<!-- 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_t1.5_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: 1.6786
- Accuracy: 0.5475
- Brier Loss: 0.5719
- Nll: 2.4381
- F1 Micro: 0.5475
- F1 Macro: 0.5432
- Ece: 0.1563
- Aurc: 0.2080
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 | 4.6969 | 0.0425 | 1.0710 | 7.3408 | 0.0425 | 0.0336 | 0.2818 | 0.9543 |
No log | 2.0 | 14 | 3.6834 | 0.1 | 0.9492 | 5.8001 | 0.1000 | 0.0878 | 0.1662 | 0.8811 |
No log | 3.0 | 21 | 3.2796 | 0.2325 | 0.8815 | 5.3289 | 0.2325 | 0.1817 | 0.1601 | 0.6217 |
No log | 4.0 | 28 | 2.7674 | 0.2975 | 0.7907 | 3.6563 | 0.2975 | 0.2855 | 0.1624 | 0.4730 |
No log | 5.0 | 35 | 2.4072 | 0.3925 | 0.7175 | 3.0843 | 0.3925 | 0.3726 | 0.1811 | 0.3501 |
No log | 6.0 | 42 | 2.2499 | 0.44 | 0.6923 | 3.0183 | 0.44 | 0.4252 | 0.2052 | 0.3017 |
No log | 7.0 | 49 | 2.2199 | 0.455 | 0.6973 | 3.0254 | 0.455 | 0.4233 | 0.2117 | 0.2958 |
No log | 8.0 | 56 | 2.1564 | 0.4725 | 0.6958 | 2.8105 | 0.4725 | 0.4637 | 0.2159 | 0.2954 |
No log | 9.0 | 63 | 2.2963 | 0.4575 | 0.7105 | 2.7884 | 0.4575 | 0.4386 | 0.2617 | 0.2866 |
No log | 10.0 | 70 | 2.3511 | 0.4475 | 0.7553 | 2.8174 | 0.4475 | 0.4278 | 0.2698 | 0.3171 |
No log | 11.0 | 77 | 2.4111 | 0.4425 | 0.7483 | 2.9313 | 0.4425 | 0.4290 | 0.2569 | 0.3056 |
No log | 12.0 | 84 | 2.2927 | 0.4825 | 0.7158 | 2.7197 | 0.4825 | 0.4799 | 0.2453 | 0.2920 |
No log | 13.0 | 91 | 2.2809 | 0.4825 | 0.7098 | 2.6808 | 0.4825 | 0.4758 | 0.2454 | 0.2811 |
No log | 14.0 | 98 | 2.1787 | 0.4825 | 0.7036 | 2.5537 | 0.4825 | 0.4801 | 0.2417 | 0.2877 |
No log | 15.0 | 105 | 2.1643 | 0.495 | 0.6934 | 2.7769 | 0.495 | 0.4892 | 0.2375 | 0.2707 |
No log | 16.0 | 112 | 2.1309 | 0.48 | 0.6866 | 2.5936 | 0.48 | 0.4647 | 0.2130 | 0.2663 |
No log | 17.0 | 119 | 2.1389 | 0.5125 | 0.6580 | 2.6313 | 0.5125 | 0.4962 | 0.1976 | 0.2424 |
No log | 18.0 | 126 | 2.0813 | 0.5025 | 0.6649 | 2.6517 | 0.5025 | 0.5019 | 0.2126 | 0.2602 |
No log | 19.0 | 133 | 2.0742 | 0.4975 | 0.6763 | 2.6690 | 0.4975 | 0.4739 | 0.2473 | 0.2575 |
No log | 20.0 | 140 | 2.0550 | 0.48 | 0.6702 | 2.5468 | 0.48 | 0.4838 | 0.2206 | 0.2745 |
No log | 21.0 | 147 | 1.9216 | 0.52 | 0.6332 | 2.6838 | 0.52 | 0.5237 | 0.1895 | 0.2410 |
No log | 22.0 | 154 | 1.9572 | 0.5025 | 0.6500 | 2.4719 | 0.5025 | 0.4877 | 0.2223 | 0.2470 |
No log | 23.0 | 161 | 1.8821 | 0.5175 | 0.6294 | 2.5337 | 0.5175 | 0.5067 | 0.2081 | 0.2332 |
No log | 24.0 | 168 | 1.9198 | 0.5125 | 0.6332 | 2.5310 | 0.5125 | 0.5028 | 0.2039 | 0.2434 |
No log | 25.0 | 175 | 1.9419 | 0.515 | 0.6358 | 2.6324 | 0.515 | 0.5005 | 0.2203 | 0.2383 |
No log | 26.0 | 182 | 1.8714 | 0.525 | 0.6145 | 2.5820 | 0.525 | 0.5005 | 0.1843 | 0.2232 |
No log | 27.0 | 189 | 1.9460 | 0.4975 | 0.6365 | 2.6119 | 0.4975 | 0.4907 | 0.2169 | 0.2395 |
No log | 28.0 | 196 | 1.9364 | 0.525 | 0.6333 | 2.5789 | 0.525 | 0.5013 | 0.1808 | 0.2437 |
No log | 29.0 | 203 | 1.9143 | 0.5425 | 0.6144 | 2.7809 | 0.5425 | 0.5181 | 0.1700 | 0.2224 |
No log | 30.0 | 210 | 1.8565 | 0.5275 | 0.6171 | 2.5610 | 0.5275 | 0.5142 | 0.1823 | 0.2304 |
No log | 31.0 | 217 | 1.8281 | 0.5325 | 0.6113 | 2.6216 | 0.5325 | 0.5191 | 0.1802 | 0.2167 |
No log | 32.0 | 224 | 1.8213 | 0.525 | 0.6192 | 2.5521 | 0.525 | 0.5278 | 0.1844 | 0.2368 |
No log | 33.0 | 231 | 1.7724 | 0.545 | 0.6030 | 2.4900 | 0.545 | 0.5460 | 0.1810 | 0.2211 |
No log | 34.0 | 238 | 1.7617 | 0.5475 | 0.5985 | 2.5994 | 0.5475 | 0.5345 | 0.1907 | 0.2161 |
No log | 35.0 | 245 | 1.8276 | 0.54 | 0.6090 | 2.5254 | 0.54 | 0.5172 | 0.1839 | 0.2249 |
No log | 36.0 | 252 | 1.7646 | 0.5475 | 0.5924 | 2.5729 | 0.5475 | 0.5458 | 0.1930 | 0.2190 |
No log | 37.0 | 259 | 1.7851 | 0.5675 | 0.5872 | 2.5036 | 0.5675 | 0.5600 | 0.1670 | 0.2056 |
No log | 38.0 | 266 | 1.7743 | 0.5225 | 0.6136 | 2.4765 | 0.5225 | 0.5163 | 0.1831 | 0.2276 |
No log | 39.0 | 273 | 1.7760 | 0.5325 | 0.5912 | 2.5692 | 0.5325 | 0.5228 | 0.1589 | 0.2108 |
No log | 40.0 | 280 | 1.7664 | 0.53 | 0.5917 | 2.5363 | 0.53 | 0.5089 | 0.1800 | 0.2095 |
No log | 41.0 | 287 | 1.8177 | 0.535 | 0.6070 | 2.5195 | 0.535 | 0.5272 | 0.1883 | 0.2268 |
No log | 42.0 | 294 | 1.7575 | 0.56 | 0.5868 | 2.5956 | 0.56 | 0.5504 | 0.1740 | 0.2091 |
No log | 43.0 | 301 | 1.7616 | 0.54 | 0.6001 | 2.3469 | 0.54 | 0.5401 | 0.1861 | 0.2272 |
No log | 44.0 | 308 | 1.7105 | 0.5425 | 0.5831 | 2.4709 | 0.5425 | 0.5380 | 0.1884 | 0.2086 |
No log | 45.0 | 315 | 1.7502 | 0.565 | 0.5880 | 2.4600 | 0.565 | 0.5546 | 0.1527 | 0.2084 |
No log | 46.0 | 322 | 1.7135 | 0.565 | 0.5834 | 2.4571 | 0.565 | 0.5631 | 0.1703 | 0.2103 |
No log | 47.0 | 329 | 1.7327 | 0.5525 | 0.5892 | 2.4573 | 0.5525 | 0.5491 | 0.1820 | 0.2178 |
No log | 48.0 | 336 | 1.7405 | 0.535 | 0.5947 | 2.5414 | 0.535 | 0.5338 | 0.1840 | 0.2180 |
No log | 49.0 | 343 | 1.7265 | 0.555 | 0.5758 | 2.4824 | 0.555 | 0.5423 | 0.1603 | 0.2025 |
No log | 50.0 | 350 | 1.7065 | 0.5525 | 0.5904 | 2.3807 | 0.5525 | 0.5576 | 0.1769 | 0.2212 |
No log | 51.0 | 357 | 1.7265 | 0.545 | 0.5833 | 2.4535 | 0.545 | 0.5413 | 0.1704 | 0.2108 |
No log | 52.0 | 364 | 1.7197 | 0.55 | 0.5740 | 2.5386 | 0.55 | 0.5347 | 0.1467 | 0.2032 |
No log | 53.0 | 371 | 1.7015 | 0.5575 | 0.5826 | 2.4159 | 0.5575 | 0.5578 | 0.1751 | 0.2138 |
No log | 54.0 | 378 | 1.7263 | 0.55 | 0.5873 | 2.4471 | 0.55 | 0.5456 | 0.1629 | 0.2144 |
No log | 55.0 | 385 | 1.6786 | 0.555 | 0.5780 | 2.3908 | 0.555 | 0.5490 | 0.1627 | 0.2106 |
No log | 56.0 | 392 | 1.7147 | 0.55 | 0.5811 | 2.3876 | 0.55 | 0.5476 | 0.1724 | 0.2122 |
No log | 57.0 | 399 | 1.6983 | 0.5525 | 0.5769 | 2.5028 | 0.5525 | 0.5415 | 0.1716 | 0.2054 |
No log | 58.0 | 406 | 1.7350 | 0.5425 | 0.5984 | 2.3835 | 0.5425 | 0.5406 | 0.1806 | 0.2282 |
No log | 59.0 | 413 | 1.7015 | 0.54 | 0.5779 | 2.4865 | 0.54 | 0.5288 | 0.1848 | 0.2079 |
No log | 60.0 | 420 | 1.6783 | 0.5525 | 0.5712 | 2.4231 | 0.5525 | 0.5497 | 0.1777 | 0.2046 |
No log | 61.0 | 427 | 1.7236 | 0.545 | 0.5856 | 2.5104 | 0.545 | 0.5379 | 0.1850 | 0.2142 |
No log | 62.0 | 434 | 1.6904 | 0.5425 | 0.5752 | 2.4626 | 0.5425 | 0.5400 | 0.1787 | 0.2084 |
No log | 63.0 | 441 | 1.7136 | 0.5375 | 0.5837 | 2.4091 | 0.5375 | 0.5304 | 0.1681 | 0.2150 |
No log | 64.0 | 448 | 1.6783 | 0.56 | 0.5762 | 2.4017 | 0.56 | 0.5594 | 0.1607 | 0.2093 |
No log | 65.0 | 455 | 1.7098 | 0.5425 | 0.5797 | 2.5062 | 0.5425 | 0.5406 | 0.1707 | 0.2120 |
No log | 66.0 | 462 | 1.6752 | 0.5525 | 0.5712 | 2.4684 | 0.5525 | 0.5478 | 0.1711 | 0.2063 |
No log | 67.0 | 469 | 1.7068 | 0.545 | 0.5761 | 2.5107 | 0.545 | 0.5415 | 0.1750 | 0.2090 |
No log | 68.0 | 476 | 1.6706 | 0.5525 | 0.5700 | 2.4322 | 0.5525 | 0.5466 | 0.1798 | 0.2068 |
No log | 69.0 | 483 | 1.6941 | 0.5525 | 0.5749 | 2.4689 | 0.5525 | 0.5501 | 0.1654 | 0.2072 |
No log | 70.0 | 490 | 1.6726 | 0.545 | 0.5711 | 2.4348 | 0.545 | 0.5416 | 0.1811 | 0.2083 |
No log | 71.0 | 497 | 1.6831 | 0.5525 | 0.5740 | 2.4659 | 0.5525 | 0.5498 | 0.1687 | 0.2061 |
0.4086 | 72.0 | 504 | 1.6749 | 0.5575 | 0.5699 | 2.4122 | 0.5575 | 0.5535 | 0.1706 | 0.2051 |
0.4086 | 73.0 | 511 | 1.6867 | 0.555 | 0.5740 | 2.4402 | 0.555 | 0.5495 | 0.1725 | 0.2069 |
0.4086 | 74.0 | 518 | 1.6768 | 0.5475 | 0.5716 | 2.4369 | 0.5475 | 0.5442 | 0.1621 | 0.2080 |
0.4086 | 75.0 | 525 | 1.6835 | 0.55 | 0.5736 | 2.4404 | 0.55 | 0.5448 | 0.1824 | 0.2075 |
0.4086 | 76.0 | 532 | 1.6737 | 0.5525 | 0.5701 | 2.4361 | 0.5525 | 0.5497 | 0.1731 | 0.2063 |
0.4086 | 77.0 | 539 | 1.6796 | 0.55 | 0.5721 | 2.4399 | 0.55 | 0.5450 | 0.1723 | 0.2066 |
0.4086 | 78.0 | 546 | 1.6774 | 0.55 | 0.5718 | 2.4362 | 0.55 | 0.5472 | 0.1732 | 0.2071 |
0.4086 | 79.0 | 553 | 1.6781 | 0.555 | 0.5711 | 2.4390 | 0.555 | 0.5506 | 0.1588 | 0.2059 |
0.4086 | 80.0 | 560 | 1.6787 | 0.555 | 0.5714 | 2.4380 | 0.555 | 0.5508 | 0.1832 | 0.2070 |
0.4086 | 81.0 | 567 | 1.6778 | 0.555 | 0.5717 | 2.4366 | 0.555 | 0.5516 | 0.1615 | 0.2060 |
0.4086 | 82.0 | 574 | 1.6788 | 0.55 | 0.5717 | 2.4380 | 0.55 | 0.5462 | 0.1680 | 0.2070 |
0.4086 | 83.0 | 581 | 1.6761 | 0.5525 | 0.5712 | 2.4366 | 0.5525 | 0.5505 | 0.1809 | 0.2064 |
0.4086 | 84.0 | 588 | 1.6778 | 0.55 | 0.5714 | 2.4388 | 0.55 | 0.5443 | 0.1571 | 0.2073 |
0.4086 | 85.0 | 595 | 1.6772 | 0.5525 | 0.5716 | 2.4367 | 0.5525 | 0.5479 | 0.1768 | 0.2073 |
0.4086 | 86.0 | 602 | 1.6789 | 0.5525 | 0.5722 | 2.4376 | 0.5525 | 0.5470 | 0.1646 | 0.2066 |
0.4086 | 87.0 | 609 | 1.6784 | 0.5525 | 0.5717 | 2.4384 | 0.5525 | 0.5486 | 0.1743 | 0.2073 |
0.4086 | 88.0 | 616 | 1.6786 | 0.55 | 0.5720 | 2.4382 | 0.55 | 0.5443 | 0.1559 | 0.2077 |
0.4086 | 89.0 | 623 | 1.6784 | 0.5525 | 0.5718 | 2.4379 | 0.5525 | 0.5479 | 0.1561 | 0.2073 |
0.4086 | 90.0 | 630 | 1.6782 | 0.5525 | 0.5718 | 2.4381 | 0.5525 | 0.5482 | 0.1561 | 0.2066 |
0.4086 | 91.0 | 637 | 1.6784 | 0.5525 | 0.5719 | 2.4375 | 0.5525 | 0.5482 | 0.1612 | 0.2065 |
0.4086 | 92.0 | 644 | 1.6783 | 0.55 | 0.5718 | 2.4377 | 0.55 | 0.5456 | 0.1702 | 0.2073 |
0.4086 | 93.0 | 651 | 1.6781 | 0.55 | 0.5718 | 2.4378 | 0.55 | 0.5456 | 0.1581 | 0.2075 |
0.4086 | 94.0 | 658 | 1.6785 | 0.55 | 0.5719 | 2.4379 | 0.55 | 0.5455 | 0.1602 | 0.2074 |
0.4086 | 95.0 | 665 | 1.6785 | 0.55 | 0.5719 | 2.4379 | 0.55 | 0.5456 | 0.1635 | 0.2075 |
0.4086 | 96.0 | 672 | 1.6785 | 0.5475 | 0.5719 | 2.4381 | 0.5475 | 0.5432 | 0.1659 | 0.2080 |
0.4086 | 97.0 | 679 | 1.6786 | 0.55 | 0.5719 | 2.4381 | 0.55 | 0.5458 | 0.1545 | 0.2071 |
0.4086 | 98.0 | 686 | 1.6786 | 0.5475 | 0.5719 | 2.4381 | 0.5475 | 0.5431 | 0.1613 | 0.2081 |
0.4086 | 99.0 | 693 | 1.6786 | 0.5475 | 0.5719 | 2.4381 | 0.5475 | 0.5431 | 0.1613 | 0.2081 |
0.4086 | 100.0 | 700 | 1.6786 | 0.5475 | 0.5719 | 2.4381 | 0.5475 | 0.5432 | 0.1563 | 0.2080 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2