<!-- 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_a0.5
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: 1.9682
- Accuracy: 0.5575
- Brier Loss: 0.5680
- Nll: 2.3526
- F1 Micro: 0.5575
- F1 Macro: 0.5516
- Ece: 0.1676
- Aurc: 0.1973
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 | 5.4578 | 0.045 | 1.0710 | 7.3211 | 0.045 | 0.0342 | 0.2911 | 0.9529 |
No log | 2.0 | 14 | 4.3718 | 0.0975 | 0.9466 | 5.6031 | 0.0975 | 0.0893 | 0.1665 | 0.8820 |
No log | 3.0 | 21 | 3.9389 | 0.2325 | 0.8820 | 5.3266 | 0.2325 | 0.1837 | 0.1656 | 0.6302 |
No log | 4.0 | 28 | 3.3251 | 0.3075 | 0.7848 | 3.6819 | 0.3075 | 0.2963 | 0.1644 | 0.4646 |
No log | 5.0 | 35 | 2.8980 | 0.4025 | 0.7180 | 2.9699 | 0.4025 | 0.3740 | 0.1867 | 0.3388 |
No log | 6.0 | 42 | 2.7055 | 0.4375 | 0.7010 | 3.2343 | 0.4375 | 0.4029 | 0.2096 | 0.3042 |
No log | 7.0 | 49 | 2.6466 | 0.4575 | 0.7040 | 2.9977 | 0.4575 | 0.4272 | 0.2082 | 0.2999 |
No log | 8.0 | 56 | 2.5399 | 0.4775 | 0.6809 | 2.8098 | 0.4775 | 0.4693 | 0.2106 | 0.2837 |
No log | 9.0 | 63 | 2.5949 | 0.49 | 0.6824 | 2.6503 | 0.49 | 0.4827 | 0.2289 | 0.2718 |
No log | 10.0 | 70 | 2.6997 | 0.4725 | 0.7318 | 2.8289 | 0.4725 | 0.4523 | 0.2536 | 0.3056 |
No log | 11.0 | 77 | 2.5694 | 0.47 | 0.7040 | 2.7969 | 0.47 | 0.4534 | 0.2348 | 0.2766 |
No log | 12.0 | 84 | 2.4539 | 0.4975 | 0.6762 | 2.6929 | 0.4975 | 0.4831 | 0.2300 | 0.2664 |
No log | 13.0 | 91 | 2.4841 | 0.5025 | 0.6664 | 2.6140 | 0.5025 | 0.4944 | 0.2098 | 0.2584 |
No log | 14.0 | 98 | 2.2755 | 0.535 | 0.6410 | 2.4991 | 0.535 | 0.5249 | 0.1905 | 0.2405 |
No log | 15.0 | 105 | 2.2998 | 0.5125 | 0.6282 | 2.6561 | 0.5125 | 0.4982 | 0.1916 | 0.2371 |
No log | 16.0 | 112 | 2.2156 | 0.525 | 0.6195 | 2.4837 | 0.525 | 0.5191 | 0.1870 | 0.2232 |
No log | 17.0 | 119 | 2.1862 | 0.5225 | 0.6096 | 2.7252 | 0.5225 | 0.5202 | 0.1747 | 0.2292 |
No log | 18.0 | 126 | 2.2058 | 0.5375 | 0.6160 | 2.5446 | 0.5375 | 0.5154 | 0.2011 | 0.2231 |
No log | 19.0 | 133 | 2.2147 | 0.5375 | 0.6143 | 2.5720 | 0.5375 | 0.5232 | 0.2028 | 0.2221 |
No log | 20.0 | 140 | 2.1791 | 0.525 | 0.6191 | 2.4505 | 0.525 | 0.5107 | 0.1918 | 0.2233 |
No log | 21.0 | 147 | 2.1165 | 0.535 | 0.5960 | 2.5369 | 0.535 | 0.5280 | 0.1867 | 0.2133 |
No log | 22.0 | 154 | 2.1193 | 0.54 | 0.6009 | 2.5568 | 0.54 | 0.5313 | 0.2049 | 0.2184 |
No log | 23.0 | 161 | 2.1082 | 0.5425 | 0.5929 | 2.5238 | 0.5425 | 0.5360 | 0.1691 | 0.2099 |
No log | 24.0 | 168 | 2.1221 | 0.535 | 0.6115 | 2.4854 | 0.535 | 0.5252 | 0.1779 | 0.2234 |
No log | 25.0 | 175 | 2.1912 | 0.52 | 0.6295 | 2.4975 | 0.52 | 0.5109 | 0.1970 | 0.2339 |
No log | 26.0 | 182 | 2.1056 | 0.5225 | 0.6150 | 2.4697 | 0.5225 | 0.5250 | 0.2020 | 0.2346 |
No log | 27.0 | 189 | 2.1017 | 0.535 | 0.6027 | 2.4992 | 0.535 | 0.5399 | 0.2003 | 0.2173 |
No log | 28.0 | 196 | 2.0999 | 0.545 | 0.5929 | 2.6313 | 0.545 | 0.5306 | 0.1844 | 0.2126 |
No log | 29.0 | 203 | 2.1188 | 0.54 | 0.6044 | 2.5420 | 0.54 | 0.5211 | 0.1745 | 0.2159 |
No log | 30.0 | 210 | 2.0670 | 0.56 | 0.5938 | 2.4868 | 0.56 | 0.5500 | 0.1849 | 0.2132 |
No log | 31.0 | 217 | 2.0709 | 0.5525 | 0.5937 | 2.4206 | 0.5525 | 0.5489 | 0.1759 | 0.2140 |
No log | 32.0 | 224 | 2.0390 | 0.5675 | 0.5794 | 2.5007 | 0.5675 | 0.5547 | 0.1745 | 0.2032 |
No log | 33.0 | 231 | 2.0725 | 0.5375 | 0.5912 | 2.4328 | 0.5375 | 0.5237 | 0.1817 | 0.2157 |
No log | 34.0 | 238 | 2.0644 | 0.565 | 0.5950 | 2.5103 | 0.565 | 0.5560 | 0.1712 | 0.2085 |
No log | 35.0 | 245 | 2.0665 | 0.5575 | 0.5918 | 2.4401 | 0.5575 | 0.5519 | 0.1604 | 0.2133 |
No log | 36.0 | 252 | 2.0485 | 0.53 | 0.5977 | 2.4321 | 0.53 | 0.5259 | 0.1927 | 0.2163 |
No log | 37.0 | 259 | 2.0352 | 0.555 | 0.5780 | 2.4665 | 0.555 | 0.5450 | 0.1701 | 0.2000 |
No log | 38.0 | 266 | 2.0423 | 0.5425 | 0.5777 | 2.5061 | 0.5425 | 0.5327 | 0.1689 | 0.2021 |
No log | 39.0 | 273 | 2.0138 | 0.5625 | 0.5799 | 2.3714 | 0.5625 | 0.5581 | 0.1705 | 0.2060 |
No log | 40.0 | 280 | 2.0161 | 0.5525 | 0.5918 | 2.3704 | 0.5525 | 0.5439 | 0.1720 | 0.2111 |
No log | 41.0 | 287 | 2.0315 | 0.545 | 0.5843 | 2.4582 | 0.545 | 0.5355 | 0.1782 | 0.2083 |
No log | 42.0 | 294 | 2.0157 | 0.545 | 0.5861 | 2.5374 | 0.545 | 0.5386 | 0.1995 | 0.2129 |
No log | 43.0 | 301 | 2.0495 | 0.555 | 0.5922 | 2.4841 | 0.555 | 0.5393 | 0.1538 | 0.2138 |
No log | 44.0 | 308 | 2.0293 | 0.5525 | 0.5824 | 2.4853 | 0.5525 | 0.5352 | 0.1745 | 0.2042 |
No log | 45.0 | 315 | 2.0253 | 0.5575 | 0.5776 | 2.4516 | 0.5575 | 0.5421 | 0.1978 | 0.2045 |
No log | 46.0 | 322 | 2.0246 | 0.5525 | 0.5953 | 2.4196 | 0.5525 | 0.5362 | 0.1715 | 0.2122 |
No log | 47.0 | 329 | 2.0114 | 0.555 | 0.5730 | 2.4431 | 0.555 | 0.5462 | 0.1759 | 0.2008 |
No log | 48.0 | 336 | 2.0046 | 0.5575 | 0.5801 | 2.3784 | 0.5575 | 0.5445 | 0.1703 | 0.2039 |
No log | 49.0 | 343 | 1.9721 | 0.565 | 0.5672 | 2.5034 | 0.565 | 0.5594 | 0.1686 | 0.1963 |
No log | 50.0 | 350 | 1.9872 | 0.565 | 0.5704 | 2.4067 | 0.565 | 0.5620 | 0.1888 | 0.2000 |
No log | 51.0 | 357 | 1.9668 | 0.5725 | 0.5695 | 2.3935 | 0.5725 | 0.5711 | 0.1534 | 0.1981 |
No log | 52.0 | 364 | 1.9796 | 0.5525 | 0.5742 | 2.3977 | 0.5525 | 0.5504 | 0.1614 | 0.2023 |
No log | 53.0 | 371 | 2.0086 | 0.56 | 0.5835 | 2.4361 | 0.56 | 0.5510 | 0.1912 | 0.2098 |
No log | 54.0 | 378 | 1.9998 | 0.54 | 0.5776 | 2.4292 | 0.54 | 0.5270 | 0.1679 | 0.2042 |
No log | 55.0 | 385 | 1.9736 | 0.555 | 0.5732 | 2.3619 | 0.555 | 0.5427 | 0.1830 | 0.2024 |
No log | 56.0 | 392 | 1.9850 | 0.57 | 0.5691 | 2.4491 | 0.57 | 0.5670 | 0.1798 | 0.1967 |
No log | 57.0 | 399 | 1.9680 | 0.5675 | 0.5680 | 2.4863 | 0.5675 | 0.5624 | 0.1644 | 0.1967 |
No log | 58.0 | 406 | 1.9633 | 0.565 | 0.5682 | 2.4407 | 0.565 | 0.5573 | 0.1576 | 0.1960 |
No log | 59.0 | 413 | 1.9715 | 0.555 | 0.5750 | 2.4037 | 0.555 | 0.5459 | 0.1753 | 0.2027 |
No log | 60.0 | 420 | 1.9791 | 0.575 | 0.5684 | 2.3886 | 0.575 | 0.5732 | 0.1647 | 0.1969 |
No log | 61.0 | 427 | 1.9587 | 0.5675 | 0.5669 | 2.4318 | 0.5675 | 0.5616 | 0.1583 | 0.1948 |
No log | 62.0 | 434 | 1.9735 | 0.5575 | 0.5699 | 2.3758 | 0.5575 | 0.5550 | 0.1442 | 0.1978 |
No log | 63.0 | 441 | 1.9720 | 0.5475 | 0.5748 | 2.4373 | 0.5475 | 0.5405 | 0.1660 | 0.2004 |
No log | 64.0 | 448 | 1.9825 | 0.55 | 0.5710 | 2.3517 | 0.55 | 0.5490 | 0.1810 | 0.2030 |
No log | 65.0 | 455 | 1.9679 | 0.5725 | 0.5705 | 2.4311 | 0.5725 | 0.5718 | 0.1690 | 0.1980 |
No log | 66.0 | 462 | 1.9735 | 0.565 | 0.5706 | 2.4810 | 0.565 | 0.5624 | 0.1740 | 0.1976 |
No log | 67.0 | 469 | 1.9752 | 0.5675 | 0.5689 | 2.3718 | 0.5675 | 0.5615 | 0.1681 | 0.1980 |
No log | 68.0 | 476 | 1.9798 | 0.565 | 0.5690 | 2.4235 | 0.565 | 0.5602 | 0.1671 | 0.1980 |
No log | 69.0 | 483 | 1.9720 | 0.5625 | 0.5681 | 2.3483 | 0.5625 | 0.5576 | 0.1845 | 0.1970 |
No log | 70.0 | 490 | 1.9640 | 0.5675 | 0.5673 | 2.3772 | 0.5675 | 0.5639 | 0.1621 | 0.1962 |
No log | 71.0 | 497 | 1.9641 | 0.5625 | 0.5680 | 2.3925 | 0.5625 | 0.5567 | 0.1670 | 0.1963 |
0.4915 | 72.0 | 504 | 1.9753 | 0.5625 | 0.5707 | 2.4507 | 0.5625 | 0.5566 | 0.1780 | 0.1986 |
0.4915 | 73.0 | 511 | 1.9792 | 0.56 | 0.5700 | 2.3604 | 0.56 | 0.5545 | 0.1580 | 0.1990 |
0.4915 | 74.0 | 518 | 1.9679 | 0.55 | 0.5700 | 2.3519 | 0.55 | 0.5412 | 0.1781 | 0.1981 |
0.4915 | 75.0 | 525 | 1.9711 | 0.57 | 0.5685 | 2.4204 | 0.57 | 0.5676 | 0.1891 | 0.1962 |
0.4915 | 76.0 | 532 | 1.9705 | 0.565 | 0.5686 | 2.3512 | 0.565 | 0.5600 | 0.1684 | 0.1967 |
0.4915 | 77.0 | 539 | 1.9673 | 0.5625 | 0.5685 | 2.3481 | 0.5625 | 0.5577 | 0.1761 | 0.1968 |
0.4915 | 78.0 | 546 | 1.9641 | 0.5625 | 0.5668 | 2.3519 | 0.5625 | 0.5577 | 0.1747 | 0.1964 |
0.4915 | 79.0 | 553 | 1.9685 | 0.5675 | 0.5677 | 2.3833 | 0.5675 | 0.5630 | 0.1490 | 0.1967 |
0.4915 | 80.0 | 560 | 1.9722 | 0.5625 | 0.5691 | 2.3864 | 0.5625 | 0.5565 | 0.1583 | 0.1979 |
0.4915 | 81.0 | 567 | 1.9641 | 0.5625 | 0.5675 | 2.3528 | 0.5625 | 0.5566 | 0.1820 | 0.1965 |
0.4915 | 82.0 | 574 | 1.9666 | 0.5625 | 0.5678 | 2.3516 | 0.5625 | 0.5564 | 0.1622 | 0.1968 |
0.4915 | 83.0 | 581 | 1.9703 | 0.56 | 0.5679 | 2.3564 | 0.56 | 0.5550 | 0.1577 | 0.1967 |
0.4915 | 84.0 | 588 | 1.9666 | 0.56 | 0.5673 | 2.3532 | 0.56 | 0.5546 | 0.1633 | 0.1967 |
0.4915 | 85.0 | 595 | 1.9683 | 0.56 | 0.5682 | 2.3506 | 0.56 | 0.5539 | 0.1526 | 0.1971 |
0.4915 | 86.0 | 602 | 1.9673 | 0.5575 | 0.5674 | 2.3523 | 0.5575 | 0.5516 | 0.1642 | 0.1972 |
0.4915 | 87.0 | 609 | 1.9678 | 0.56 | 0.5678 | 2.3514 | 0.56 | 0.5539 | 0.1543 | 0.1970 |
0.4915 | 88.0 | 616 | 1.9680 | 0.56 | 0.5678 | 2.3531 | 0.56 | 0.5539 | 0.1581 | 0.1970 |
0.4915 | 89.0 | 623 | 1.9683 | 0.5575 | 0.5681 | 2.3532 | 0.5575 | 0.5516 | 0.1732 | 0.1976 |
0.4915 | 90.0 | 630 | 1.9677 | 0.56 | 0.5677 | 2.3536 | 0.56 | 0.5539 | 0.1702 | 0.1970 |
0.4915 | 91.0 | 637 | 1.9680 | 0.5575 | 0.5679 | 2.3528 | 0.5575 | 0.5515 | 0.1760 | 0.1974 |
0.4915 | 92.0 | 644 | 1.9681 | 0.5575 | 0.5679 | 2.3522 | 0.5575 | 0.5515 | 0.1584 | 0.1975 |
0.4915 | 93.0 | 651 | 1.9680 | 0.5575 | 0.5678 | 2.3524 | 0.5575 | 0.5515 | 0.1610 | 0.1973 |
0.4915 | 94.0 | 658 | 1.9680 | 0.5575 | 0.5680 | 2.3521 | 0.5575 | 0.5515 | 0.1653 | 0.1973 |
0.4915 | 95.0 | 665 | 1.9681 | 0.5575 | 0.5679 | 2.3524 | 0.5575 | 0.5515 | 0.1663 | 0.1973 |
0.4915 | 96.0 | 672 | 1.9682 | 0.5575 | 0.5680 | 2.3528 | 0.5575 | 0.5516 | 0.1663 | 0.1973 |
0.4915 | 97.0 | 679 | 1.9682 | 0.5575 | 0.5680 | 2.3526 | 0.5575 | 0.5515 | 0.1625 | 0.1973 |
0.4915 | 98.0 | 686 | 1.9681 | 0.5575 | 0.5679 | 2.3526 | 0.5575 | 0.5516 | 0.1676 | 0.1973 |
0.4915 | 99.0 | 693 | 1.9682 | 0.5575 | 0.5679 | 2.3526 | 0.5575 | 0.5516 | 0.1635 | 0.1973 |
0.4915 | 100.0 | 700 | 1.9682 | 0.5575 | 0.5680 | 2.3526 | 0.5575 | 0.5516 | 0.1676 | 0.1973 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2