<!-- 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. -->
vit-small_tobacco3482_kd_CEKD_t2.5_a0.9
This model is a fine-tuned version of WinKawaks/vit-small-patch16-224 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5446
- Accuracy: 0.85
- Brier Loss: 0.2446
- Nll: 1.0816
- F1 Micro: 0.85
- F1 Macro: 0.8348
- Ece: 0.1474
- Aurc: 0.0436
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: 100
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
---|---|---|---|---|---|---|---|---|---|---|
No log | 1.0 | 7 | 2.1216 | 0.215 | 0.8751 | 5.2864 | 0.2150 | 0.1264 | 0.2697 | 0.6907 |
No log | 2.0 | 14 | 1.7056 | 0.405 | 0.7400 | 3.5721 | 0.405 | 0.2275 | 0.2995 | 0.4011 |
No log | 3.0 | 21 | 1.1857 | 0.62 | 0.5612 | 2.0143 | 0.62 | 0.5712 | 0.2994 | 0.2024 |
No log | 4.0 | 28 | 0.8767 | 0.705 | 0.4085 | 1.6918 | 0.705 | 0.6436 | 0.2231 | 0.1152 |
No log | 5.0 | 35 | 0.8620 | 0.72 | 0.3878 | 1.7931 | 0.72 | 0.7294 | 0.2233 | 0.1076 |
No log | 6.0 | 42 | 0.7517 | 0.775 | 0.3252 | 1.5573 | 0.775 | 0.7600 | 0.1970 | 0.0790 |
No log | 7.0 | 49 | 0.7280 | 0.79 | 0.3175 | 1.5140 | 0.79 | 0.7742 | 0.1903 | 0.0826 |
No log | 8.0 | 56 | 0.6848 | 0.8 | 0.2942 | 1.4438 | 0.8000 | 0.7902 | 0.1828 | 0.0866 |
No log | 9.0 | 63 | 0.6744 | 0.81 | 0.2889 | 1.4703 | 0.81 | 0.7969 | 0.1989 | 0.0692 |
No log | 10.0 | 70 | 0.8432 | 0.74 | 0.3859 | 1.3134 | 0.74 | 0.7206 | 0.1959 | 0.1051 |
No log | 11.0 | 77 | 0.7424 | 0.765 | 0.3294 | 1.5162 | 0.765 | 0.7792 | 0.2005 | 0.1048 |
No log | 12.0 | 84 | 0.6953 | 0.79 | 0.3194 | 1.2233 | 0.79 | 0.7850 | 0.1800 | 0.0922 |
No log | 13.0 | 91 | 0.5703 | 0.845 | 0.2538 | 1.2355 | 0.845 | 0.8372 | 0.1739 | 0.0447 |
No log | 14.0 | 98 | 0.6439 | 0.795 | 0.2924 | 1.2777 | 0.795 | 0.7743 | 0.1771 | 0.0534 |
No log | 15.0 | 105 | 0.5895 | 0.825 | 0.2650 | 1.2086 | 0.825 | 0.8071 | 0.1665 | 0.0566 |
No log | 16.0 | 112 | 0.5973 | 0.81 | 0.2753 | 1.0959 | 0.81 | 0.8013 | 0.1839 | 0.0534 |
No log | 17.0 | 119 | 0.5825 | 0.795 | 0.2722 | 1.1565 | 0.795 | 0.7886 | 0.1855 | 0.0534 |
No log | 18.0 | 126 | 0.5854 | 0.845 | 0.2661 | 1.1223 | 0.845 | 0.8424 | 0.1981 | 0.0549 |
No log | 19.0 | 133 | 0.5514 | 0.82 | 0.2553 | 0.9585 | 0.82 | 0.8150 | 0.1600 | 0.0481 |
No log | 20.0 | 140 | 0.5600 | 0.835 | 0.2443 | 1.2692 | 0.835 | 0.8232 | 0.1657 | 0.0469 |
No log | 21.0 | 147 | 0.5592 | 0.845 | 0.2473 | 1.1658 | 0.845 | 0.8331 | 0.1683 | 0.0493 |
No log | 22.0 | 154 | 0.5507 | 0.845 | 0.2411 | 1.1403 | 0.845 | 0.8311 | 0.1797 | 0.0450 |
No log | 23.0 | 161 | 0.5305 | 0.84 | 0.2361 | 1.1509 | 0.8400 | 0.8287 | 0.1650 | 0.0409 |
No log | 24.0 | 168 | 0.5352 | 0.835 | 0.2378 | 1.2208 | 0.835 | 0.8201 | 0.1515 | 0.0420 |
No log | 25.0 | 175 | 0.5425 | 0.845 | 0.2420 | 1.1208 | 0.845 | 0.8321 | 0.1776 | 0.0430 |
No log | 26.0 | 182 | 0.5396 | 0.84 | 0.2409 | 1.1230 | 0.8400 | 0.8286 | 0.1647 | 0.0446 |
No log | 27.0 | 189 | 0.5436 | 0.85 | 0.2401 | 1.1179 | 0.85 | 0.8387 | 0.1568 | 0.0427 |
No log | 28.0 | 196 | 0.5373 | 0.835 | 0.2415 | 1.1092 | 0.835 | 0.8141 | 0.1641 | 0.0427 |
No log | 29.0 | 203 | 0.5420 | 0.845 | 0.2436 | 1.0988 | 0.845 | 0.8326 | 0.1551 | 0.0444 |
No log | 30.0 | 210 | 0.5413 | 0.845 | 0.2420 | 1.1064 | 0.845 | 0.8312 | 0.1486 | 0.0440 |
No log | 31.0 | 217 | 0.5411 | 0.84 | 0.2418 | 1.1024 | 0.8400 | 0.8286 | 0.1565 | 0.0435 |
No log | 32.0 | 224 | 0.5426 | 0.845 | 0.2429 | 1.0993 | 0.845 | 0.8322 | 0.1631 | 0.0433 |
No log | 33.0 | 231 | 0.5424 | 0.85 | 0.2426 | 1.0989 | 0.85 | 0.8348 | 0.1615 | 0.0436 |
No log | 34.0 | 238 | 0.5406 | 0.84 | 0.2419 | 1.0979 | 0.8400 | 0.8251 | 0.1640 | 0.0440 |
No log | 35.0 | 245 | 0.5438 | 0.85 | 0.2436 | 1.0953 | 0.85 | 0.8348 | 0.1595 | 0.0438 |
No log | 36.0 | 252 | 0.5429 | 0.85 | 0.2429 | 1.0970 | 0.85 | 0.8348 | 0.1495 | 0.0433 |
No log | 37.0 | 259 | 0.5431 | 0.85 | 0.2427 | 1.0951 | 0.85 | 0.8348 | 0.1617 | 0.0435 |
No log | 38.0 | 266 | 0.5424 | 0.85 | 0.2426 | 1.0959 | 0.85 | 0.8348 | 0.1587 | 0.0434 |
No log | 39.0 | 273 | 0.5428 | 0.85 | 0.2432 | 1.0924 | 0.85 | 0.8348 | 0.1512 | 0.0433 |
No log | 40.0 | 280 | 0.5437 | 0.85 | 0.2438 | 1.0911 | 0.85 | 0.8348 | 0.1726 | 0.0438 |
No log | 41.0 | 287 | 0.5438 | 0.85 | 0.2434 | 1.0925 | 0.85 | 0.8348 | 0.1704 | 0.0433 |
No log | 42.0 | 294 | 0.5428 | 0.85 | 0.2432 | 1.0927 | 0.85 | 0.8348 | 0.1585 | 0.0436 |
No log | 43.0 | 301 | 0.5455 | 0.85 | 0.2443 | 1.0907 | 0.85 | 0.8348 | 0.1756 | 0.0437 |
No log | 44.0 | 308 | 0.5427 | 0.85 | 0.2433 | 1.0908 | 0.85 | 0.8348 | 0.1616 | 0.0433 |
No log | 45.0 | 315 | 0.5456 | 0.85 | 0.2446 | 1.0878 | 0.85 | 0.8348 | 0.1767 | 0.0437 |
No log | 46.0 | 322 | 0.5439 | 0.85 | 0.2438 | 1.0895 | 0.85 | 0.8348 | 0.1503 | 0.0435 |
No log | 47.0 | 329 | 0.5448 | 0.85 | 0.2443 | 1.0891 | 0.85 | 0.8348 | 0.1674 | 0.0439 |
No log | 48.0 | 336 | 0.5440 | 0.85 | 0.2437 | 1.0898 | 0.85 | 0.8348 | 0.1768 | 0.0437 |
No log | 49.0 | 343 | 0.5443 | 0.85 | 0.2441 | 1.0883 | 0.85 | 0.8348 | 0.1433 | 0.0432 |
No log | 50.0 | 350 | 0.5449 | 0.85 | 0.2444 | 1.0877 | 0.85 | 0.8348 | 0.1722 | 0.0436 |
No log | 51.0 | 357 | 0.5443 | 0.85 | 0.2442 | 1.0871 | 0.85 | 0.8348 | 0.1606 | 0.0434 |
No log | 52.0 | 364 | 0.5453 | 0.85 | 0.2444 | 1.0865 | 0.85 | 0.8348 | 0.1729 | 0.0436 |
No log | 53.0 | 371 | 0.5433 | 0.845 | 0.2438 | 1.0873 | 0.845 | 0.8287 | 0.1570 | 0.0434 |
No log | 54.0 | 378 | 0.5453 | 0.85 | 0.2447 | 1.0854 | 0.85 | 0.8348 | 0.1606 | 0.0435 |
No log | 55.0 | 385 | 0.5438 | 0.85 | 0.2439 | 1.0868 | 0.85 | 0.8348 | 0.1721 | 0.0434 |
No log | 56.0 | 392 | 0.5455 | 0.85 | 0.2447 | 1.0853 | 0.85 | 0.8348 | 0.1710 | 0.0437 |
No log | 57.0 | 399 | 0.5435 | 0.85 | 0.2439 | 1.0864 | 0.85 | 0.8348 | 0.1540 | 0.0434 |
No log | 58.0 | 406 | 0.5451 | 0.85 | 0.2447 | 1.0844 | 0.85 | 0.8348 | 0.1636 | 0.0436 |
No log | 59.0 | 413 | 0.5442 | 0.85 | 0.2441 | 1.0858 | 0.85 | 0.8348 | 0.1556 | 0.0435 |
No log | 60.0 | 420 | 0.5453 | 0.85 | 0.2447 | 1.0843 | 0.85 | 0.8348 | 0.1717 | 0.0437 |
No log | 61.0 | 427 | 0.5439 | 0.85 | 0.2442 | 1.0847 | 0.85 | 0.8348 | 0.1541 | 0.0432 |
No log | 62.0 | 434 | 0.5455 | 0.85 | 0.2449 | 1.0839 | 0.85 | 0.8348 | 0.1550 | 0.0435 |
No log | 63.0 | 441 | 0.5446 | 0.85 | 0.2445 | 1.0843 | 0.85 | 0.8348 | 0.1553 | 0.0435 |
No log | 64.0 | 448 | 0.5448 | 0.85 | 0.2446 | 1.0833 | 0.85 | 0.8348 | 0.1634 | 0.0435 |
No log | 65.0 | 455 | 0.5443 | 0.85 | 0.2443 | 1.0847 | 0.85 | 0.8348 | 0.1554 | 0.0435 |
No log | 66.0 | 462 | 0.5448 | 0.85 | 0.2447 | 1.0831 | 0.85 | 0.8348 | 0.1547 | 0.0436 |
No log | 67.0 | 469 | 0.5452 | 0.85 | 0.2448 | 1.0828 | 0.85 | 0.8348 | 0.1563 | 0.0436 |
No log | 68.0 | 476 | 0.5443 | 0.85 | 0.2444 | 1.0834 | 0.85 | 0.8348 | 0.1472 | 0.0434 |
No log | 69.0 | 483 | 0.5447 | 0.85 | 0.2445 | 1.0832 | 0.85 | 0.8348 | 0.1632 | 0.0434 |
No log | 70.0 | 490 | 0.5447 | 0.85 | 0.2446 | 1.0831 | 0.85 | 0.8348 | 0.1559 | 0.0435 |
No log | 71.0 | 497 | 0.5447 | 0.85 | 0.2446 | 1.0829 | 0.85 | 0.8348 | 0.1473 | 0.0435 |
0.1823 | 72.0 | 504 | 0.5443 | 0.85 | 0.2444 | 1.0828 | 0.85 | 0.8348 | 0.1559 | 0.0434 |
0.1823 | 73.0 | 511 | 0.5447 | 0.85 | 0.2447 | 1.0825 | 0.85 | 0.8348 | 0.1472 | 0.0434 |
0.1823 | 74.0 | 518 | 0.5444 | 0.85 | 0.2444 | 1.0829 | 0.85 | 0.8348 | 0.1559 | 0.0436 |
0.1823 | 75.0 | 525 | 0.5446 | 0.85 | 0.2445 | 1.0829 | 0.85 | 0.8348 | 0.1557 | 0.0435 |
0.1823 | 76.0 | 532 | 0.5448 | 0.85 | 0.2445 | 1.0825 | 0.85 | 0.8348 | 0.1559 | 0.0435 |
0.1823 | 77.0 | 539 | 0.5443 | 0.85 | 0.2444 | 1.0827 | 0.85 | 0.8348 | 0.1558 | 0.0435 |
0.1823 | 78.0 | 546 | 0.5446 | 0.85 | 0.2446 | 1.0824 | 0.85 | 0.8348 | 0.1560 | 0.0436 |
0.1823 | 79.0 | 553 | 0.5450 | 0.85 | 0.2448 | 1.0821 | 0.85 | 0.8348 | 0.1637 | 0.0436 |
0.1823 | 80.0 | 560 | 0.5447 | 0.85 | 0.2446 | 1.0823 | 0.85 | 0.8348 | 0.1638 | 0.0436 |
0.1823 | 81.0 | 567 | 0.5446 | 0.85 | 0.2446 | 1.0820 | 0.85 | 0.8348 | 0.1560 | 0.0435 |
0.1823 | 82.0 | 574 | 0.5447 | 0.85 | 0.2446 | 1.0819 | 0.85 | 0.8348 | 0.1561 | 0.0435 |
0.1823 | 83.0 | 581 | 0.5448 | 0.85 | 0.2446 | 1.0822 | 0.85 | 0.8348 | 0.1550 | 0.0436 |
0.1823 | 84.0 | 588 | 0.5445 | 0.85 | 0.2446 | 1.0819 | 0.85 | 0.8348 | 0.1551 | 0.0435 |
0.1823 | 85.0 | 595 | 0.5446 | 0.85 | 0.2446 | 1.0818 | 0.85 | 0.8348 | 0.1560 | 0.0436 |
0.1823 | 86.0 | 602 | 0.5446 | 0.85 | 0.2446 | 1.0818 | 0.85 | 0.8348 | 0.1560 | 0.0435 |
0.1823 | 87.0 | 609 | 0.5448 | 0.85 | 0.2447 | 1.0820 | 0.85 | 0.8348 | 0.1560 | 0.0435 |
0.1823 | 88.0 | 616 | 0.5447 | 0.85 | 0.2446 | 1.0819 | 0.85 | 0.8348 | 0.1551 | 0.0435 |
0.1823 | 89.0 | 623 | 0.5446 | 0.85 | 0.2446 | 1.0819 | 0.85 | 0.8348 | 0.1560 | 0.0435 |
0.1823 | 90.0 | 630 | 0.5446 | 0.85 | 0.2446 | 1.0816 | 0.85 | 0.8348 | 0.1638 | 0.0436 |
0.1823 | 91.0 | 637 | 0.5446 | 0.85 | 0.2445 | 1.0817 | 0.85 | 0.8348 | 0.1474 | 0.0435 |
0.1823 | 92.0 | 644 | 0.5445 | 0.85 | 0.2445 | 1.0818 | 0.85 | 0.8348 | 0.1551 | 0.0436 |
0.1823 | 93.0 | 651 | 0.5447 | 0.85 | 0.2446 | 1.0818 | 0.85 | 0.8348 | 0.1560 | 0.0436 |
0.1823 | 94.0 | 658 | 0.5447 | 0.85 | 0.2446 | 1.0816 | 0.85 | 0.8348 | 0.1561 | 0.0436 |
0.1823 | 95.0 | 665 | 0.5447 | 0.85 | 0.2446 | 1.0816 | 0.85 | 0.8348 | 0.1550 | 0.0435 |
0.1823 | 96.0 | 672 | 0.5446 | 0.85 | 0.2446 | 1.0816 | 0.85 | 0.8348 | 0.1474 | 0.0436 |
0.1823 | 97.0 | 679 | 0.5446 | 0.85 | 0.2446 | 1.0817 | 0.85 | 0.8348 | 0.1551 | 0.0436 |
0.1823 | 98.0 | 686 | 0.5446 | 0.85 | 0.2446 | 1.0817 | 0.85 | 0.8348 | 0.1474 | 0.0436 |
0.1823 | 99.0 | 693 | 0.5446 | 0.85 | 0.2446 | 1.0816 | 0.85 | 0.8348 | 0.1474 | 0.0436 |
0.1823 | 100.0 | 700 | 0.5446 | 0.85 | 0.2446 | 1.0816 | 0.85 | 0.8348 | 0.1474 | 0.0436 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2