |
+
"loss": 0.0783,
|
| 957 |
+
"step": 950
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 42.67,
|
| 961 |
+
"learning_rate": 3.535353535353535e-05,
|
| 962 |
+
"loss": 0.1045,
|
| 963 |
+
"step": 960
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 42.98,
|
| 967 |
+
"eval_accuracy": 0.6739130434782609,
|
| 968 |
+
"eval_loss": 1.7306798696517944,
|
| 969 |
+
"eval_runtime": 0.6341,
|
| 970 |
+
"eval_samples_per_second": 72.539,
|
| 971 |
+
"eval_steps_per_second": 4.731,
|
| 972 |
+
"step": 967
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 43.11,
|
| 976 |
+
"learning_rate": 3.514309764309765e-05,
|
| 977 |
+
"loss": 0.0713,
|
| 978 |
+
"step": 970
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 43.56,
|
| 982 |
+
"learning_rate": 3.4932659932659934e-05,
|
| 983 |
+
"loss": 0.1254,
|
| 984 |
+
"step": 980
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 44.0,
|
| 988 |
+
"learning_rate": 3.472222222222222e-05,
|
| 989 |
+
"loss": 0.094,
|
| 990 |
+
"step": 990
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 44.0,
|
| 994 |
+
"eval_accuracy": 0.7391304347826086,
|
| 995 |
+
"eval_loss": 1.4969877004623413,
|
| 996 |
+
"eval_runtime": 0.6602,
|
| 997 |
+
"eval_samples_per_second": 69.678,
|
| 998 |
+
"eval_steps_per_second": 4.544,
|
| 999 |
+
"step": 990
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 44.44,
|
| 1003 |
+
"learning_rate": 3.451178451178451e-05,
|
| 1004 |
+
"loss": 0.108,
|
| 1005 |
+
"step": 1000
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 44.89,
|
| 1009 |
+
"learning_rate": 3.43013468013468e-05,
|
| 1010 |
+
"loss": 0.092,
|
| 1011 |
+
"step": 1010
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 44.98,
|
| 1015 |
+
"eval_accuracy": 0.717391304347826,
|
| 1016 |
+
"eval_loss": 1.5766059160232544,
|
| 1017 |
+
"eval_runtime": 0.6201,
|
| 1018 |
+
"eval_samples_per_second": 74.177,
|
| 1019 |
+
"eval_steps_per_second": 4.838,
|
| 1020 |
+
"step": 1012
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 45.33,
|
| 1024 |
+
"learning_rate": 3.409090909090909e-05,
|
| 1025 |
+
"loss": 0.1059,
|
| 1026 |
+
"step": 1020
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 45.78,
|
| 1030 |
+
"learning_rate": 3.388047138047139e-05,
|
| 1031 |
+
"loss": 0.0863,
|
| 1032 |
+
"step": 1030
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 46.0,
|
| 1036 |
+
"eval_accuracy": 0.6521739130434783,
|
| 1037 |
+
"eval_loss": 1.7600295543670654,
|
| 1038 |
+
"eval_runtime": 0.7127,
|
| 1039 |
+
"eval_samples_per_second": 64.546,
|
| 1040 |
+
"eval_steps_per_second": 4.21,
|
| 1041 |
+
"step": 1035
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 46.22,
|
| 1045 |
+
"learning_rate": 3.3670033670033675e-05,
|
| 1046 |
+
"loss": 0.106,
|
| 1047 |
+
"step": 1040
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 46.67,
|
| 1051 |
+
"learning_rate": 3.345959595959596e-05,
|
| 1052 |
+
"loss": 0.101,
|
| 1053 |
+
"step": 1050
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 46.98,
|
| 1057 |
+
"eval_accuracy": 0.6956521739130435,
|
| 1058 |
+
"eval_loss": 1.4763251543045044,
|
| 1059 |
+
"eval_runtime": 0.6232,
|
| 1060 |
+
"eval_samples_per_second": 73.817,
|
| 1061 |
+
"eval_steps_per_second": 4.814,
|
| 1062 |
+
"step": 1057
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 47.11,
|
| 1066 |
+
"learning_rate": 3.324915824915825e-05,
|
| 1067 |
+
"loss": 0.0873,
|
| 1068 |
+
"step": 1060
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 47.56,
|
| 1072 |
+
"learning_rate": 3.3038720538720545e-05,
|
| 1073 |
+
"loss": 0.0731,
|
| 1074 |
+
"step": 1070
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 48.0,
|
| 1078 |
+
"learning_rate": 3.282828282828283e-05,
|
| 1079 |
+
"loss": 0.0995,
|
| 1080 |
+
"step": 1080
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 48.0,
|
| 1084 |
+
"eval_accuracy": 0.6521739130434783,
|
| 1085 |
+
"eval_loss": 2.0017707347869873,
|
| 1086 |
+
"eval_runtime": 0.8091,
|
| 1087 |
+
"eval_samples_per_second": 56.852,
|
| 1088 |
+
"eval_steps_per_second": 3.708,
|
| 1089 |
+
"step": 1080
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 48.44,
|
| 1093 |
+
"learning_rate": 3.261784511784512e-05,
|
| 1094 |
+
"loss": 0.1169,
|
| 1095 |
+
"step": 1090
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 48.89,
|
| 1099 |
+
"learning_rate": 3.240740740740741e-05,
|
| 1100 |
+
"loss": 0.0893,
|
| 1101 |
+
"step": 1100
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 48.98,
|
| 1105 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1106 |
+
"eval_loss": 1.4871541261672974,
|
| 1107 |
+
"eval_runtime": 0.6386,
|
| 1108 |
+
"eval_samples_per_second": 72.027,
|
| 1109 |
+
"eval_steps_per_second": 4.697,
|
| 1110 |
+
"step": 1102
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 49.33,
|
| 1114 |
+
"learning_rate": 3.2196969696969696e-05,
|
| 1115 |
+
"loss": 0.0607,
|
| 1116 |
+
"step": 1110
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 49.78,
|
| 1120 |
+
"learning_rate": 3.198653198653199e-05,
|
| 1121 |
+
"loss": 0.0965,
|
| 1122 |
+
"step": 1120
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 50.0,
|
| 1126 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1127 |
+
"eval_loss": 1.6165415048599243,
|
| 1128 |
+
"eval_runtime": 0.6157,
|
| 1129 |
+
"eval_samples_per_second": 74.713,
|
| 1130 |
+
"eval_steps_per_second": 4.873,
|
| 1131 |
+
"step": 1125
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 50.22,
|
| 1135 |
+
"learning_rate": 3.177609427609428e-05,
|
| 1136 |
+
"loss": 0.079,
|
| 1137 |
+
"step": 1130
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 50.67,
|
| 1141 |
+
"learning_rate": 3.1565656565656566e-05,
|
| 1142 |
+
"loss": 0.0595,
|
| 1143 |
+
"step": 1140
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 50.98,
|
| 1147 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1148 |
+
"eval_loss": 1.66079580783844,
|
| 1149 |
+
"eval_runtime": 0.6159,
|
| 1150 |
+
"eval_samples_per_second": 74.687,
|
| 1151 |
+
"eval_steps_per_second": 4.871,
|
| 1152 |
+
"step": 1147
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 51.11,
|
| 1156 |
+
"learning_rate": 3.135521885521885e-05,
|
| 1157 |
+
"loss": 0.0755,
|
| 1158 |
+
"step": 1150
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 51.56,
|
| 1162 |
+
"learning_rate": 3.114478114478115e-05,
|
| 1163 |
+
"loss": 0.0702,
|
| 1164 |
+
"step": 1160
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 52.0,
|
| 1168 |
+
"learning_rate": 3.0934343434343436e-05,
|
| 1169 |
+
"loss": 0.0606,
|
| 1170 |
+
"step": 1170
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 52.0,
|
| 1174 |
+
"eval_accuracy": 0.782608695652174,
|
| 1175 |
+
"eval_loss": 1.6604318618774414,
|
| 1176 |
+
"eval_runtime": 0.6672,
|
| 1177 |
+
"eval_samples_per_second": 68.949,
|
| 1178 |
+
"eval_steps_per_second": 4.497,
|
| 1179 |
+
"step": 1170
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 52.44,
|
| 1183 |
+
"learning_rate": 3.072390572390573e-05,
|
| 1184 |
+
"loss": 0.0722,
|
| 1185 |
+
"step": 1180
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 52.89,
|
| 1189 |
+
"learning_rate": 3.0513468013468015e-05,
|
| 1190 |
+
"loss": 0.0794,
|
| 1191 |
+
"step": 1190
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 52.98,
|
| 1195 |
+
"eval_accuracy": 0.717391304347826,
|
| 1196 |
+
"eval_loss": 1.9967399835586548,
|
| 1197 |
+
"eval_runtime": 0.6242,
|
| 1198 |
+
"eval_samples_per_second": 73.688,
|
| 1199 |
+
"eval_steps_per_second": 4.806,
|
| 1200 |
+
"step": 1192
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 53.33,
|
| 1204 |
+
"learning_rate": 3.0303030303030306e-05,
|
| 1205 |
+
"loss": 0.0731,
|
| 1206 |
+
"step": 1200
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 53.78,
|
| 1210 |
+
"learning_rate": 3.0092592592592593e-05,
|
| 1211 |
+
"loss": 0.0919,
|
| 1212 |
+
"step": 1210
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 54.0,
|
| 1216 |
+
"eval_accuracy": 0.6956521739130435,
|
| 1217 |
+
"eval_loss": 1.7727715969085693,
|
| 1218 |
+
"eval_runtime": 0.6228,
|
| 1219 |
+
"eval_samples_per_second": 73.864,
|
| 1220 |
+
"eval_steps_per_second": 4.817,
|
| 1221 |
+
"step": 1215
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 54.22,
|
| 1225 |
+
"learning_rate": 2.9882154882154885e-05,
|
| 1226 |
+
"loss": 0.0776,
|
| 1227 |
+
"step": 1220
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 54.67,
|
| 1231 |
+
"learning_rate": 2.9671717171717172e-05,
|
| 1232 |
+
"loss": 0.0666,
|
| 1233 |
+
"step": 1230
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 54.98,
|
| 1237 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1238 |
+
"eval_loss": 1.7363907098770142,
|
| 1239 |
+
"eval_runtime": 0.6531,
|
| 1240 |
+
"eval_samples_per_second": 70.43,
|
| 1241 |
+
"eval_steps_per_second": 4.593,
|
| 1242 |
+
"step": 1237
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 55.11,
|
| 1246 |
+
"learning_rate": 2.946127946127946e-05,
|
| 1247 |
+
"loss": 0.0663,
|
| 1248 |
+
"step": 1240
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 55.56,
|
| 1252 |
+
"learning_rate": 2.925084175084175e-05,
|
| 1253 |
+
"loss": 0.0574,
|
| 1254 |
+
"step": 1250
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 56.0,
|
| 1258 |
+
"learning_rate": 2.904040404040404e-05,
|
| 1259 |
+
"loss": 0.0842,
|
| 1260 |
+
"step": 1260
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 56.0,
|
| 1264 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1265 |
+
"eval_loss": 1.6661043167114258,
|
| 1266 |
+
"eval_runtime": 0.6312,
|
| 1267 |
+
"eval_samples_per_second": 72.88,
|
| 1268 |
+
"eval_steps_per_second": 4.753,
|
| 1269 |
+
"step": 1260
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 56.44,
|
| 1273 |
+
"learning_rate": 2.882996632996633e-05,
|
| 1274 |
+
"loss": 0.0545,
|
| 1275 |
+
"step": 1270
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 56.89,
|
| 1279 |
+
"learning_rate": 2.8619528619528618e-05,
|
| 1280 |
+
"loss": 0.0781,
|
| 1281 |
+
"step": 1280
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 56.98,
|
| 1285 |
+
"eval_accuracy": 0.717391304347826,
|
| 1286 |
+
"eval_loss": 1.9340128898620605,
|
| 1287 |
+
"eval_runtime": 0.9102,
|
| 1288 |
+
"eval_samples_per_second": 50.537,
|
| 1289 |
+
"eval_steps_per_second": 3.296,
|
| 1290 |
+
"step": 1282
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 57.33,
|
| 1294 |
+
"learning_rate": 2.8409090909090912e-05,
|
| 1295 |
+
"loss": 0.0455,
|
| 1296 |
+
"step": 1290
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 57.78,
|
| 1300 |
+
"learning_rate": 2.8198653198653204e-05,
|
| 1301 |
+
"loss": 0.0565,
|
| 1302 |
+
"step": 1300
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 58.0,
|
| 1306 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1307 |
+
"eval_loss": 1.7399216890335083,
|
| 1308 |
+
"eval_runtime": 0.6902,
|
| 1309 |
+
"eval_samples_per_second": 66.647,
|
| 1310 |
+
"eval_steps_per_second": 4.347,
|
| 1311 |
+
"step": 1305
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 58.22,
|
| 1315 |
+
"learning_rate": 2.798821548821549e-05,
|
| 1316 |
+
"loss": 0.0867,
|
| 1317 |
+
"step": 1310
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 58.67,
|
| 1321 |
+
"learning_rate": 2.777777777777778e-05,
|
| 1322 |
+
"loss": 0.0939,
|
| 1323 |
+
"step": 1320
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 58.98,
|
| 1327 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1328 |
+
"eval_loss": 1.664444088935852,
|
| 1329 |
+
"eval_runtime": 0.6351,
|
| 1330 |
+
"eval_samples_per_second": 72.424,
|
| 1331 |
+
"eval_steps_per_second": 4.723,
|
| 1332 |
+
"step": 1327
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 59.11,
|
| 1336 |
+
"learning_rate": 2.756734006734007e-05,
|
| 1337 |
+
"loss": 0.0603,
|
| 1338 |
+
"step": 1330
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 59.56,
|
| 1342 |
+
"learning_rate": 2.7356902356902358e-05,
|
| 1343 |
+
"loss": 0.0605,
|
| 1344 |
+
"step": 1340
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 60.0,
|
| 1348 |
+
"learning_rate": 2.714646464646465e-05,
|
| 1349 |
+
"loss": 0.0666,
|
| 1350 |
+
"step": 1350
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 60.0,
|
| 1354 |
+
"eval_accuracy": 0.6956521739130435,
|
| 1355 |
+
"eval_loss": 1.6804474592208862,
|
| 1356 |
+
"eval_runtime": 0.6187,
|
| 1357 |
+
"eval_samples_per_second": 74.347,
|
| 1358 |
+
"eval_steps_per_second": 4.849,
|
| 1359 |
+
"step": 1350
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 60.44,
|
| 1363 |
+
"learning_rate": 2.6936026936026937e-05,
|
| 1364 |
+
"loss": 0.045,
|
| 1365 |
+
"step": 1360
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 60.89,
|
| 1369 |
+
"learning_rate": 2.6725589225589225e-05,
|
| 1370 |
+
"loss": 0.0577,
|
| 1371 |
+
"step": 1370
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 60.98,
|
| 1375 |
+
"eval_accuracy": 0.6956521739130435,
|
| 1376 |
+
"eval_loss": 1.8968379497528076,
|
| 1377 |
+
"eval_runtime": 0.6547,
|
| 1378 |
+
"eval_samples_per_second": 70.263,
|
| 1379 |
+
"eval_steps_per_second": 4.582,
|
| 1380 |
+
"step": 1372
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 61.33,
|
| 1384 |
+
"learning_rate": 2.6515151515151516e-05,
|
| 1385 |
+
"loss": 0.0895,
|
| 1386 |
+
"step": 1380
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 61.78,
|
| 1390 |
+
"learning_rate": 2.6304713804713804e-05,
|
| 1391 |
+
"loss": 0.0534,
|
| 1392 |
+
"step": 1390
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 62.0,
|
| 1396 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1397 |
+
"eval_loss": 1.8967163562774658,
|
| 1398 |
+
"eval_runtime": 0.7952,
|
| 1399 |
+
"eval_samples_per_second": 57.844,
|
| 1400 |
+
"eval_steps_per_second": 3.772,
|
| 1401 |
+
"step": 1395
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 62.22,
|
| 1405 |
+
"learning_rate": 2.6094276094276095e-05,
|
| 1406 |
+
"loss": 0.1,
|
| 1407 |
+
"step": 1400
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 62.67,
|
| 1411 |
+
"learning_rate": 2.5883838383838382e-05,
|
| 1412 |
+
"loss": 0.0592,
|
| 1413 |
+
"step": 1410
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 62.98,
|
| 1417 |
+
"eval_accuracy": 0.717391304347826,
|
| 1418 |
+
"eval_loss": 2.0113070011138916,
|
| 1419 |
+
"eval_runtime": 0.6402,
|
| 1420 |
+
"eval_samples_per_second": 71.858,
|
| 1421 |
+
"eval_steps_per_second": 4.686,
|
| 1422 |
+
"step": 1417
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 63.11,
|
| 1426 |
+
"learning_rate": 2.5673400673400677e-05,
|
| 1427 |
+
"loss": 0.0862,
|
| 1428 |
+
"step": 1420
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 63.56,
|
| 1432 |
+
"learning_rate": 2.5462962962962965e-05,
|
| 1433 |
+
"loss": 0.0761,
|
| 1434 |
+
"step": 1430
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 64.0,
|
| 1438 |
+
"learning_rate": 2.5252525252525256e-05,
|
| 1439 |
+
"loss": 0.0732,
|
| 1440 |
+
"step": 1440
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 64.0,
|
| 1444 |
+
"eval_accuracy": 0.6521739130434783,
|
| 1445 |
+
"eval_loss": 1.9213429689407349,
|
| 1446 |
+
"eval_runtime": 0.6532,
|
| 1447 |
+
"eval_samples_per_second": 70.426,
|
| 1448 |
+
"eval_steps_per_second": 4.593,
|
| 1449 |
+
"step": 1440
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 64.44,
|
| 1453 |
+
"learning_rate": 2.5042087542087544e-05,
|
| 1454 |
+
"loss": 0.0412,
|
| 1455 |
+
"step": 1450
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 64.89,
|
| 1459 |
+
"learning_rate": 2.4831649831649835e-05,
|
| 1460 |
+
"loss": 0.0768,
|
| 1461 |
+
"step": 1460
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 64.98,
|
| 1465 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1466 |
+
"eval_loss": 1.8911714553833008,
|
| 1467 |
+
"eval_runtime": 0.6397,
|
| 1468 |
+
"eval_samples_per_second": 71.914,
|
| 1469 |
+
"eval_steps_per_second": 4.69,
|
| 1470 |
+
"step": 1462
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 65.33,
|
| 1474 |
+
"learning_rate": 2.4621212121212123e-05,
|
| 1475 |
+
"loss": 0.0724,
|
| 1476 |
+
"step": 1470
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 65.78,
|
| 1480 |
+
"learning_rate": 2.441077441077441e-05,
|
| 1481 |
+
"loss": 0.0415,
|
| 1482 |
+
"step": 1480
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 66.0,
|
| 1486 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1487 |
+
"eval_loss": 1.7955325841903687,
|
| 1488 |
+
"eval_runtime": 0.6727,
|
| 1489 |
+
"eval_samples_per_second": 68.382,
|
| 1490 |
+
"eval_steps_per_second": 4.46,
|
| 1491 |
+
"step": 1485
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 66.22,
|
| 1495 |
+
"learning_rate": 2.42003367003367e-05,
|
| 1496 |
+
"loss": 0.0646,
|
| 1497 |
+
"step": 1490
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 66.67,
|
| 1501 |
+
"learning_rate": 2.398989898989899e-05,
|
| 1502 |
+
"loss": 0.036,
|
| 1503 |
+
"step": 1500
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 66.98,
|
| 1507 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1508 |
+
"eval_loss": 1.6583586931228638,
|
| 1509 |
+
"eval_runtime": 0.7332,
|
| 1510 |
+
"eval_samples_per_second": 62.739,
|
| 1511 |
+
"eval_steps_per_second": 4.092,
|
| 1512 |
+
"step": 1507
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 67.11,
|
| 1516 |
+
"learning_rate": 2.377946127946128e-05,
|
| 1517 |
+
"loss": 0.0679,
|
| 1518 |
+
"step": 1510
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 67.56,
|
| 1522 |
+
"learning_rate": 2.356902356902357e-05,
|
| 1523 |
+
"loss": 0.0601,
|
| 1524 |
+
"step": 1520
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 68.0,
|
| 1528 |
+
"learning_rate": 2.335858585858586e-05,
|
| 1529 |
+
"loss": 0.0617,
|
| 1530 |
+
"step": 1530
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 68.0,
|
| 1534 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1535 |
+
"eval_loss": 1.9461021423339844,
|
| 1536 |
+
"eval_runtime": 0.6382,
|
| 1537 |
+
"eval_samples_per_second": 72.082,
|
| 1538 |
+
"eval_steps_per_second": 4.701,
|
| 1539 |
+
"step": 1530
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 68.44,
|
| 1543 |
+
"learning_rate": 2.314814814814815e-05,
|
| 1544 |
+
"loss": 0.0743,
|
| 1545 |
+
"step": 1540
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 68.89,
|
| 1549 |
+
"learning_rate": 2.2937710437710438e-05,
|
| 1550 |
+
"loss": 0.0622,
|
| 1551 |
+
"step": 1550
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 68.98,
|
| 1555 |
+
"eval_accuracy": 0.782608695652174,
|
| 1556 |
+
"eval_loss": 1.7301748991012573,
|
| 1557 |
+
"eval_runtime": 0.6432,
|
| 1558 |
+
"eval_samples_per_second": 71.522,
|
| 1559 |
+
"eval_steps_per_second": 4.664,
|
| 1560 |
+
"step": 1552
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 69.33,
|
| 1564 |
+
"learning_rate": 2.272727272727273e-05,
|
| 1565 |
+
"loss": 0.0592,
|
| 1566 |
+
"step": 1560
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 69.78,
|
| 1570 |
+
"learning_rate": 2.2516835016835017e-05,
|
| 1571 |
+
"loss": 0.0362,
|
| 1572 |
+
"step": 1570
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 70.0,
|
| 1576 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1577 |
+
"eval_loss": 1.7996348142623901,
|
| 1578 |
+
"eval_runtime": 0.6392,
|
| 1579 |
+
"eval_samples_per_second": 71.968,
|
| 1580 |
+
"eval_steps_per_second": 4.694,
|
| 1581 |
+
"step": 1575
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 70.22,
|
| 1585 |
+
"learning_rate": 2.2306397306397308e-05,
|
| 1586 |
+
"loss": 0.0514,
|
| 1587 |
+
"step": 1580
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 70.67,
|
| 1591 |
+
"learning_rate": 2.20959595959596e-05,
|
| 1592 |
+
"loss": 0.0526,
|
| 1593 |
+
"step": 1590
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 70.98,
|
| 1597 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1598 |
+
"eval_loss": 1.6478713750839233,
|
| 1599 |
+
"eval_runtime": 0.6542,
|
| 1600 |
+
"eval_samples_per_second": 70.32,
|
| 1601 |
+
"eval_steps_per_second": 4.586,
|
| 1602 |
+
"step": 1597
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 71.11,
|
| 1606 |
+
"learning_rate": 2.1885521885521887e-05,
|
| 1607 |
+
"loss": 0.0944,
|
| 1608 |
+
"step": 1600
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 71.56,
|
| 1612 |
+
"learning_rate": 2.1675084175084175e-05,
|
| 1613 |
+
"loss": 0.0458,
|
| 1614 |
+
"step": 1610
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 72.0,
|
| 1618 |
+
"learning_rate": 2.1464646464646466e-05,
|
| 1619 |
+
"loss": 0.0493,
|
| 1620 |
+
"step": 1620
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 72.0,
|
| 1624 |
+
"eval_accuracy": 0.6739130434782609,
|
| 1625 |
+
"eval_loss": 1.725121021270752,
|
| 1626 |
+
"eval_runtime": 0.6347,
|
| 1627 |
+
"eval_samples_per_second": 72.476,
|
| 1628 |
+
"eval_steps_per_second": 4.727,
|
| 1629 |
+
"step": 1620
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 72.44,
|
| 1633 |
+
"learning_rate": 2.1254208754208754e-05,
|
| 1634 |
+
"loss": 0.073,
|
| 1635 |
+
"step": 1630
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 72.89,
|
| 1639 |
+
"learning_rate": 2.1043771043771045e-05,
|
| 1640 |
+
"loss": 0.0703,
|
| 1641 |
+
"step": 1640
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 72.98,
|
| 1645 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1646 |
+
"eval_loss": 2.0377814769744873,
|
| 1647 |
+
"eval_runtime": 0.7602,
|
| 1648 |
+
"eval_samples_per_second": 60.513,
|
| 1649 |
+
"eval_steps_per_second": 3.946,
|
| 1650 |
+
"step": 1642
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 73.33,
|
| 1654 |
+
"learning_rate": 2.0833333333333336e-05,
|
| 1655 |
+
"loss": 0.0561,
|
| 1656 |
+
"step": 1650
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 73.78,
|
| 1660 |
+
"learning_rate": 2.0622895622895624e-05,
|
| 1661 |
+
"loss": 0.0692,
|
| 1662 |
+
"step": 1660
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 74.0,
|
| 1666 |
+
"eval_accuracy": 0.6521739130434783,
|
| 1667 |
+
"eval_loss": 2.099923610687256,
|
| 1668 |
+
"eval_runtime": 0.6142,
|
| 1669 |
+
"eval_samples_per_second": 74.893,
|
| 1670 |
+
"eval_steps_per_second": 4.884,
|
| 1671 |
+
"step": 1665
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 74.22,
|
| 1675 |
+
"learning_rate": 2.0412457912457915e-05,
|
| 1676 |
+
"loss": 0.0532,
|
| 1677 |
+
"step": 1670
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 74.67,
|
| 1681 |
+
"learning_rate": 2.0202020202020203e-05,
|
| 1682 |
+
"loss": 0.0396,
|
| 1683 |
+
"step": 1680
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 74.98,
|
| 1687 |
+
"eval_accuracy": 0.6739130434782609,
|
| 1688 |
+
"eval_loss": 2.00740122795105,
|
| 1689 |
+
"eval_runtime": 0.6082,
|
| 1690 |
+
"eval_samples_per_second": 75.638,
|
| 1691 |
+
"eval_steps_per_second": 4.933,
|
| 1692 |
+
"step": 1687
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 75.11,
|
| 1696 |
+
"learning_rate": 1.999158249158249e-05,
|
| 1697 |
+
"loss": 0.038,
|
| 1698 |
+
"step": 1690
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 75.56,
|
| 1702 |
+
"learning_rate": 1.978114478114478e-05,
|
| 1703 |
+
"loss": 0.0209,
|
| 1704 |
+
"step": 1700
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 76.0,
|
| 1708 |
+
"learning_rate": 1.9570707070707073e-05,
|
| 1709 |
+
"loss": 0.0505,
|
| 1710 |
+
"step": 1710
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 76.0,
|
| 1714 |
+
"eval_accuracy": 0.717391304347826,
|
| 1715 |
+
"eval_loss": 1.7462636232376099,
|
| 1716 |
+
"eval_runtime": 0.6273,
|
| 1717 |
+
"eval_samples_per_second": 73.325,
|
| 1718 |
+
"eval_steps_per_second": 4.782,
|
| 1719 |
+
"step": 1710
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 76.44,
|
| 1723 |
+
"learning_rate": 1.936026936026936e-05,
|
| 1724 |
+
"loss": 0.0392,
|
| 1725 |
+
"step": 1720
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 76.89,
|
| 1729 |
+
"learning_rate": 1.914983164983165e-05,
|
| 1730 |
+
"loss": 0.0512,
|
| 1731 |
+
"step": 1730
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 76.98,
|
| 1735 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1736 |
+
"eval_loss": 1.640069603919983,
|
| 1737 |
+
"eval_runtime": 0.6271,
|
| 1738 |
+
"eval_samples_per_second": 73.348,
|
| 1739 |
+
"eval_steps_per_second": 4.784,
|
| 1740 |
+
"step": 1732
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 77.33,
|
| 1744 |
+
"learning_rate": 1.893939393939394e-05,
|
| 1745 |
+
"loss": 0.0625,
|
| 1746 |
+
"step": 1740
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 77.78,
|
| 1750 |
+
"learning_rate": 1.872895622895623e-05,
|
| 1751 |
+
"loss": 0.0653,
|
| 1752 |
+
"step": 1750
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 78.0,
|
| 1756 |
+
"eval_accuracy": 0.6956521739130435,
|
| 1757 |
+
"eval_loss": 1.7835583686828613,
|
| 1758 |
+
"eval_runtime": 0.6087,
|
| 1759 |
+
"eval_samples_per_second": 75.57,
|
| 1760 |
+
"eval_steps_per_second": 4.928,
|
| 1761 |
+
"step": 1755
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 78.22,
|
| 1765 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 1766 |
+
"loss": 0.0487,
|
| 1767 |
+
"step": 1760
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 78.67,
|
| 1771 |
+
"learning_rate": 1.830808080808081e-05,
|
| 1772 |
+
"loss": 0.0764,
|
| 1773 |
+
"step": 1770
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 78.98,
|
| 1777 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1778 |
+
"eval_loss": 1.7904179096221924,
|
| 1779 |
+
"eval_runtime": 0.6272,
|
| 1780 |
+
"eval_samples_per_second": 73.344,
|
| 1781 |
+
"eval_steps_per_second": 4.783,
|
| 1782 |
+
"step": 1777
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 79.11,
|
| 1786 |
+
"learning_rate": 1.80976430976431e-05,
|
| 1787 |
+
"loss": 0.0599,
|
| 1788 |
+
"step": 1780
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 79.56,
|
| 1792 |
+
"learning_rate": 1.788720538720539e-05,
|
| 1793 |
+
"loss": 0.0629,
|
| 1794 |
+
"step": 1790
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 80.0,
|
| 1798 |
+
"learning_rate": 1.7676767676767676e-05,
|
| 1799 |
+
"loss": 0.0598,
|
| 1800 |
+
"step": 1800
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 80.0,
|
| 1804 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1805 |
+
"eval_loss": 1.8720263242721558,
|
| 1806 |
+
"eval_runtime": 0.6152,
|
| 1807 |
+
"eval_samples_per_second": 74.778,
|
| 1808 |
+
"eval_steps_per_second": 4.877,
|
| 1809 |
+
"step": 1800
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 80.44,
|
| 1813 |
+
"learning_rate": 1.7466329966329967e-05,
|
| 1814 |
+
"loss": 0.0561,
|
| 1815 |
+
"step": 1810
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 80.89,
|
| 1819 |
+
"learning_rate": 1.7255892255892255e-05,
|
| 1820 |
+
"loss": 0.0523,
|
| 1821 |
+
"step": 1820
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 80.98,
|
| 1825 |
+
"eval_accuracy": 0.717391304347826,
|
| 1826 |
+
"eval_loss": 1.7355570793151855,
|
| 1827 |
+
"eval_runtime": 0.6281,
|
| 1828 |
+
"eval_samples_per_second": 73.232,
|
| 1829 |
+
"eval_steps_per_second": 4.776,
|
| 1830 |
+
"step": 1822
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 81.33,
|
| 1834 |
+
"learning_rate": 1.7045454545454546e-05,
|
| 1835 |
+
"loss": 0.0344,
|
| 1836 |
+
"step": 1830
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 81.78,
|
| 1840 |
+
"learning_rate": 1.6835016835016837e-05,
|
| 1841 |
+
"loss": 0.0499,
|
| 1842 |
+
"step": 1840
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 82.0,
|
| 1846 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1847 |
+
"eval_loss": 1.9222629070281982,
|
| 1848 |
+
"eval_runtime": 0.6387,
|
| 1849 |
+
"eval_samples_per_second": 72.02,
|
| 1850 |
+
"eval_steps_per_second": 4.697,
|
| 1851 |
+
"step": 1845
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 82.22,
|
| 1855 |
+
"learning_rate": 1.6624579124579125e-05,
|
| 1856 |
+
"loss": 0.0562,
|
| 1857 |
+
"step": 1850
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 82.67,
|
| 1861 |
+
"learning_rate": 1.6414141414141416e-05,
|
| 1862 |
+
"loss": 0.0841,
|
| 1863 |
+
"step": 1860
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 82.98,
|
| 1867 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1868 |
+
"eval_loss": 1.9060295820236206,
|
| 1869 |
+
"eval_runtime": 0.6047,
|
| 1870 |
+
"eval_samples_per_second": 76.075,
|
| 1871 |
+
"eval_steps_per_second": 4.961,
|
| 1872 |
+
"step": 1867
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 83.11,
|
| 1876 |
+
"learning_rate": 1.6203703703703704e-05,
|
| 1877 |
+
"loss": 0.0401,
|
| 1878 |
+
"step": 1870
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 83.56,
|
| 1882 |
+
"learning_rate": 1.5993265993265995e-05,
|
| 1883 |
+
"loss": 0.042,
|
| 1884 |
+
"step": 1880
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 84.0,
|
| 1888 |
+
"learning_rate": 1.5782828282828283e-05,
|
| 1889 |
+
"loss": 0.0597,
|
| 1890 |
+
"step": 1890
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 84.0,
|
| 1894 |
+
"eval_accuracy": 0.717391304347826,
|
| 1895 |
+
"eval_loss": 1.9091507196426392,
|
| 1896 |
+
"eval_runtime": 0.6107,
|
| 1897 |
+
"eval_samples_per_second": 75.323,
|
| 1898 |
+
"eval_steps_per_second": 4.912,
|
| 1899 |
+
"step": 1890
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 84.44,
|
| 1903 |
+
"learning_rate": 1.5572390572390574e-05,
|
| 1904 |
+
"loss": 0.0563,
|
| 1905 |
+
"step": 1900
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 84.89,
|
| 1909 |
+
"learning_rate": 1.5361952861952865e-05,
|
| 1910 |
+
"loss": 0.0696,
|
| 1911 |
+
"step": 1910
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 84.98,
|
| 1915 |
+
"eval_accuracy": 0.6956521739130435,
|
| 1916 |
+
"eval_loss": 1.9534136056900024,
|
| 1917 |
+
"eval_runtime": 0.6437,
|
| 1918 |
+
"eval_samples_per_second": 71.465,
|
| 1919 |
+
"eval_steps_per_second": 4.661,
|
| 1920 |
+
"step": 1912
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 85.33,
|
| 1924 |
+
"learning_rate": 1.5151515151515153e-05,
|
| 1925 |
+
"loss": 0.026,
|
| 1926 |
+
"step": 1920
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 85.78,
|
| 1930 |
+
"learning_rate": 1.4941077441077442e-05,
|
| 1931 |
+
"loss": 0.0566,
|
| 1932 |
+
"step": 1930
|
| 1933 |
+
},
|
| 1934 |
+
{
|
| 1935 |
+
"epoch": 86.0,
|
| 1936 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1937 |
+
"eval_loss": 1.935554027557373,
|
| 1938 |
+
"eval_runtime": 0.6054,
|
| 1939 |
+
"eval_samples_per_second": 75.977,
|
| 1940 |
+
"eval_steps_per_second": 4.955,
|
| 1941 |
+
"step": 1935
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 86.22,
|
| 1945 |
+
"learning_rate": 1.473063973063973e-05,
|
| 1946 |
+
"loss": 0.0326,
|
| 1947 |
+
"step": 1940
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 86.67,
|
| 1951 |
+
"learning_rate": 1.452020202020202e-05,
|
| 1952 |
+
"loss": 0.0435,
|
| 1953 |
+
"step": 1950
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 86.98,
|
| 1957 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1958 |
+
"eval_loss": 2.056638717651367,
|
| 1959 |
+
"eval_runtime": 0.6052,
|
| 1960 |
+
"eval_samples_per_second": 76.011,
|
| 1961 |
+
"eval_steps_per_second": 4.957,
|
| 1962 |
+
"step": 1957
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 87.11,
|
| 1966 |
+
"learning_rate": 1.4309764309764309e-05,
|
| 1967 |
+
"loss": 0.0598,
|
| 1968 |
+
"step": 1960
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 87.56,
|
| 1972 |
+
"learning_rate": 1.4099326599326602e-05,
|
| 1973 |
+
"loss": 0.0348,
|
| 1974 |
+
"step": 1970
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 88.0,
|
| 1978 |
+
"learning_rate": 1.388888888888889e-05,
|
| 1979 |
+
"loss": 0.024,
|
| 1980 |
+
"step": 1980
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 88.0,
|
| 1984 |
+
"eval_accuracy": 0.717391304347826,
|
| 1985 |
+
"eval_loss": 1.8717210292816162,
|
| 1986 |
+
"eval_runtime": 0.6198,
|
| 1987 |
+
"eval_samples_per_second": 74.217,
|
| 1988 |
+
"eval_steps_per_second": 4.84,
|
| 1989 |
+
"step": 1980
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 88.44,
|
| 1993 |
+
"learning_rate": 1.3678451178451179e-05,
|
| 1994 |
+
"loss": 0.0265,
|
| 1995 |
+
"step": 1990
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 88.89,
|
| 1999 |
+
"learning_rate": 1.3468013468013468e-05,
|
| 2000 |
+
"loss": 0.0137,
|
| 2001 |
+
"step": 2000
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 88.98,
|
| 2005 |
+
"eval_accuracy": 0.717391304347826,
|
| 2006 |
+
"eval_loss": 2.027980089187622,
|
| 2007 |
+
"eval_runtime": 0.6032,
|
| 2008 |
+
"eval_samples_per_second": 76.263,
|
| 2009 |
+
"eval_steps_per_second": 4.974,
|
| 2010 |
+
"step": 2002
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 89.33,
|
| 2014 |
+
"learning_rate": 1.3257575757575758e-05,
|
| 2015 |
+
"loss": 0.0516,
|
| 2016 |
+
"step": 2010
|
| 2017 |
+
},
|
| 2018 |
+
{
|
| 2019 |
+
"epoch": 89.78,
|
| 2020 |
+
"learning_rate": 1.3047138047138047e-05,
|
| 2021 |
+
"loss": 0.0663,
|
| 2022 |
+
"step": 2020
|
| 2023 |
+
},
|
| 2024 |
+
{
|
| 2025 |
+
"epoch": 90.0,
|
| 2026 |
+
"eval_accuracy": 0.717391304347826,
|
| 2027 |
+
"eval_loss": 1.8828924894332886,
|
| 2028 |
+
"eval_runtime": 0.6197,
|
| 2029 |
+
"eval_samples_per_second": 74.233,
|
| 2030 |
+
"eval_steps_per_second": 4.841,
|
| 2031 |
+
"step": 2025
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 90.22,
|
| 2035 |
+
"learning_rate": 1.2836700336700338e-05,
|
| 2036 |
+
"loss": 0.0603,
|
| 2037 |
+
"step": 2030
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 90.67,
|
| 2041 |
+
"learning_rate": 1.2626262626262628e-05,
|
| 2042 |
+
"loss": 0.035,
|
| 2043 |
+
"step": 2040
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 90.98,
|
| 2047 |
+
"eval_accuracy": 0.717391304347826,
|
| 2048 |
+
"eval_loss": 1.9687623977661133,
|
| 2049 |
+
"eval_runtime": 0.6247,
|
| 2050 |
+
"eval_samples_per_second": 73.641,
|
| 2051 |
+
"eval_steps_per_second": 4.803,
|
| 2052 |
+
"step": 2047
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 91.11,
|
| 2056 |
+
"learning_rate": 1.2415824915824917e-05,
|
| 2057 |
+
"loss": 0.0285,
|
| 2058 |
+
"step": 2050
|
| 2059 |
+
},
|
| 2060 |
+
{
|
| 2061 |
+
"epoch": 91.56,
|
| 2062 |
+
"learning_rate": 1.2205387205387205e-05,
|
| 2063 |
+
"loss": 0.0435,
|
| 2064 |
+
"step": 2060
|
| 2065 |
+
},
|
| 2066 |
+
{
|
| 2067 |
+
"epoch": 92.0,
|
| 2068 |
+
"learning_rate": 1.1994949494949495e-05,
|
| 2069 |
+
"loss": 0.0504,
|
| 2070 |
+
"step": 2070
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 92.0,
|
| 2074 |
+
"eval_accuracy": 0.7391304347826086,
|
| 2075 |
+
"eval_loss": 2.0455873012542725,
|
| 2076 |
+
"eval_runtime": 0.6222,
|
| 2077 |
+
"eval_samples_per_second": 73.935,
|
| 2078 |
+
"eval_steps_per_second": 4.822,
|
| 2079 |
+
"step": 2070
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 92.44,
|
| 2083 |
+
"learning_rate": 1.1784511784511786e-05,
|
| 2084 |
+
"loss": 0.039,
|
| 2085 |
+
"step": 2080
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 92.89,
|
| 2089 |
+
"learning_rate": 1.1574074074074075e-05,
|
| 2090 |
+
"loss": 0.025,
|
| 2091 |
+
"step": 2090
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 92.98,
|
| 2095 |
+
"eval_accuracy": 0.6304347826086957,
|
| 2096 |
+
"eval_loss": 2.1567869186401367,
|
| 2097 |
+
"eval_runtime": 0.5982,
|
| 2098 |
+
"eval_samples_per_second": 76.901,
|
| 2099 |
+
"eval_steps_per_second": 5.015,
|
| 2100 |
+
"step": 2092
|
| 2101 |
+
},
|
| 2102 |
+
{
|
| 2103 |
+
"epoch": 93.33,
|
| 2104 |
+
"learning_rate": 1.1363636363636365e-05,
|
| 2105 |
+
"loss": 0.0177,
|
| 2106 |
+
"step": 2100
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 93.78,
|
| 2110 |
+
"learning_rate": 1.1153198653198654e-05,
|
| 2111 |
+
"loss": 0.0405,
|
| 2112 |
+
"step": 2110
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 94.0,
|
| 2116 |
+
"eval_accuracy": 0.6304347826086957,
|
| 2117 |
+
"eval_loss": 2.0627565383911133,
|
| 2118 |
+
"eval_runtime": 0.6632,
|
| 2119 |
+
"eval_samples_per_second": 69.363,
|
| 2120 |
+
"eval_steps_per_second": 4.524,
|
| 2121 |
+
"step": 2115
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 94.22,
|
| 2125 |
+
"learning_rate": 1.0942760942760944e-05,
|
| 2126 |
+
"loss": 0.0358,
|
| 2127 |
+
"step": 2120
|
| 2128 |
+
},
|
| 2129 |
+
{
|
| 2130 |
+
"epoch": 94.67,
|
| 2131 |
+
"learning_rate": 1.0732323232323233e-05,
|
| 2132 |
+
"loss": 0.0247,
|
| 2133 |
+
"step": 2130
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 94.98,
|
| 2137 |
+
"eval_accuracy": 0.6956521739130435,
|
| 2138 |
+
"eval_loss": 2.068596124649048,
|
| 2139 |
+
"eval_runtime": 0.6331,
|
| 2140 |
+
"eval_samples_per_second": 72.653,
|
| 2141 |
+
"eval_steps_per_second": 4.738,
|
| 2142 |
+
"step": 2137
|
| 2143 |
+
},
|
| 2144 |
+
{
|
| 2145 |
+
"epoch": 95.11,
|
| 2146 |
+
"learning_rate": 1.0521885521885522e-05,
|
| 2147 |
+
"loss": 0.0242,
|
| 2148 |
+
"step": 2140
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 95.56,
|
| 2152 |
+
"learning_rate": 1.0311447811447812e-05,
|
| 2153 |
+
"loss": 0.0675,
|
| 2154 |
+
"step": 2150
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 96.0,
|
| 2158 |
+
"learning_rate": 1.0101010101010101e-05,
|
| 2159 |
+
"loss": 0.0429,
|
| 2160 |
+
"step": 2160
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 96.0,
|
| 2164 |
+
"eval_accuracy": 0.717391304347826,
|
| 2165 |
+
"eval_loss": 2.1125054359436035,
|
| 2166 |
+
"eval_runtime": 0.6082,
|
| 2167 |
+
"eval_samples_per_second": 75.638,
|
| 2168 |
+
"eval_steps_per_second": 4.933,
|
| 2169 |
+
"step": 2160
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 96.44,
|
| 2173 |
+
"learning_rate": 9.89057239057239e-06,
|
| 2174 |
+
"loss": 0.0287,
|
| 2175 |
+
"step": 2170
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 96.89,
|
| 2179 |
+
"learning_rate": 9.68013468013468e-06,
|
| 2180 |
+
"loss": 0.0408,
|
| 2181 |
+
"step": 2180
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 96.98,
|
| 2185 |
+
"eval_accuracy": 0.7391304347826086,
|
| 2186 |
+
"eval_loss": 2.1003143787384033,
|
| 2187 |
+
"eval_runtime": 0.6066,
|
| 2188 |
+
"eval_samples_per_second": 75.827,
|
| 2189 |
+
"eval_steps_per_second": 4.945,
|
| 2190 |
+
"step": 2182
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 97.33,
|
| 2194 |
+
"learning_rate": 9.46969696969697e-06,
|
| 2195 |
+
"loss": 0.0576,
|
| 2196 |
+
"step": 2190
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 97.78,
|
| 2200 |
+
"learning_rate": 9.259259259259259e-06,
|
| 2201 |
+
"loss": 0.0385,
|
| 2202 |
+
"step": 2200
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 98.0,
|
| 2206 |
+
"eval_accuracy": 0.6739130434782609,
|
| 2207 |
+
"eval_loss": 2.2997324466705322,
|
| 2208 |
+
"eval_runtime": 0.6152,
|
| 2209 |
+
"eval_samples_per_second": 74.778,
|
| 2210 |
+
"eval_steps_per_second": 4.877,
|
| 2211 |
+
"step": 2205
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 98.22,
|
| 2215 |
+
"learning_rate": 9.04882154882155e-06,
|
| 2216 |
+
"loss": 0.0345,
|
| 2217 |
+
"step": 2210
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 98.67,
|
| 2221 |
+
"learning_rate": 8.838383838383838e-06,
|
| 2222 |
+
"loss": 0.0364,
|
| 2223 |
+
"step": 2220
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 98.98,
|
| 2227 |
+
"eval_accuracy": 0.717391304347826,
|
| 2228 |
+
"eval_loss": 2.144188165664673,
|
| 2229 |
+
"eval_runtime": 0.6051,
|
| 2230 |
+
"eval_samples_per_second": 76.015,
|
| 2231 |
+
"eval_steps_per_second": 4.957,
|
| 2232 |
+
"step": 2227
|
| 2233 |
+
},
|
| 2234 |
+
{
|
| 2235 |
+
"epoch": 99.11,
|
| 2236 |
+
"learning_rate": 8.627946127946127e-06,
|
| 2237 |
+
"loss": 0.0505,
|
| 2238 |
+
"step": 2230
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 99.56,
|
| 2242 |
+
"learning_rate": 8.417508417508419e-06,
|
| 2243 |
+
"loss": 0.0271,
|
| 2244 |
+
"step": 2240
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 100.0,
|
| 2248 |
+
"learning_rate": 8.207070707070708e-06,
|
| 2249 |
+
"loss": 0.0415,
|
| 2250 |
+
"step": 2250
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 100.0,
|
| 2254 |
+
"eval_accuracy": 0.717391304347826,
|
| 2255 |
+
"eval_loss": 2.104325294494629,
|
| 2256 |
+
"eval_runtime": 0.6057,
|
| 2257 |
+
"eval_samples_per_second": 75.943,
|
| 2258 |
+
"eval_steps_per_second": 4.953,
|
| 2259 |
+
"step": 2250
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 100.44,
|
| 2263 |
+
"learning_rate": 7.996632996632998e-06,
|
| 2264 |
+
"loss": 0.024,
|
| 2265 |
+
"step": 2260
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 100.89,
|
| 2269 |
+
"learning_rate": 7.786195286195287e-06,
|
| 2270 |
+
"loss": 0.0175,
|
| 2271 |
+
"step": 2270
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 100.98,
|
| 2275 |
+
"eval_accuracy": 0.6739130434782609,
|
| 2276 |
+
"eval_loss": 2.1847035884857178,
|
| 2277 |
+
"eval_runtime": 0.6572,
|
| 2278 |
+
"eval_samples_per_second": 69.993,
|
| 2279 |
+
"eval_steps_per_second": 4.565,
|
| 2280 |
+
"step": 2272
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 101.33,
|
| 2284 |
+
"learning_rate": 7.5757575757575764e-06,
|
| 2285 |
+
"loss": 0.0488,
|
| 2286 |
+
"step": 2280
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 101.78,
|
| 2290 |
+
"learning_rate": 7.365319865319865e-06,
|
| 2291 |
+
"loss": 0.0281,
|
| 2292 |
+
"step": 2290
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 102.0,
|
| 2296 |
+
"eval_accuracy": 0.6521739130434783,
|
| 2297 |
+
"eval_loss": 2.3262391090393066,
|
| 2298 |
+
"eval_runtime": 0.6817,
|
| 2299 |
+
"eval_samples_per_second": 67.477,
|
| 2300 |
+
"eval_steps_per_second": 4.401,
|
| 2301 |
+
"step": 2295
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 102.22,
|
| 2305 |
+
"learning_rate": 7.1548821548821545e-06,
|
| 2306 |
+
"loss": 0.0485,
|
| 2307 |
+
"step": 2300
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 102.67,
|
| 2311 |
+
"learning_rate": 6.944444444444445e-06,
|
| 2312 |
+
"loss": 0.0268,
|
| 2313 |
+
"step": 2310
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 102.98,
|
| 2317 |
+
"eval_accuracy": 0.6956521739130435,
|
| 2318 |
+
"eval_loss": 2.2843432426452637,
|
| 2319 |
+
"eval_runtime": 0.6162,
|
| 2320 |
+
"eval_samples_per_second": 74.654,
|
| 2321 |
+
"eval_steps_per_second": 4.869,
|
| 2322 |
+
"step": 2317
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 103.11,
|
| 2326 |
+
"learning_rate": 6.734006734006734e-06,
|
| 2327 |
+
"loss": 0.0389,
|
| 2328 |
+
"step": 2320
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 103.56,
|
| 2332 |
+
"learning_rate": 6.523569023569024e-06,
|
| 2333 |
+
"loss": 0.0389,
|
| 2334 |
+
"step": 2330
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 104.0,
|
| 2338 |
+
"learning_rate": 6.313131313131314e-06,
|
| 2339 |
+
"loss": 0.022,
|
| 2340 |
+
"step": 2340
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 104.0,
|
| 2344 |
+
"eval_accuracy": 0.6956521739130435,
|
| 2345 |
+
"eval_loss": 2.352231502532959,
|
| 2346 |
+
"eval_runtime": 0.5942,
|
| 2347 |
+
"eval_samples_per_second": 77.419,
|
| 2348 |
+
"eval_steps_per_second": 5.049,
|
| 2349 |
+
"step": 2340
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 104.44,
|
| 2353 |
+
"learning_rate": 6.1026936026936026e-06,
|
| 2354 |
+
"loss": 0.0284,
|
| 2355 |
+
"step": 2350
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 104.89,
|
| 2359 |
+
"learning_rate": 5.892255892255893e-06,
|
| 2360 |
+
"loss": 0.0279,
|
| 2361 |
+
"step": 2360
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 104.98,
|
| 2365 |
+
"eval_accuracy": 0.6739130434782609,
|
| 2366 |
+
"eval_loss": 2.4826581478118896,
|
| 2367 |
+
"eval_runtime": 0.6241,
|
| 2368 |
+
"eval_samples_per_second": 73.701,
|
| 2369 |
+
"eval_steps_per_second": 4.807,
|
| 2370 |
+
"step": 2362
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 105.33,
|
| 2374 |
+
"learning_rate": 5.681818181818182e-06,
|
| 2375 |
+
"loss": 0.0472,
|
| 2376 |
+
"step": 2370
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 105.78,
|
| 2380 |
+
"learning_rate": 5.471380471380472e-06,
|
| 2381 |
+
"loss": 0.0188,
|
| 2382 |
+
"step": 2380
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 106.0,
|
| 2386 |
+
"eval_accuracy": 0.6521739130434783,
|
| 2387 |
+
"eval_loss": 2.5688304901123047,
|
| 2388 |
+
"eval_runtime": 0.6101,
|
| 2389 |
+
"eval_samples_per_second": 75.392,
|
| 2390 |
+
"eval_steps_per_second": 4.917,
|
| 2391 |
+
"step": 2385
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 106.22,
|
| 2395 |
+
"learning_rate": 5.260942760942761e-06,
|
| 2396 |
+
"loss": 0.0404,
|
| 2397 |
+
"step": 2390
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 106.67,
|
| 2401 |
+
"learning_rate": 5.050505050505051e-06,
|
| 2402 |
+
"loss": 0.0303,
|
| 2403 |
+
"step": 2400
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 106.98,
|
| 2407 |
+
"eval_accuracy": 0.6521739130434783,
|
| 2408 |
+
"eval_loss": 2.4357047080993652,
|
| 2409 |
+
"eval_runtime": 0.6062,
|
| 2410 |
+
"eval_samples_per_second": 75.884,
|
| 2411 |
+
"eval_steps_per_second": 4.949,
|
| 2412 |
+
"step": 2407
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 107.11,
|
| 2416 |
+
"learning_rate": 4.84006734006734e-06,
|
| 2417 |
+
"loss": 0.0397,
|
| 2418 |
+
"step": 2410
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 107.56,
|
| 2422 |
+
"learning_rate": 4.6296296296296296e-06,
|
| 2423 |
+
"loss": 0.0592,
|
| 2424 |
+
"step": 2420
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 108.0,
|
| 2428 |
+
"learning_rate": 4.419191919191919e-06,
|
| 2429 |
+
"loss": 0.0439,
|
| 2430 |
+
"step": 2430
|
| 2431 |
+
},
|
| 2432 |
+
{
|
| 2433 |
+
"epoch": 108.0,
|
| 2434 |
+
"eval_accuracy": 0.6521739130434783,
|
| 2435 |
+
"eval_loss": 2.4359428882598877,
|
| 2436 |
+
"eval_runtime": 0.6317,
|
| 2437 |
+
"eval_samples_per_second": 72.823,
|
| 2438 |
+
"eval_steps_per_second": 4.749,
|
| 2439 |
+
"step": 2430
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 108.44,
|
| 2443 |
+
"learning_rate": 4.208754208754209e-06,
|
| 2444 |
+
"loss": 0.0501,
|
| 2445 |
+
"step": 2440
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 108.89,
|
| 2449 |
+
"learning_rate": 3.998316498316499e-06,
|
| 2450 |
+
"loss": 0.0422,
|
| 2451 |
+
"step": 2450
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 108.98,
|
| 2455 |
+
"eval_accuracy": 0.6739130434782609,
|
| 2456 |
+
"eval_loss": 2.472536563873291,
|
| 2457 |
+
"eval_runtime": 0.6173,
|
| 2458 |
+
"eval_samples_per_second": 74.523,
|
| 2459 |
+
"eval_steps_per_second": 4.86,
|
| 2460 |
+
"step": 2452
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 109.33,
|
| 2464 |
+
"learning_rate": 3.7878787878787882e-06,
|
| 2465 |
+
"loss": 0.0158,
|
| 2466 |
+
"step": 2460
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 109.78,
|
| 2470 |
+
"learning_rate": 3.5774410774410772e-06,
|
| 2471 |
+
"loss": 0.032,
|
| 2472 |
+
"step": 2470
|
| 2473 |
+
},
|
| 2474 |
+
{
|
| 2475 |
+
"epoch": 110.0,
|
| 2476 |
+
"eval_accuracy": 0.6521739130434783,
|
| 2477 |
+
"eval_loss": 2.289853811264038,
|
| 2478 |
+
"eval_runtime": 0.6113,
|
| 2479 |
+
"eval_samples_per_second": 75.246,
|
| 2480 |
+
"eval_steps_per_second": 4.907,
|
| 2481 |
+
"step": 2475
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 110.22,
|
| 2485 |
+
"learning_rate": 3.367003367003367e-06,
|
| 2486 |
+
"loss": 0.051,
|
| 2487 |
+
"step": 2480
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 110.67,
|
| 2491 |
+
"learning_rate": 3.156565656565657e-06,
|
| 2492 |
+
"loss": 0.0414,
|
| 2493 |
+
"step": 2490
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 110.98,
|
| 2497 |
+
"eval_accuracy": 0.6956521739130435,
|
| 2498 |
+
"eval_loss": 2.268512010574341,
|
| 2499 |
+
"eval_runtime": 0.6027,
|
| 2500 |
+
"eval_samples_per_second": 76.328,
|
| 2501 |
+
"eval_steps_per_second": 4.978,
|
| 2502 |
+
"step": 2497
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 111.11,
|
| 2506 |
+
"learning_rate": 2.9461279461279464e-06,
|
| 2507 |
+
"loss": 0.0532,
|
| 2508 |
+
"step": 2500
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 111.56,
|
| 2512 |
+
"learning_rate": 2.735690235690236e-06,
|
| 2513 |
+
"loss": 0.0461,
|
| 2514 |
+
"step": 2510
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 112.0,
|
| 2518 |
+
"learning_rate": 2.5252525252525253e-06,
|
| 2519 |
+
"loss": 0.03,
|
| 2520 |
+
"step": 2520
|
| 2521 |
+
},
|
| 2522 |
+
{
|
| 2523 |
+
"epoch": 112.0,
|
| 2524 |
+
"eval_accuracy": 0.6739130434782609,
|
| 2525 |
+
"eval_loss": 2.306300163269043,
|
| 2526 |
+
"eval_runtime": 0.6202,
|
| 2527 |
+
"eval_samples_per_second": 74.172,
|
| 2528 |
+
"eval_steps_per_second": 4.837,
|
| 2529 |
+
"step": 2520
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 112.44,
|
| 2533 |
+
"learning_rate": 2.3148148148148148e-06,
|
| 2534 |
+
"loss": 0.0454,
|
| 2535 |
+
"step": 2530
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 112.89,
|
| 2539 |
+
"learning_rate": 2.1043771043771047e-06,
|
| 2540 |
+
"loss": 0.0293,
|
| 2541 |
+
"step": 2540
|
| 2542 |
+
},
|
| 2543 |
+
{
|
| 2544 |
+
"epoch": 112.98,
|
| 2545 |
+
"eval_accuracy": 0.6739130434782609,
|
| 2546 |
+
"eval_loss": 2.352445602416992,
|
| 2547 |
+
"eval_runtime": 0.6112,
|
| 2548 |
+
"eval_samples_per_second": 75.268,
|
| 2549 |
+
"eval_steps_per_second": 4.909,
|
| 2550 |
+
"step": 2542
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 113.33,
|
| 2554 |
+
"learning_rate": 1.8939393939393941e-06,
|
| 2555 |
+
"loss": 0.037,
|
| 2556 |
+
"step": 2550
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 113.78,
|
| 2560 |
+
"learning_rate": 1.6835016835016836e-06,
|
| 2561 |
+
"loss": 0.0514,
|
| 2562 |
+
"step": 2560
|
| 2563 |
+
},
|
| 2564 |
+
{
|
| 2565 |
+
"epoch": 114.0,
|
| 2566 |
+
"eval_accuracy": 0.6739130434782609,
|
| 2567 |
+
"eval_loss": 2.361192464828491,
|
| 2568 |
+
"eval_runtime": 0.6632,
|
| 2569 |
+
"eval_samples_per_second": 69.363,
|
| 2570 |
+
"eval_steps_per_second": 4.524,
|
| 2571 |
+
"step": 2565
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 114.22,
|
| 2575 |
+
"learning_rate": 1.4730639730639732e-06,
|
| 2576 |
+
"loss": 0.0786,
|
| 2577 |
+
"step": 2570
|
| 2578 |
+
},
|
| 2579 |
+
{
|
| 2580 |
+
"epoch": 114.67,
|
| 2581 |
+
"learning_rate": 1.2626262626262627e-06,
|
| 2582 |
+
"loss": 0.0234,
|
| 2583 |
+
"step": 2580
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 114.98,
|
| 2587 |
+
"eval_accuracy": 0.6521739130434783,
|
| 2588 |
+
"eval_loss": 2.371103286743164,
|
| 2589 |
+
"eval_runtime": 0.6022,
|
| 2590 |
+
"eval_samples_per_second": 76.39,
|
| 2591 |
+
"eval_steps_per_second": 4.982,
|
| 2592 |
+
"step": 2587
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 115.11,
|
| 2596 |
+
"learning_rate": 1.0521885521885523e-06,
|
| 2597 |
+
"loss": 0.0222,
|
| 2598 |
+
"step": 2590
|
| 2599 |
+
},
|
| 2600 |
+
{
|
| 2601 |
+
"epoch": 115.56,
|
| 2602 |
+
"learning_rate": 8.417508417508418e-07,
|
| 2603 |
+
"loss": 0.0355,
|
| 2604 |
+
"step": 2600
|
| 2605 |
+
},
|
| 2606 |
+
{
|
| 2607 |
+
"epoch": 116.0,
|
| 2608 |
+
"learning_rate": 6.313131313131313e-07,
|
| 2609 |
+
"loss": 0.0476,
|
| 2610 |
+
"step": 2610
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 116.0,
|
| 2614 |
+
"eval_accuracy": 0.6739130434782609,
|
| 2615 |
+
"eval_loss": 2.354818344116211,
|
| 2616 |
+
"eval_runtime": 0.6402,
|
| 2617 |
+
"eval_samples_per_second": 71.855,
|
| 2618 |
+
"eval_steps_per_second": 4.686,
|
| 2619 |
+
"step": 2610
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 116.44,
|
| 2623 |
+
"learning_rate": 4.208754208754209e-07,
|
| 2624 |
+
"loss": 0.0291,
|
| 2625 |
+
"step": 2620
|
| 2626 |
+
},
|
| 2627 |
+
{
|
| 2628 |
+
"epoch": 116.89,
|
| 2629 |
+
"learning_rate": 2.1043771043771044e-07,
|
| 2630 |
+
"loss": 0.0307,
|
| 2631 |
+
"step": 2630
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 116.98,
|
| 2635 |
+
"eval_accuracy": 0.6739130434782609,
|
| 2636 |
+
"eval_loss": 2.343219041824341,
|
| 2637 |
+
"eval_runtime": 0.6017,
|
| 2638 |
+
"eval_samples_per_second": 76.455,
|
| 2639 |
+
"eval_steps_per_second": 4.986,
|
| 2640 |
+
"step": 2632
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 117.33,
|
| 2644 |
+
"learning_rate": 0.0,
|
| 2645 |
+
"loss": 0.028,
|
| 2646 |
+
"step": 2640
|
| 2647 |
+
},
|
| 2648 |
+
{
|
| 2649 |
+
"epoch": 117.33,
|
| 2650 |
+
"eval_accuracy": 0.6739130434782609,
|
| 2651 |
+
"eval_loss": 2.3432161808013916,
|
| 2652 |
+
"eval_runtime": 0.6057,
|
| 2653 |
+
"eval_samples_per_second": 75.951,
|
| 2654 |
+
"eval_steps_per_second": 4.953,
|
| 2655 |
+
"step": 2640
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 117.33,
|
| 2659 |
+
"step": 2640,
|
| 2660 |
+
"total_flos": 5.466852859010089e+18,
|
| 2661 |
+
"train_loss": 0.1724454693552671,
|
| 2662 |
+
"train_runtime": 3352.4175,
|
| 2663 |
+
"train_samples_per_second": 51.259,
|
| 2664 |
+
"train_steps_per_second": 0.787
|
| 2665 |
+
}
|
| 2666 |
+
],
|
| 2667 |
+
"logging_steps": 10,
|
| 2668 |
+
"max_steps": 2640,
|
| 2669 |
+
"num_input_tokens_seen": 0,
|
| 2670 |
+
"num_train_epochs": 120,
|
| 2671 |
+
"save_steps": 500,
|
| 2672 |
+
"total_flos": 5.466852859010089e+18,
|
| 2673 |
+
"train_batch_size": 16,
|
| 2674 |
+
"trial_name": null,
|
| 2675 |
+
"trial_params": null
|
| 2676 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:41494d048bf2389e0e35bf1dba77777e6fd51a2b20686299c29ef120a962c400
|
| 3 |
+
size 4728
|