<!-- 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_kd_CEKD_t5.0_a0.5
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: 1.6906
- Accuracy: 0.5675
- Brier Loss: 0.5696
- Nll: 2.4654
- F1 Micro: 0.5675
- F1 Macro: 0.5648
- Ece: 0.1638
- Aurc: 0.1969
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 | 4.6645 | 0.0425 | 1.0732 | 7.4792 | 0.0425 | 0.0329 | 0.2822 | 0.9534 |
No log | 2.0 | 14 | 3.6090 | 0.1 | 0.9462 | 5.7033 | 0.1000 | 0.0909 | 0.1715 | 0.8908 |
No log | 3.0 | 21 | 3.2938 | 0.1975 | 0.8945 | 5.2434 | 0.1975 | 0.1585 | 0.1594 | 0.6803 |
No log | 4.0 | 28 | 2.8501 | 0.29 | 0.7966 | 3.6125 | 0.29 | 0.2703 | 0.1664 | 0.4859 |
No log | 5.0 | 35 | 2.5122 | 0.39 | 0.7162 | 3.1122 | 0.39 | 0.3562 | 0.1838 | 0.3484 |
No log | 6.0 | 42 | 2.3667 | 0.4175 | 0.6955 | 3.3546 | 0.4175 | 0.3757 | 0.1683 | 0.3138 |
No log | 7.0 | 49 | 2.2629 | 0.4725 | 0.6766 | 2.9737 | 0.4725 | 0.4417 | 0.1917 | 0.2939 |
No log | 8.0 | 56 | 2.1495 | 0.485 | 0.6530 | 2.7257 | 0.485 | 0.4798 | 0.1901 | 0.2762 |
No log | 9.0 | 63 | 2.1491 | 0.495 | 0.6517 | 2.7709 | 0.495 | 0.4982 | 0.1934 | 0.2664 |
No log | 10.0 | 70 | 2.2541 | 0.465 | 0.6794 | 2.6880 | 0.465 | 0.4654 | 0.1923 | 0.2870 |
No log | 11.0 | 77 | 2.1607 | 0.4925 | 0.6564 | 2.7773 | 0.4925 | 0.4769 | 0.1904 | 0.2545 |
No log | 12.0 | 84 | 2.1581 | 0.505 | 0.6714 | 2.7017 | 0.505 | 0.4876 | 0.2023 | 0.2669 |
No log | 13.0 | 91 | 2.0025 | 0.515 | 0.6324 | 2.5557 | 0.515 | 0.5048 | 0.2194 | 0.2348 |
No log | 14.0 | 98 | 2.0267 | 0.5025 | 0.6474 | 2.4764 | 0.5025 | 0.4876 | 0.2071 | 0.2474 |
No log | 15.0 | 105 | 1.9258 | 0.535 | 0.6106 | 2.4995 | 0.535 | 0.5310 | 0.1960 | 0.2251 |
No log | 16.0 | 112 | 1.8650 | 0.545 | 0.6031 | 2.4944 | 0.545 | 0.5383 | 0.1847 | 0.2187 |
No log | 17.0 | 119 | 1.8902 | 0.5375 | 0.6118 | 2.5468 | 0.5375 | 0.5284 | 0.1852 | 0.2278 |
No log | 18.0 | 126 | 1.9187 | 0.53 | 0.6188 | 2.5580 | 0.53 | 0.5174 | 0.1792 | 0.2285 |
No log | 19.0 | 133 | 1.8399 | 0.54 | 0.6049 | 2.5246 | 0.54 | 0.5362 | 0.1758 | 0.2191 |
No log | 20.0 | 140 | 1.8912 | 0.53 | 0.6117 | 2.5403 | 0.53 | 0.5139 | 0.1819 | 0.2210 |
No log | 21.0 | 147 | 1.8930 | 0.5325 | 0.6099 | 2.6739 | 0.5325 | 0.5294 | 0.1736 | 0.2281 |
No log | 22.0 | 154 | 1.8930 | 0.5275 | 0.6070 | 2.6060 | 0.5275 | 0.5244 | 0.1832 | 0.2214 |
No log | 23.0 | 161 | 1.8596 | 0.5375 | 0.6030 | 2.5066 | 0.5375 | 0.5260 | 0.1864 | 0.2185 |
No log | 24.0 | 168 | 1.8328 | 0.5375 | 0.5942 | 2.5127 | 0.5375 | 0.5302 | 0.1807 | 0.2079 |
No log | 25.0 | 175 | 1.8069 | 0.53 | 0.6004 | 2.4747 | 0.53 | 0.5241 | 0.1910 | 0.2189 |
No log | 26.0 | 182 | 1.7937 | 0.5425 | 0.5912 | 2.5083 | 0.5425 | 0.5366 | 0.1841 | 0.2123 |
No log | 27.0 | 189 | 1.7986 | 0.55 | 0.5945 | 2.5267 | 0.55 | 0.5476 | 0.1901 | 0.2134 |
No log | 28.0 | 196 | 1.8090 | 0.5525 | 0.5918 | 2.4467 | 0.5525 | 0.5466 | 0.1765 | 0.2125 |
No log | 29.0 | 203 | 1.7911 | 0.55 | 0.5931 | 2.4536 | 0.55 | 0.5462 | 0.1619 | 0.2140 |
No log | 30.0 | 210 | 1.7702 | 0.5425 | 0.5911 | 2.5037 | 0.5425 | 0.5340 | 0.1819 | 0.2080 |
No log | 31.0 | 217 | 1.8248 | 0.53 | 0.6063 | 2.4965 | 0.53 | 0.5145 | 0.1757 | 0.2217 |
No log | 32.0 | 224 | 1.7866 | 0.545 | 0.5920 | 2.5543 | 0.545 | 0.5365 | 0.1745 | 0.2082 |
No log | 33.0 | 231 | 1.7793 | 0.555 | 0.5867 | 2.4751 | 0.555 | 0.5503 | 0.1657 | 0.2107 |
No log | 34.0 | 238 | 1.7417 | 0.5525 | 0.5799 | 2.4174 | 0.5525 | 0.5517 | 0.1778 | 0.2049 |
No log | 35.0 | 245 | 1.7557 | 0.54 | 0.5869 | 2.4065 | 0.54 | 0.5379 | 0.1640 | 0.2119 |
No log | 36.0 | 252 | 1.7567 | 0.54 | 0.5884 | 2.5031 | 0.54 | 0.5374 | 0.1519 | 0.2122 |
No log | 37.0 | 259 | 1.7306 | 0.56 | 0.5781 | 2.5313 | 0.56 | 0.5551 | 0.1906 | 0.2042 |
No log | 38.0 | 266 | 1.7615 | 0.5425 | 0.5888 | 2.4510 | 0.5425 | 0.5423 | 0.1783 | 0.2121 |
No log | 39.0 | 273 | 1.7607 | 0.555 | 0.5855 | 2.5044 | 0.555 | 0.5473 | 0.1808 | 0.2062 |
No log | 40.0 | 280 | 1.7430 | 0.5275 | 0.5840 | 2.4265 | 0.5275 | 0.5228 | 0.1822 | 0.2125 |
No log | 41.0 | 287 | 1.7515 | 0.5625 | 0.5800 | 2.5432 | 0.5625 | 0.5500 | 0.1812 | 0.2019 |
No log | 42.0 | 294 | 1.7229 | 0.5475 | 0.5778 | 2.4597 | 0.5475 | 0.5416 | 0.1656 | 0.2042 |
No log | 43.0 | 301 | 1.7566 | 0.555 | 0.5818 | 2.4305 | 0.555 | 0.5517 | 0.1521 | 0.2036 |
No log | 44.0 | 308 | 1.7093 | 0.575 | 0.5745 | 2.4087 | 0.575 | 0.5761 | 0.1678 | 0.2019 |
No log | 45.0 | 315 | 1.7319 | 0.56 | 0.5788 | 2.5339 | 0.56 | 0.5550 | 0.1647 | 0.2028 |
No log | 46.0 | 322 | 1.7331 | 0.5475 | 0.5863 | 2.4386 | 0.5475 | 0.5470 | 0.1737 | 0.2113 |
No log | 47.0 | 329 | 1.7243 | 0.56 | 0.5776 | 2.5336 | 0.56 | 0.5548 | 0.1642 | 0.2037 |
No log | 48.0 | 336 | 1.7205 | 0.5675 | 0.5817 | 2.4578 | 0.5675 | 0.5661 | 0.1800 | 0.2091 |
No log | 49.0 | 343 | 1.7210 | 0.5725 | 0.5721 | 2.4792 | 0.5725 | 0.5656 | 0.1550 | 0.1970 |
No log | 50.0 | 350 | 1.6994 | 0.58 | 0.5719 | 2.4645 | 0.58 | 0.5773 | 0.1580 | 0.1957 |
No log | 51.0 | 357 | 1.7082 | 0.5775 | 0.5737 | 2.4805 | 0.5775 | 0.5701 | 0.1734 | 0.1992 |
No log | 52.0 | 364 | 1.7144 | 0.5575 | 0.5693 | 2.4501 | 0.5575 | 0.5519 | 0.1812 | 0.1974 |
No log | 53.0 | 371 | 1.7196 | 0.57 | 0.5796 | 2.4555 | 0.57 | 0.5618 | 0.1728 | 0.2022 |
No log | 54.0 | 378 | 1.7149 | 0.5675 | 0.5746 | 2.4363 | 0.5675 | 0.5605 | 0.1901 | 0.1999 |
No log | 55.0 | 385 | 1.7030 | 0.555 | 0.5775 | 2.4253 | 0.555 | 0.5528 | 0.1823 | 0.2029 |
No log | 56.0 | 392 | 1.7209 | 0.5675 | 0.5747 | 2.4926 | 0.5675 | 0.5634 | 0.1644 | 0.2008 |
No log | 57.0 | 399 | 1.7103 | 0.5525 | 0.5760 | 2.4528 | 0.5525 | 0.5487 | 0.1738 | 0.2025 |
No log | 58.0 | 406 | 1.7005 | 0.5575 | 0.5771 | 2.4678 | 0.5575 | 0.5547 | 0.1514 | 0.2041 |
No log | 59.0 | 413 | 1.7098 | 0.56 | 0.5763 | 2.4368 | 0.56 | 0.5585 | 0.1650 | 0.2022 |
No log | 60.0 | 420 | 1.6976 | 0.5775 | 0.5681 | 2.4633 | 0.5775 | 0.5737 | 0.1729 | 0.1963 |
No log | 61.0 | 427 | 1.7057 | 0.5575 | 0.5739 | 2.4717 | 0.5575 | 0.5522 | 0.1932 | 0.2007 |
No log | 62.0 | 434 | 1.6884 | 0.5725 | 0.5646 | 2.4538 | 0.5725 | 0.5693 | 0.1692 | 0.1930 |
No log | 63.0 | 441 | 1.6979 | 0.56 | 0.5731 | 2.4635 | 0.56 | 0.5562 | 0.1690 | 0.2016 |
No log | 64.0 | 448 | 1.6848 | 0.55 | 0.5686 | 2.4583 | 0.55 | 0.5452 | 0.1782 | 0.1980 |
No log | 65.0 | 455 | 1.7072 | 0.5575 | 0.5774 | 2.4596 | 0.5575 | 0.5549 | 0.1627 | 0.2036 |
No log | 66.0 | 462 | 1.7057 | 0.5625 | 0.5759 | 2.4644 | 0.5625 | 0.5586 | 0.1780 | 0.2009 |
No log | 67.0 | 469 | 1.7016 | 0.56 | 0.5719 | 2.4976 | 0.56 | 0.5546 | 0.1734 | 0.2004 |
No log | 68.0 | 476 | 1.6951 | 0.5675 | 0.5725 | 2.4025 | 0.5675 | 0.5648 | 0.1866 | 0.1980 |
No log | 69.0 | 483 | 1.7012 | 0.555 | 0.5750 | 2.4970 | 0.555 | 0.5527 | 0.1958 | 0.2022 |
No log | 70.0 | 490 | 1.6983 | 0.575 | 0.5689 | 2.4763 | 0.575 | 0.5708 | 0.1648 | 0.1950 |
No log | 71.0 | 497 | 1.6954 | 0.5675 | 0.5729 | 2.4642 | 0.5675 | 0.5638 | 0.1762 | 0.2004 |
0.437 | 72.0 | 504 | 1.6973 | 0.5625 | 0.5718 | 2.4639 | 0.5625 | 0.5605 | 0.1680 | 0.1991 |
0.437 | 73.0 | 511 | 1.6942 | 0.565 | 0.5714 | 2.4629 | 0.565 | 0.5610 | 0.1776 | 0.1980 |
0.437 | 74.0 | 518 | 1.6923 | 0.5725 | 0.5694 | 2.4717 | 0.5725 | 0.5698 | 0.1676 | 0.1967 |
0.437 | 75.0 | 525 | 1.6926 | 0.5675 | 0.5699 | 2.4674 | 0.5675 | 0.5633 | 0.1741 | 0.1975 |
0.437 | 76.0 | 532 | 1.6906 | 0.5675 | 0.5692 | 2.4673 | 0.5675 | 0.5641 | 0.1785 | 0.1962 |
0.437 | 77.0 | 539 | 1.6912 | 0.565 | 0.5692 | 2.4671 | 0.565 | 0.5617 | 0.1568 | 0.1958 |
0.437 | 78.0 | 546 | 1.6879 | 0.565 | 0.5685 | 2.4629 | 0.565 | 0.5620 | 0.1860 | 0.1954 |
0.437 | 79.0 | 553 | 1.6886 | 0.565 | 0.5695 | 2.4650 | 0.565 | 0.5625 | 0.1777 | 0.1968 |
0.437 | 80.0 | 560 | 1.6882 | 0.5625 | 0.5700 | 2.4632 | 0.5625 | 0.5583 | 0.1791 | 0.1982 |
0.437 | 81.0 | 567 | 1.6918 | 0.565 | 0.5704 | 2.4638 | 0.565 | 0.5622 | 0.1630 | 0.1978 |
0.437 | 82.0 | 574 | 1.6909 | 0.5675 | 0.5697 | 2.4646 | 0.5675 | 0.5641 | 0.1862 | 0.1959 |
0.437 | 83.0 | 581 | 1.6881 | 0.565 | 0.5687 | 2.4665 | 0.565 | 0.5621 | 0.1494 | 0.1956 |
0.437 | 84.0 | 588 | 1.6897 | 0.565 | 0.5692 | 2.4648 | 0.565 | 0.5626 | 0.1716 | 0.1969 |
0.437 | 85.0 | 595 | 1.6910 | 0.5675 | 0.5697 | 2.4654 | 0.5675 | 0.5652 | 0.1747 | 0.1971 |
0.437 | 86.0 | 602 | 1.6905 | 0.57 | 0.5694 | 2.4648 | 0.57 | 0.5667 | 0.1659 | 0.1959 |
0.437 | 87.0 | 609 | 1.6896 | 0.5675 | 0.5693 | 2.4648 | 0.5675 | 0.5642 | 0.1681 | 0.1958 |
0.437 | 88.0 | 616 | 1.6902 | 0.5675 | 0.5695 | 2.4655 | 0.5675 | 0.5642 | 0.1709 | 0.1961 |
0.437 | 89.0 | 623 | 1.6907 | 0.5675 | 0.5697 | 2.4655 | 0.5675 | 0.5648 | 0.1676 | 0.1971 |
0.437 | 90.0 | 630 | 1.6903 | 0.5675 | 0.5694 | 2.4657 | 0.5675 | 0.5648 | 0.1699 | 0.1966 |
0.437 | 91.0 | 637 | 1.6906 | 0.565 | 0.5696 | 2.4656 | 0.565 | 0.5624 | 0.1689 | 0.1970 |
0.437 | 92.0 | 644 | 1.6904 | 0.565 | 0.5695 | 2.4654 | 0.565 | 0.5622 | 0.1668 | 0.1970 |
0.437 | 93.0 | 651 | 1.6904 | 0.5675 | 0.5696 | 2.4655 | 0.5675 | 0.5649 | 0.1619 | 0.1970 |
0.437 | 94.0 | 658 | 1.6903 | 0.565 | 0.5694 | 2.4655 | 0.565 | 0.5622 | 0.1662 | 0.1969 |
0.437 | 95.0 | 665 | 1.6905 | 0.565 | 0.5695 | 2.4653 | 0.565 | 0.5622 | 0.1710 | 0.1969 |
0.437 | 96.0 | 672 | 1.6905 | 0.5675 | 0.5695 | 2.4655 | 0.5675 | 0.5648 | 0.1638 | 0.1968 |
0.437 | 97.0 | 679 | 1.6906 | 0.565 | 0.5695 | 2.4654 | 0.565 | 0.5622 | 0.1664 | 0.1970 |
0.437 | 98.0 | 686 | 1.6906 | 0.5675 | 0.5695 | 2.4655 | 0.5675 | 0.5648 | 0.1638 | 0.1969 |
0.437 | 99.0 | 693 | 1.6906 | 0.5675 | 0.5696 | 2.4654 | 0.5675 | 0.5648 | 0.1638 | 0.1969 |
0.437 | 100.0 | 700 | 1.6906 | 0.5675 | 0.5696 | 2.4654 | 0.5675 | 0.5648 | 0.1638 | 0.1969 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2