<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
rvlcdip-tiny_rvl_cdip-NK1000_kd_NKD_t1.0_g1.5
This model is a fine-tuned version of WinKawaks/vit-tiny-patch16-224 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 5.4561
- Accuracy: 0.802
- Brier Loss: 0.3399
- Nll: 1.6335
- F1 Micro: 0.802
- F1 Macro: 0.8037
- Ece: 0.1478
- Aurc: 0.0576
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
---|---|---|---|---|---|---|---|---|---|---|
No log | 1.0 | 125 | 5.6533 | 0.5288 | 0.6696 | 3.8280 | 0.5288 | 0.4993 | 0.2170 | 0.2294 |
No log | 2.0 | 250 | 5.3016 | 0.6285 | 0.5364 | 2.7651 | 0.6285 | 0.6089 | 0.1861 | 0.1465 |
No log | 3.0 | 375 | 5.1153 | 0.696 | 0.4775 | 2.4052 | 0.696 | 0.6956 | 0.2003 | 0.1078 |
5.7003 | 4.0 | 500 | 4.9491 | 0.7358 | 0.3968 | 2.1532 | 0.7358 | 0.7375 | 0.1406 | 0.0813 |
5.7003 | 5.0 | 625 | 4.8556 | 0.754 | 0.3676 | 1.8243 | 0.754 | 0.7472 | 0.1001 | 0.0756 |
5.7003 | 6.0 | 750 | 4.8060 | 0.7625 | 0.3475 | 1.8558 | 0.7625 | 0.7636 | 0.0808 | 0.0696 |
5.7003 | 7.0 | 875 | 4.8301 | 0.7648 | 0.3320 | 1.7367 | 0.7648 | 0.7663 | 0.0434 | 0.0677 |
4.6459 | 8.0 | 1000 | 4.7883 | 0.7692 | 0.3305 | 1.8366 | 0.7692 | 0.7728 | 0.0532 | 0.0666 |
4.6459 | 9.0 | 1125 | 4.8347 | 0.7762 | 0.3282 | 1.7122 | 0.7762 | 0.7789 | 0.0610 | 0.0675 |
4.6459 | 10.0 | 1250 | 4.8679 | 0.7682 | 0.3338 | 1.8225 | 0.7682 | 0.7713 | 0.0634 | 0.0672 |
4.6459 | 11.0 | 1375 | 4.9875 | 0.7655 | 0.3521 | 1.9651 | 0.7655 | 0.7647 | 0.0914 | 0.0692 |
4.2436 | 12.0 | 1500 | 4.9708 | 0.77 | 0.3410 | 2.0195 | 0.7700 | 0.7694 | 0.0838 | 0.0684 |
4.2436 | 13.0 | 1625 | 4.9246 | 0.7752 | 0.3349 | 1.8150 | 0.7752 | 0.7758 | 0.0801 | 0.0666 |
4.2436 | 14.0 | 1750 | 4.9235 | 0.776 | 0.3327 | 1.8364 | 0.776 | 0.7782 | 0.0896 | 0.0628 |
4.2436 | 15.0 | 1875 | 4.9149 | 0.7817 | 0.3348 | 1.9243 | 0.7817 | 0.7857 | 0.0917 | 0.0650 |
4.0997 | 16.0 | 2000 | 4.8998 | 0.7837 | 0.3255 | 1.8326 | 0.7837 | 0.7874 | 0.0901 | 0.0637 |
4.0997 | 17.0 | 2125 | 4.9658 | 0.7792 | 0.3358 | 1.8156 | 0.7792 | 0.7815 | 0.1025 | 0.0640 |
4.0997 | 18.0 | 2250 | 4.9819 | 0.7905 | 0.3256 | 1.8605 | 0.7905 | 0.7919 | 0.1016 | 0.0613 |
4.0997 | 19.0 | 2375 | 5.0040 | 0.778 | 0.3417 | 1.9392 | 0.778 | 0.7800 | 0.1095 | 0.0638 |
4.0325 | 20.0 | 2500 | 5.0084 | 0.7817 | 0.3387 | 1.9882 | 0.7817 | 0.7833 | 0.1043 | 0.0642 |
4.0325 | 21.0 | 2625 | 5.0680 | 0.7805 | 0.3473 | 1.8641 | 0.7805 | 0.7803 | 0.1200 | 0.0631 |
4.0325 | 22.0 | 2750 | 5.0324 | 0.7808 | 0.3395 | 1.8541 | 0.7808 | 0.7835 | 0.1124 | 0.0620 |
4.0325 | 23.0 | 2875 | 5.0734 | 0.7845 | 0.3446 | 1.9087 | 0.7845 | 0.7884 | 0.1170 | 0.0625 |
3.99 | 24.0 | 3000 | 5.2144 | 0.782 | 0.3564 | 1.9540 | 0.782 | 0.7845 | 0.1293 | 0.0640 |
3.99 | 25.0 | 3125 | 5.0299 | 0.7873 | 0.3387 | 1.8106 | 0.7873 | 0.7887 | 0.1167 | 0.0614 |
3.99 | 26.0 | 3250 | 5.0673 | 0.792 | 0.3318 | 1.7538 | 0.792 | 0.7930 | 0.1134 | 0.0599 |
3.99 | 27.0 | 3375 | 5.0854 | 0.791 | 0.3379 | 1.8144 | 0.791 | 0.7932 | 0.1253 | 0.0586 |
3.9606 | 28.0 | 3500 | 5.0962 | 0.787 | 0.3403 | 1.7780 | 0.787 | 0.7884 | 0.1224 | 0.0592 |
3.9606 | 29.0 | 3625 | 5.0812 | 0.7877 | 0.3379 | 1.7721 | 0.7877 | 0.7900 | 0.1247 | 0.0592 |
3.9606 | 30.0 | 3750 | 5.1318 | 0.7905 | 0.3359 | 1.8105 | 0.7905 | 0.7931 | 0.1290 | 0.0597 |
3.9606 | 31.0 | 3875 | 5.0330 | 0.7953 | 0.3276 | 1.7361 | 0.7953 | 0.7978 | 0.1144 | 0.0584 |
3.9355 | 32.0 | 4000 | 5.0843 | 0.7975 | 0.3276 | 1.7556 | 0.7975 | 0.7990 | 0.1236 | 0.0560 |
3.9355 | 33.0 | 4125 | 5.1843 | 0.7995 | 0.3315 | 1.7084 | 0.7995 | 0.8004 | 0.1297 | 0.0575 |
3.9355 | 34.0 | 4250 | 5.1703 | 0.7987 | 0.3333 | 1.6918 | 0.7987 | 0.8000 | 0.1257 | 0.0580 |
3.9355 | 35.0 | 4375 | 5.1933 | 0.7937 | 0.3372 | 1.7084 | 0.7937 | 0.7941 | 0.1307 | 0.0561 |
3.9148 | 36.0 | 4500 | 5.1404 | 0.7987 | 0.3275 | 1.6423 | 0.7987 | 0.8011 | 0.1308 | 0.0547 |
3.9148 | 37.0 | 4625 | 5.1734 | 0.8017 | 0.3272 | 1.6836 | 0.8017 | 0.8034 | 0.1272 | 0.0572 |
3.9148 | 38.0 | 4750 | 5.2479 | 0.802 | 0.3322 | 1.7081 | 0.802 | 0.8032 | 0.1353 | 0.0550 |
3.9148 | 39.0 | 4875 | 5.1921 | 0.8 | 0.3320 | 1.6554 | 0.8000 | 0.8012 | 0.1334 | 0.0538 |
3.9001 | 40.0 | 5000 | 5.2477 | 0.801 | 0.3353 | 1.6333 | 0.801 | 0.8022 | 0.1390 | 0.0539 |
3.9001 | 41.0 | 5125 | 5.2140 | 0.801 | 0.3299 | 1.6370 | 0.801 | 0.8017 | 0.1340 | 0.0544 |
3.9001 | 42.0 | 5250 | 5.2660 | 0.807 | 0.3303 | 1.6090 | 0.807 | 0.8079 | 0.1339 | 0.0545 |
3.9001 | 43.0 | 5375 | 5.2884 | 0.8007 | 0.3319 | 1.6816 | 0.8007 | 0.8022 | 0.1394 | 0.0547 |
3.8892 | 44.0 | 5500 | 5.3358 | 0.804 | 0.3352 | 1.6399 | 0.804 | 0.8049 | 0.1387 | 0.0560 |
3.8892 | 45.0 | 5625 | 5.3545 | 0.8043 | 0.3349 | 1.6445 | 0.8043 | 0.8060 | 0.1408 | 0.0555 |
3.8892 | 46.0 | 5750 | 5.4026 | 0.8033 | 0.3373 | 1.6493 | 0.8033 | 0.8049 | 0.1439 | 0.0567 |
3.8892 | 47.0 | 5875 | 5.4195 | 0.8015 | 0.3386 | 1.6393 | 0.8015 | 0.8031 | 0.1468 | 0.0570 |
3.8834 | 48.0 | 6000 | 5.4409 | 0.803 | 0.3396 | 1.6392 | 0.803 | 0.8046 | 0.1458 | 0.0574 |
3.8834 | 49.0 | 6125 | 5.4501 | 0.8023 | 0.3395 | 1.6367 | 0.8023 | 0.8039 | 0.1468 | 0.0574 |
3.8834 | 50.0 | 6250 | 5.4561 | 0.802 | 0.3399 | 1.6335 | 0.802 | 0.8037 | 0.1478 | 0.0576 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2