<!-- 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-tiny_rvl_cdip_100_examples_per_class
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: 1.1181
- Accuracy: 0.555
- Brier Loss: 0.5639
- Nll: 2.5058
- F1 Micro: 0.555
- F1 Macro: 0.5524
- Ece: 0.1705
- Aurc: 0.2031
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 | 3.4972 | 0.045 | 1.0713 | 7.2880 | 0.045 | 0.0370 | 0.2764 | 0.9554 |
No log | 2.0 | 14 | 2.5557 | 0.0925 | 0.9501 | 5.9529 | 0.0925 | 0.0781 | 0.1601 | 0.8830 |
No log | 3.0 | 21 | 2.2287 | 0.215 | 0.8879 | 5.6730 | 0.2150 | 0.1651 | 0.1774 | 0.6393 |
No log | 4.0 | 28 | 1.8874 | 0.29 | 0.8001 | 3.7085 | 0.29 | 0.2600 | 0.1619 | 0.4897 |
No log | 5.0 | 35 | 1.6351 | 0.375 | 0.7226 | 3.1565 | 0.375 | 0.3592 | 0.1647 | 0.3632 |
No log | 6.0 | 42 | 1.5371 | 0.4275 | 0.6865 | 3.0837 | 0.4275 | 0.4211 | 0.1661 | 0.3123 |
No log | 7.0 | 49 | 1.5664 | 0.4575 | 0.6933 | 2.9854 | 0.4575 | 0.4296 | 0.2078 | 0.2973 |
No log | 8.0 | 56 | 1.5244 | 0.44 | 0.6953 | 2.7931 | 0.44 | 0.4374 | 0.2067 | 0.3131 |
No log | 9.0 | 63 | 1.5573 | 0.47 | 0.6933 | 2.7079 | 0.47 | 0.4583 | 0.2023 | 0.2977 |
No log | 10.0 | 70 | 1.6359 | 0.455 | 0.7367 | 2.7425 | 0.455 | 0.4251 | 0.2519 | 0.3015 |
No log | 11.0 | 77 | 1.6207 | 0.465 | 0.7274 | 2.6891 | 0.465 | 0.4426 | 0.2454 | 0.3086 |
No log | 12.0 | 84 | 1.6268 | 0.4775 | 0.6996 | 2.7118 | 0.4775 | 0.4638 | 0.2370 | 0.2887 |
No log | 13.0 | 91 | 1.6162 | 0.4725 | 0.7065 | 2.9655 | 0.4725 | 0.4522 | 0.2355 | 0.2868 |
No log | 14.0 | 98 | 1.6040 | 0.4875 | 0.6796 | 2.8589 | 0.4875 | 0.4736 | 0.2148 | 0.2755 |
No log | 15.0 | 105 | 1.6294 | 0.4825 | 0.7123 | 2.8243 | 0.4825 | 0.4714 | 0.2390 | 0.2867 |
No log | 16.0 | 112 | 1.5734 | 0.49 | 0.6894 | 2.7436 | 0.49 | 0.4731 | 0.2306 | 0.2652 |
No log | 17.0 | 119 | 1.5915 | 0.465 | 0.7045 | 2.6023 | 0.465 | 0.4547 | 0.2404 | 0.2905 |
No log | 18.0 | 126 | 1.4637 | 0.4875 | 0.6723 | 2.7700 | 0.4875 | 0.4637 | 0.2162 | 0.2644 |
No log | 19.0 | 133 | 1.4599 | 0.48 | 0.6739 | 2.5843 | 0.48 | 0.4625 | 0.2375 | 0.2671 |
No log | 20.0 | 140 | 1.5233 | 0.4925 | 0.6821 | 2.8319 | 0.4925 | 0.4759 | 0.2331 | 0.2620 |
No log | 21.0 | 147 | 1.4234 | 0.5025 | 0.6644 | 2.5679 | 0.5025 | 0.5051 | 0.2214 | 0.2746 |
No log | 22.0 | 154 | 1.5117 | 0.5 | 0.6770 | 2.6852 | 0.5 | 0.4805 | 0.2201 | 0.2662 |
No log | 23.0 | 161 | 1.4680 | 0.4875 | 0.6798 | 2.5928 | 0.4875 | 0.4799 | 0.1973 | 0.2714 |
No log | 24.0 | 168 | 1.4762 | 0.4825 | 0.6827 | 2.7070 | 0.4825 | 0.4690 | 0.2411 | 0.2845 |
No log | 25.0 | 175 | 1.4672 | 0.4875 | 0.6778 | 2.6189 | 0.4875 | 0.4746 | 0.2272 | 0.2733 |
No log | 26.0 | 182 | 1.5126 | 0.5075 | 0.6671 | 2.8784 | 0.5075 | 0.5018 | 0.2404 | 0.2523 |
No log | 27.0 | 189 | 1.4648 | 0.495 | 0.6722 | 2.6680 | 0.495 | 0.4959 | 0.2250 | 0.2773 |
No log | 28.0 | 196 | 1.4679 | 0.5175 | 0.6463 | 2.7459 | 0.5175 | 0.4851 | 0.2134 | 0.2377 |
No log | 29.0 | 203 | 1.4982 | 0.475 | 0.6930 | 2.7531 | 0.4750 | 0.4666 | 0.2446 | 0.2901 |
No log | 30.0 | 210 | 1.3873 | 0.5075 | 0.6346 | 2.5718 | 0.5075 | 0.4857 | 0.1991 | 0.2456 |
No log | 31.0 | 217 | 1.2963 | 0.54 | 0.6140 | 2.6319 | 0.54 | 0.5431 | 0.1806 | 0.2316 |
No log | 32.0 | 224 | 1.3467 | 0.505 | 0.6361 | 2.7359 | 0.505 | 0.4792 | 0.2133 | 0.2409 |
No log | 33.0 | 231 | 1.2819 | 0.55 | 0.6017 | 2.6009 | 0.55 | 0.5423 | 0.1700 | 0.2182 |
No log | 34.0 | 238 | 1.3502 | 0.525 | 0.6402 | 2.4999 | 0.525 | 0.5166 | 0.1945 | 0.2492 |
No log | 35.0 | 245 | 1.3020 | 0.5275 | 0.6155 | 2.6222 | 0.5275 | 0.5173 | 0.1844 | 0.2243 |
No log | 36.0 | 252 | 1.4007 | 0.5175 | 0.6457 | 2.5934 | 0.5175 | 0.5160 | 0.2112 | 0.2588 |
No log | 37.0 | 259 | 1.2637 | 0.5175 | 0.6187 | 2.6172 | 0.5175 | 0.5031 | 0.1879 | 0.2314 |
No log | 38.0 | 266 | 1.3416 | 0.54 | 0.6142 | 2.6727 | 0.54 | 0.5300 | 0.1827 | 0.2262 |
No log | 39.0 | 273 | 1.2449 | 0.5375 | 0.6086 | 2.6247 | 0.5375 | 0.5359 | 0.1847 | 0.2294 |
No log | 40.0 | 280 | 1.2930 | 0.5325 | 0.6131 | 2.7111 | 0.5325 | 0.5141 | 0.1865 | 0.2262 |
No log | 41.0 | 287 | 1.2288 | 0.535 | 0.5988 | 2.5900 | 0.535 | 0.5277 | 0.1827 | 0.2146 |
No log | 42.0 | 294 | 1.2507 | 0.545 | 0.6031 | 2.5254 | 0.545 | 0.5289 | 0.1754 | 0.2224 |
No log | 43.0 | 301 | 1.2154 | 0.5475 | 0.5986 | 2.6049 | 0.5475 | 0.5414 | 0.1730 | 0.2229 |
No log | 44.0 | 308 | 1.1948 | 0.5575 | 0.5824 | 2.5753 | 0.5575 | 0.5492 | 0.1904 | 0.2047 |
No log | 45.0 | 315 | 1.1773 | 0.555 | 0.5846 | 2.6682 | 0.555 | 0.5564 | 0.1918 | 0.2171 |
No log | 46.0 | 322 | 1.1860 | 0.56 | 0.5763 | 2.5625 | 0.56 | 0.5549 | 0.1951 | 0.2018 |
No log | 47.0 | 329 | 1.1970 | 0.5575 | 0.5850 | 2.5154 | 0.5575 | 0.5436 | 0.1819 | 0.2114 |
No log | 48.0 | 336 | 1.1805 | 0.5425 | 0.5862 | 2.5828 | 0.5425 | 0.5423 | 0.1885 | 0.2132 |
No log | 49.0 | 343 | 1.1902 | 0.555 | 0.5862 | 2.5043 | 0.555 | 0.5446 | 0.1705 | 0.2075 |
No log | 50.0 | 350 | 1.1496 | 0.5725 | 0.5748 | 2.5429 | 0.5725 | 0.5752 | 0.1647 | 0.2034 |
No log | 51.0 | 357 | 1.1749 | 0.57 | 0.5802 | 2.5185 | 0.57 | 0.5564 | 0.1713 | 0.2055 |
No log | 52.0 | 364 | 1.1485 | 0.5625 | 0.5717 | 2.5957 | 0.5625 | 0.5570 | 0.1752 | 0.1993 |
No log | 53.0 | 371 | 1.1721 | 0.5525 | 0.5853 | 2.6264 | 0.5525 | 0.5488 | 0.1632 | 0.2165 |
No log | 54.0 | 378 | 1.1592 | 0.5525 | 0.5783 | 2.5631 | 0.5525 | 0.5433 | 0.1743 | 0.2057 |
No log | 55.0 | 385 | 1.1473 | 0.5725 | 0.5731 | 2.6200 | 0.5725 | 0.5689 | 0.1667 | 0.2047 |
No log | 56.0 | 392 | 1.1643 | 0.5575 | 0.5771 | 2.5922 | 0.5575 | 0.5458 | 0.1699 | 0.2098 |
No log | 57.0 | 399 | 1.1510 | 0.5775 | 0.5740 | 2.5202 | 0.5775 | 0.5798 | 0.1633 | 0.2042 |
No log | 58.0 | 406 | 1.1627 | 0.5725 | 0.5755 | 2.5974 | 0.5725 | 0.5619 | 0.1685 | 0.2053 |
No log | 59.0 | 413 | 1.1524 | 0.56 | 0.5806 | 2.5180 | 0.56 | 0.5581 | 0.1826 | 0.2119 |
No log | 60.0 | 420 | 1.1526 | 0.5675 | 0.5712 | 2.6118 | 0.5675 | 0.5624 | 0.1714 | 0.2025 |
No log | 61.0 | 427 | 1.1861 | 0.5525 | 0.5845 | 2.5332 | 0.5525 | 0.5429 | 0.1634 | 0.2111 |
No log | 62.0 | 434 | 1.1313 | 0.5675 | 0.5648 | 2.5085 | 0.5675 | 0.5690 | 0.1589 | 0.1952 |
No log | 63.0 | 441 | 1.1651 | 0.5625 | 0.5796 | 2.5044 | 0.5625 | 0.5495 | 0.1666 | 0.2078 |
No log | 64.0 | 448 | 1.1352 | 0.565 | 0.5689 | 2.4902 | 0.565 | 0.5652 | 0.1703 | 0.2032 |
No log | 65.0 | 455 | 1.1379 | 0.5725 | 0.5723 | 2.5435 | 0.5725 | 0.5666 | 0.1593 | 0.2026 |
No log | 66.0 | 462 | 1.1507 | 0.575 | 0.5763 | 2.4687 | 0.575 | 0.5697 | 0.1551 | 0.2056 |
No log | 67.0 | 469 | 1.1354 | 0.5625 | 0.5708 | 2.5495 | 0.5625 | 0.5564 | 0.1754 | 0.2009 |
No log | 68.0 | 476 | 1.1286 | 0.5775 | 0.5671 | 2.4870 | 0.5775 | 0.5746 | 0.1795 | 0.2008 |
No log | 69.0 | 483 | 1.1371 | 0.5675 | 0.5701 | 2.5777 | 0.5675 | 0.5642 | 0.1769 | 0.2047 |
No log | 70.0 | 490 | 1.1152 | 0.56 | 0.5613 | 2.5036 | 0.56 | 0.5599 | 0.1370 | 0.1985 |
No log | 71.0 | 497 | 1.1289 | 0.5725 | 0.5657 | 2.5572 | 0.5725 | 0.5666 | 0.1496 | 0.2014 |
0.2584 | 72.0 | 504 | 1.1270 | 0.57 | 0.5653 | 2.5292 | 0.57 | 0.5688 | 0.1570 | 0.1973 |
0.2584 | 73.0 | 511 | 1.1297 | 0.57 | 0.5680 | 2.5565 | 0.57 | 0.5649 | 0.1467 | 0.2042 |
0.2584 | 74.0 | 518 | 1.1246 | 0.5625 | 0.5646 | 2.5033 | 0.5625 | 0.5618 | 0.1581 | 0.2004 |
0.2584 | 75.0 | 525 | 1.1186 | 0.57 | 0.5635 | 2.5465 | 0.57 | 0.5671 | 0.1454 | 0.1999 |
0.2584 | 76.0 | 532 | 1.1210 | 0.56 | 0.5654 | 2.5094 | 0.56 | 0.5587 | 0.1510 | 0.2031 |
0.2584 | 77.0 | 539 | 1.1212 | 0.5675 | 0.5635 | 2.5170 | 0.5675 | 0.5630 | 0.1631 | 0.1996 |
0.2584 | 78.0 | 546 | 1.1190 | 0.56 | 0.5642 | 2.5074 | 0.56 | 0.5592 | 0.1506 | 0.2027 |
0.2584 | 79.0 | 553 | 1.1215 | 0.5625 | 0.5643 | 2.5112 | 0.5625 | 0.5599 | 0.1573 | 0.2024 |
0.2584 | 80.0 | 560 | 1.1181 | 0.56 | 0.5635 | 2.5064 | 0.56 | 0.5595 | 0.1601 | 0.2009 |
0.2584 | 81.0 | 567 | 1.1201 | 0.5675 | 0.5639 | 2.5096 | 0.5675 | 0.5669 | 0.1602 | 0.2008 |
0.2584 | 82.0 | 574 | 1.1195 | 0.5625 | 0.5645 | 2.5054 | 0.5625 | 0.5610 | 0.1759 | 0.2025 |
0.2584 | 83.0 | 581 | 1.1182 | 0.5625 | 0.5641 | 2.5062 | 0.5625 | 0.5619 | 0.1830 | 0.2018 |
0.2584 | 84.0 | 588 | 1.1195 | 0.5575 | 0.5637 | 2.5121 | 0.5575 | 0.5556 | 0.1838 | 0.2026 |
0.2584 | 85.0 | 595 | 1.1192 | 0.56 | 0.5641 | 2.5058 | 0.56 | 0.5588 | 0.1716 | 0.2026 |
0.2584 | 86.0 | 602 | 1.1186 | 0.5625 | 0.5639 | 2.5060 | 0.5625 | 0.5619 | 0.1662 | 0.2022 |
0.2584 | 87.0 | 609 | 1.1181 | 0.5575 | 0.5637 | 2.5070 | 0.5575 | 0.5557 | 0.1596 | 0.2027 |
0.2584 | 88.0 | 616 | 1.1178 | 0.56 | 0.5637 | 2.5060 | 0.56 | 0.5577 | 0.1655 | 0.2017 |
0.2584 | 89.0 | 623 | 1.1183 | 0.56 | 0.5639 | 2.5057 | 0.56 | 0.5580 | 0.1542 | 0.2021 |
0.2584 | 90.0 | 630 | 1.1184 | 0.56 | 0.5640 | 2.5060 | 0.56 | 0.5581 | 0.1841 | 0.2021 |
0.2584 | 91.0 | 637 | 1.1183 | 0.5575 | 0.5638 | 2.5060 | 0.5575 | 0.5547 | 0.1672 | 0.2029 |
0.2584 | 92.0 | 644 | 1.1182 | 0.5575 | 0.5638 | 2.5059 | 0.5575 | 0.5547 | 0.1608 | 0.2029 |
0.2584 | 93.0 | 651 | 1.1183 | 0.56 | 0.5639 | 2.5057 | 0.56 | 0.5576 | 0.1648 | 0.2021 |
0.2584 | 94.0 | 658 | 1.1185 | 0.555 | 0.5639 | 2.5061 | 0.555 | 0.5524 | 0.1648 | 0.2031 |
0.2584 | 95.0 | 665 | 1.1183 | 0.555 | 0.5640 | 2.5058 | 0.555 | 0.5524 | 0.1614 | 0.2031 |
0.2584 | 96.0 | 672 | 1.1182 | 0.555 | 0.5639 | 2.5057 | 0.555 | 0.5524 | 0.1769 | 0.2030 |
0.2584 | 97.0 | 679 | 1.1182 | 0.555 | 0.5639 | 2.5057 | 0.555 | 0.5524 | 0.1733 | 0.2031 |
0.2584 | 98.0 | 686 | 1.1181 | 0.555 | 0.5639 | 2.5058 | 0.555 | 0.5524 | 0.1754 | 0.2031 |
0.2584 | 99.0 | 693 | 1.1181 | 0.555 | 0.5639 | 2.5058 | 0.555 | 0.5524 | 0.1705 | 0.2031 |
0.2584 | 100.0 | 700 | 1.1181 | 0.555 | 0.5639 | 2.5058 | 0.555 | 0.5524 | 0.1705 | 0.2031 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2