<!-- 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_t2.5_a0.7
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: 1.7789
- Accuracy: 0.565
- Brier Loss: 0.5798
- Nll: 2.3548
- F1 Micro: 0.565
- F1 Macro: 0.5569
- Ece: 0.1677
- Aurc: 0.2032
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 | 4.7719 | 0.04 | 1.0712 | 7.3456 | 0.04 | 0.0315 | 0.2816 | 0.9543 |
No log | 2.0 | 14 | 3.7636 | 0.1 | 0.9486 | 5.7989 | 0.1000 | 0.0863 | 0.1625 | 0.8854 |
No log | 3.0 | 21 | 3.3871 | 0.235 | 0.8838 | 5.3315 | 0.235 | 0.1857 | 0.1589 | 0.6274 |
No log | 4.0 | 28 | 2.8780 | 0.2975 | 0.7900 | 3.5467 | 0.2975 | 0.2884 | 0.1672 | 0.4712 |
No log | 5.0 | 35 | 2.5223 | 0.3875 | 0.7164 | 3.0188 | 0.3875 | 0.3630 | 0.1596 | 0.3495 |
No log | 6.0 | 42 | 2.3634 | 0.43 | 0.6927 | 3.1378 | 0.4300 | 0.4031 | 0.1877 | 0.3055 |
No log | 7.0 | 49 | 2.3487 | 0.445 | 0.7009 | 3.0084 | 0.445 | 0.4123 | 0.2026 | 0.3003 |
No log | 8.0 | 56 | 2.2521 | 0.47 | 0.6857 | 2.8634 | 0.47 | 0.4607 | 0.1977 | 0.2918 |
No log | 9.0 | 63 | 2.3597 | 0.4775 | 0.6955 | 2.6768 | 0.4775 | 0.4642 | 0.2215 | 0.2787 |
No log | 10.0 | 70 | 2.3524 | 0.46 | 0.7130 | 2.7830 | 0.46 | 0.4441 | 0.2429 | 0.2941 |
No log | 11.0 | 77 | 2.4233 | 0.46 | 0.7108 | 3.0867 | 0.46 | 0.4418 | 0.2247 | 0.2828 |
No log | 12.0 | 84 | 2.2723 | 0.485 | 0.6901 | 2.7313 | 0.485 | 0.4742 | 0.2293 | 0.2774 |
No log | 13.0 | 91 | 2.2818 | 0.49 | 0.7007 | 2.5829 | 0.49 | 0.4859 | 0.2473 | 0.2791 |
No log | 14.0 | 98 | 2.1682 | 0.4975 | 0.6695 | 2.8292 | 0.4975 | 0.4815 | 0.2170 | 0.2653 |
No log | 15.0 | 105 | 2.0652 | 0.52 | 0.6520 | 2.4319 | 0.52 | 0.5120 | 0.2079 | 0.2617 |
No log | 16.0 | 112 | 2.0524 | 0.5225 | 0.6384 | 2.5273 | 0.5225 | 0.5128 | 0.1980 | 0.2391 |
No log | 17.0 | 119 | 2.1736 | 0.48 | 0.6765 | 2.4887 | 0.48 | 0.4809 | 0.2112 | 0.2834 |
No log | 18.0 | 126 | 2.0096 | 0.515 | 0.6496 | 2.4230 | 0.515 | 0.5215 | 0.1957 | 0.2513 |
No log | 19.0 | 133 | 2.0207 | 0.5175 | 0.6417 | 2.4591 | 0.5175 | 0.5104 | 0.2011 | 0.2526 |
No log | 20.0 | 140 | 1.9152 | 0.5425 | 0.6044 | 2.5069 | 0.5425 | 0.5405 | 0.1826 | 0.2147 |
No log | 21.0 | 147 | 1.9600 | 0.52 | 0.6332 | 2.5016 | 0.52 | 0.5204 | 0.1871 | 0.2427 |
No log | 22.0 | 154 | 1.9325 | 0.515 | 0.6226 | 2.4981 | 0.515 | 0.5078 | 0.1860 | 0.2323 |
No log | 23.0 | 161 | 1.9172 | 0.53 | 0.6144 | 2.4601 | 0.53 | 0.5221 | 0.1941 | 0.2311 |
No log | 24.0 | 168 | 1.8891 | 0.5425 | 0.6091 | 2.4653 | 0.5425 | 0.5399 | 0.1933 | 0.2260 |
No log | 25.0 | 175 | 1.9460 | 0.5175 | 0.6214 | 2.4785 | 0.5175 | 0.5168 | 0.1694 | 0.2335 |
No log | 26.0 | 182 | 1.9060 | 0.5525 | 0.5970 | 2.4789 | 0.5525 | 0.5430 | 0.1934 | 0.2137 |
No log | 27.0 | 189 | 1.9421 | 0.5375 | 0.6205 | 2.4814 | 0.5375 | 0.5313 | 0.2135 | 0.2402 |
No log | 28.0 | 196 | 2.0195 | 0.545 | 0.6187 | 2.5330 | 0.545 | 0.5256 | 0.1800 | 0.2360 |
No log | 29.0 | 203 | 1.9428 | 0.535 | 0.6167 | 2.4894 | 0.535 | 0.5111 | 0.1862 | 0.2322 |
No log | 30.0 | 210 | 1.8996 | 0.5225 | 0.6207 | 2.4810 | 0.5225 | 0.5137 | 0.1994 | 0.2330 |
No log | 31.0 | 217 | 1.8462 | 0.54 | 0.6119 | 2.4201 | 0.54 | 0.5341 | 0.1817 | 0.2190 |
No log | 32.0 | 224 | 1.8324 | 0.55 | 0.5988 | 2.4230 | 0.55 | 0.5427 | 0.1888 | 0.2171 |
No log | 33.0 | 231 | 1.8393 | 0.545 | 0.5977 | 2.3943 | 0.545 | 0.5339 | 0.1838 | 0.2172 |
No log | 34.0 | 238 | 1.8704 | 0.5475 | 0.6081 | 2.4488 | 0.5475 | 0.5427 | 0.1768 | 0.2200 |
No log | 35.0 | 245 | 1.8546 | 0.54 | 0.6022 | 2.3273 | 0.54 | 0.5316 | 0.1847 | 0.2226 |
No log | 36.0 | 252 | 1.8608 | 0.53 | 0.5972 | 2.5153 | 0.53 | 0.5139 | 0.1810 | 0.2202 |
No log | 37.0 | 259 | 1.8663 | 0.5325 | 0.6057 | 2.4642 | 0.5325 | 0.5243 | 0.1836 | 0.2205 |
No log | 38.0 | 266 | 1.8300 | 0.545 | 0.5954 | 2.5101 | 0.545 | 0.5418 | 0.1890 | 0.2141 |
No log | 39.0 | 273 | 1.8121 | 0.5625 | 0.5853 | 2.4397 | 0.5625 | 0.5550 | 0.1704 | 0.2110 |
No log | 40.0 | 280 | 1.7916 | 0.54 | 0.5884 | 2.3565 | 0.54 | 0.5361 | 0.1685 | 0.2135 |
No log | 41.0 | 287 | 1.8353 | 0.5575 | 0.5929 | 2.4252 | 0.5575 | 0.5451 | 0.1823 | 0.2116 |
No log | 42.0 | 294 | 1.7999 | 0.5675 | 0.5839 | 2.4820 | 0.5675 | 0.5631 | 0.1729 | 0.2045 |
No log | 43.0 | 301 | 1.8622 | 0.52 | 0.6106 | 2.4823 | 0.52 | 0.5028 | 0.1948 | 0.2270 |
No log | 44.0 | 308 | 1.7892 | 0.55 | 0.5892 | 2.3342 | 0.55 | 0.5450 | 0.1798 | 0.2126 |
No log | 45.0 | 315 | 1.7978 | 0.545 | 0.5868 | 2.4345 | 0.545 | 0.5439 | 0.1894 | 0.2094 |
No log | 46.0 | 322 | 1.7697 | 0.56 | 0.5772 | 2.4272 | 0.56 | 0.5585 | 0.1601 | 0.1997 |
No log | 47.0 | 329 | 1.7754 | 0.5475 | 0.5835 | 2.3977 | 0.5475 | 0.5438 | 0.1759 | 0.2059 |
No log | 48.0 | 336 | 1.7922 | 0.545 | 0.5929 | 2.4119 | 0.545 | 0.5390 | 0.1891 | 0.2131 |
No log | 49.0 | 343 | 1.8055 | 0.5625 | 0.5872 | 2.3654 | 0.5625 | 0.5497 | 0.1759 | 0.2073 |
No log | 50.0 | 350 | 1.7972 | 0.56 | 0.5894 | 2.3366 | 0.56 | 0.5487 | 0.1803 | 0.2083 |
No log | 51.0 | 357 | 1.7890 | 0.555 | 0.5815 | 2.3858 | 0.555 | 0.5501 | 0.1693 | 0.2067 |
No log | 52.0 | 364 | 1.7958 | 0.5475 | 0.5883 | 2.4244 | 0.5475 | 0.5355 | 0.1910 | 0.2105 |
No log | 53.0 | 371 | 1.7881 | 0.5675 | 0.5834 | 2.4135 | 0.5675 | 0.5603 | 0.1836 | 0.2028 |
No log | 54.0 | 378 | 1.7675 | 0.555 | 0.5766 | 2.4043 | 0.555 | 0.5563 | 0.1653 | 0.2047 |
No log | 55.0 | 385 | 1.7688 | 0.55 | 0.5843 | 2.3641 | 0.55 | 0.5505 | 0.1729 | 0.2092 |
No log | 56.0 | 392 | 1.7796 | 0.55 | 0.5861 | 2.3404 | 0.55 | 0.5458 | 0.1808 | 0.2114 |
No log | 57.0 | 399 | 1.7861 | 0.54 | 0.5885 | 2.3460 | 0.54 | 0.5323 | 0.1902 | 0.2073 |
No log | 58.0 | 406 | 1.7746 | 0.56 | 0.5818 | 2.3715 | 0.56 | 0.5557 | 0.1643 | 0.2034 |
No log | 59.0 | 413 | 1.7828 | 0.5575 | 0.5868 | 2.3086 | 0.5575 | 0.5526 | 0.1956 | 0.2088 |
No log | 60.0 | 420 | 1.7735 | 0.565 | 0.5825 | 2.3405 | 0.565 | 0.5619 | 0.1696 | 0.2058 |
No log | 61.0 | 427 | 1.7651 | 0.5675 | 0.5760 | 2.4771 | 0.5675 | 0.5636 | 0.1847 | 0.2027 |
No log | 62.0 | 434 | 1.7751 | 0.5575 | 0.5834 | 2.3727 | 0.5575 | 0.5524 | 0.1638 | 0.2052 |
No log | 63.0 | 441 | 1.7900 | 0.56 | 0.5834 | 2.3635 | 0.56 | 0.5502 | 0.1789 | 0.2061 |
No log | 64.0 | 448 | 1.7729 | 0.56 | 0.5821 | 2.3797 | 0.56 | 0.5554 | 0.1676 | 0.2046 |
No log | 65.0 | 455 | 1.7743 | 0.5625 | 0.5826 | 2.4174 | 0.5625 | 0.5581 | 0.1538 | 0.2052 |
No log | 66.0 | 462 | 1.7749 | 0.5625 | 0.5801 | 2.3799 | 0.5625 | 0.5592 | 0.1709 | 0.2036 |
No log | 67.0 | 469 | 1.7795 | 0.5625 | 0.5814 | 2.3169 | 0.5625 | 0.5533 | 0.1883 | 0.2037 |
No log | 68.0 | 476 | 1.7773 | 0.5675 | 0.5794 | 2.3588 | 0.5675 | 0.5622 | 0.1779 | 0.2013 |
No log | 69.0 | 483 | 1.7762 | 0.56 | 0.5793 | 2.3514 | 0.56 | 0.5566 | 0.1707 | 0.2039 |
No log | 70.0 | 490 | 1.7762 | 0.5625 | 0.5787 | 2.3620 | 0.5625 | 0.5529 | 0.1607 | 0.2017 |
No log | 71.0 | 497 | 1.7740 | 0.5675 | 0.5798 | 2.3235 | 0.5675 | 0.5612 | 0.1637 | 0.2046 |
0.4215 | 72.0 | 504 | 1.7739 | 0.56 | 0.5790 | 2.3235 | 0.56 | 0.5542 | 0.1583 | 0.2023 |
0.4215 | 73.0 | 511 | 1.7783 | 0.56 | 0.5806 | 2.4187 | 0.56 | 0.5545 | 0.1674 | 0.2040 |
0.4215 | 74.0 | 518 | 1.7785 | 0.56 | 0.5805 | 2.3302 | 0.56 | 0.5544 | 0.1748 | 0.2033 |
0.4215 | 75.0 | 525 | 1.7777 | 0.5625 | 0.5795 | 2.3321 | 0.5625 | 0.5548 | 0.1754 | 0.2029 |
0.4215 | 76.0 | 532 | 1.7785 | 0.565 | 0.5799 | 2.3249 | 0.565 | 0.5586 | 0.1696 | 0.2023 |
0.4215 | 77.0 | 539 | 1.7763 | 0.565 | 0.5790 | 2.3561 | 0.565 | 0.5573 | 0.1574 | 0.2022 |
0.4215 | 78.0 | 546 | 1.7767 | 0.565 | 0.5790 | 2.3296 | 0.565 | 0.5572 | 0.1633 | 0.2024 |
0.4215 | 79.0 | 553 | 1.7763 | 0.565 | 0.5790 | 2.3555 | 0.565 | 0.5580 | 0.1687 | 0.2016 |
0.4215 | 80.0 | 560 | 1.7783 | 0.565 | 0.5800 | 2.3254 | 0.565 | 0.5576 | 0.1752 | 0.2032 |
0.4215 | 81.0 | 567 | 1.7773 | 0.5675 | 0.5796 | 2.3530 | 0.5675 | 0.5605 | 0.1519 | 0.2023 |
0.4215 | 82.0 | 574 | 1.7774 | 0.5625 | 0.5797 | 2.3253 | 0.5625 | 0.5549 | 0.1911 | 0.2028 |
0.4215 | 83.0 | 581 | 1.7784 | 0.5625 | 0.5794 | 2.3554 | 0.5625 | 0.5544 | 0.1659 | 0.2030 |
0.4215 | 84.0 | 588 | 1.7769 | 0.565 | 0.5793 | 2.3527 | 0.565 | 0.5585 | 0.1588 | 0.2024 |
0.4215 | 85.0 | 595 | 1.7787 | 0.565 | 0.5799 | 2.3549 | 0.565 | 0.5576 | 0.1687 | 0.2032 |
0.4215 | 86.0 | 602 | 1.7778 | 0.565 | 0.5795 | 2.3548 | 0.565 | 0.5574 | 0.1577 | 0.2029 |
0.4215 | 87.0 | 609 | 1.7787 | 0.5625 | 0.5798 | 2.3545 | 0.5625 | 0.5549 | 0.1643 | 0.2032 |
0.4215 | 88.0 | 616 | 1.7786 | 0.565 | 0.5796 | 2.3554 | 0.565 | 0.5574 | 0.1667 | 0.2031 |
0.4215 | 89.0 | 623 | 1.7785 | 0.565 | 0.5799 | 2.3546 | 0.565 | 0.5574 | 0.1691 | 0.2032 |
0.4215 | 90.0 | 630 | 1.7784 | 0.565 | 0.5797 | 2.3548 | 0.565 | 0.5574 | 0.1656 | 0.2031 |
0.4215 | 91.0 | 637 | 1.7784 | 0.565 | 0.5797 | 2.3550 | 0.565 | 0.5569 | 0.1753 | 0.2032 |
0.4215 | 92.0 | 644 | 1.7786 | 0.565 | 0.5797 | 2.3545 | 0.565 | 0.5574 | 0.1744 | 0.2030 |
0.4215 | 93.0 | 651 | 1.7785 | 0.565 | 0.5797 | 2.3545 | 0.565 | 0.5574 | 0.1709 | 0.2031 |
0.4215 | 94.0 | 658 | 1.7787 | 0.565 | 0.5797 | 2.3543 | 0.565 | 0.5574 | 0.1704 | 0.2032 |
0.4215 | 95.0 | 665 | 1.7787 | 0.565 | 0.5798 | 2.3545 | 0.565 | 0.5574 | 0.1713 | 0.2032 |
0.4215 | 96.0 | 672 | 1.7788 | 0.565 | 0.5798 | 2.3549 | 0.565 | 0.5569 | 0.1777 | 0.2031 |
0.4215 | 97.0 | 679 | 1.7789 | 0.565 | 0.5798 | 2.3550 | 0.565 | 0.5569 | 0.1677 | 0.2032 |
0.4215 | 98.0 | 686 | 1.7789 | 0.565 | 0.5798 | 2.3549 | 0.565 | 0.5569 | 0.1648 | 0.2032 |
0.4215 | 99.0 | 693 | 1.7789 | 0.565 | 0.5798 | 2.3548 | 0.565 | 0.5569 | 0.1728 | 0.2032 |
0.4215 | 100.0 | 700 | 1.7789 | 0.565 | 0.5798 | 2.3548 | 0.565 | 0.5569 | 0.1677 | 0.2032 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2