<!-- 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_a0.9
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: 1.5750
- Accuracy: 0.5325
- Brier Loss: 0.5990
- Nll: 2.5263
- F1 Micro: 0.5325
- F1 Macro: 0.5240
- Ece: 0.1659
- Aurc: 0.2152
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 | 3.9285 | 0.04 | 1.0722 | 7.3572 | 0.04 | 0.0319 | 0.2792 | 0.9556 |
No log | 2.0 | 14 | 3.0027 | 0.095 | 0.9510 | 5.8779 | 0.095 | 0.0766 | 0.1668 | 0.8900 |
No log | 3.0 | 21 | 2.6988 | 0.225 | 0.8896 | 5.3750 | 0.225 | 0.1801 | 0.1669 | 0.6473 |
No log | 4.0 | 28 | 2.3179 | 0.285 | 0.8016 | 3.5658 | 0.285 | 0.2657 | 0.1679 | 0.4916 |
No log | 5.0 | 35 | 2.0566 | 0.37 | 0.7203 | 2.9834 | 0.37 | 0.3493 | 0.1684 | 0.3612 |
No log | 6.0 | 42 | 1.9505 | 0.4325 | 0.6892 | 3.0719 | 0.4325 | 0.4127 | 0.1775 | 0.3084 |
No log | 7.0 | 49 | 1.9995 | 0.4375 | 0.7008 | 3.1569 | 0.4375 | 0.4084 | 0.2032 | 0.3103 |
No log | 8.0 | 56 | 1.9133 | 0.445 | 0.6906 | 2.8574 | 0.445 | 0.4464 | 0.2016 | 0.3062 |
No log | 9.0 | 63 | 1.9876 | 0.4625 | 0.6918 | 2.9267 | 0.4625 | 0.4538 | 0.2228 | 0.2868 |
No log | 10.0 | 70 | 2.0051 | 0.4725 | 0.6971 | 2.9249 | 0.4725 | 0.4553 | 0.2234 | 0.2814 |
No log | 11.0 | 77 | 2.1834 | 0.465 | 0.7319 | 2.9998 | 0.465 | 0.4426 | 0.2444 | 0.3006 |
No log | 12.0 | 84 | 1.9953 | 0.4825 | 0.7087 | 2.7128 | 0.4825 | 0.4731 | 0.2386 | 0.2980 |
No log | 13.0 | 91 | 1.8834 | 0.4975 | 0.6771 | 2.6879 | 0.4975 | 0.4954 | 0.2240 | 0.2748 |
No log | 14.0 | 98 | 1.9647 | 0.4675 | 0.6987 | 2.8305 | 0.4675 | 0.4429 | 0.2409 | 0.2902 |
No log | 15.0 | 105 | 1.8810 | 0.5 | 0.6785 | 2.6402 | 0.5 | 0.4847 | 0.2171 | 0.2725 |
No log | 16.0 | 112 | 1.8777 | 0.4875 | 0.6877 | 2.6940 | 0.4875 | 0.4871 | 0.2210 | 0.2846 |
No log | 17.0 | 119 | 1.9260 | 0.4925 | 0.6796 | 2.7055 | 0.4925 | 0.4834 | 0.2012 | 0.2744 |
No log | 18.0 | 126 | 1.7864 | 0.505 | 0.6547 | 2.6724 | 0.505 | 0.4912 | 0.2081 | 0.2434 |
No log | 19.0 | 133 | 1.7618 | 0.4975 | 0.6430 | 2.5951 | 0.4975 | 0.4915 | 0.2172 | 0.2490 |
No log | 20.0 | 140 | 1.7496 | 0.515 | 0.6513 | 2.5263 | 0.515 | 0.5025 | 0.1975 | 0.2502 |
No log | 21.0 | 147 | 1.7082 | 0.5275 | 0.6438 | 2.4039 | 0.5275 | 0.5224 | 0.2017 | 0.2450 |
No log | 22.0 | 154 | 1.7482 | 0.4975 | 0.6682 | 2.5194 | 0.4975 | 0.4911 | 0.2247 | 0.2571 |
No log | 23.0 | 161 | 1.7377 | 0.5075 | 0.6482 | 2.4136 | 0.5075 | 0.4900 | 0.2221 | 0.2396 |
No log | 24.0 | 168 | 1.7094 | 0.515 | 0.6372 | 2.5605 | 0.515 | 0.5083 | 0.2137 | 0.2474 |
No log | 25.0 | 175 | 1.6884 | 0.5175 | 0.6422 | 2.5270 | 0.5175 | 0.5104 | 0.2111 | 0.2444 |
No log | 26.0 | 182 | 1.6489 | 0.5275 | 0.6246 | 2.5344 | 0.5275 | 0.5211 | 0.2066 | 0.2333 |
No log | 27.0 | 189 | 1.6165 | 0.53 | 0.6191 | 2.5418 | 0.53 | 0.5256 | 0.2021 | 0.2305 |
No log | 28.0 | 196 | 1.6316 | 0.5275 | 0.6181 | 2.6568 | 0.5275 | 0.5212 | 0.2004 | 0.2300 |
No log | 29.0 | 203 | 1.6595 | 0.5175 | 0.6306 | 2.4298 | 0.5175 | 0.5096 | 0.2020 | 0.2427 |
No log | 30.0 | 210 | 1.6193 | 0.5325 | 0.6157 | 2.5455 | 0.5325 | 0.5272 | 0.1779 | 0.2278 |
No log | 31.0 | 217 | 1.6517 | 0.5325 | 0.6274 | 2.4579 | 0.5325 | 0.5259 | 0.2006 | 0.2362 |
No log | 32.0 | 224 | 1.6434 | 0.5325 | 0.6167 | 2.5805 | 0.5325 | 0.5229 | 0.1995 | 0.2273 |
No log | 33.0 | 231 | 1.6660 | 0.5225 | 0.6269 | 2.6794 | 0.5225 | 0.5132 | 0.2244 | 0.2283 |
No log | 34.0 | 238 | 1.6353 | 0.515 | 0.6194 | 2.6085 | 0.515 | 0.5069 | 0.1839 | 0.2303 |
No log | 35.0 | 245 | 1.5920 | 0.5325 | 0.6051 | 2.5645 | 0.5325 | 0.5248 | 0.1868 | 0.2208 |
No log | 36.0 | 252 | 1.5909 | 0.54 | 0.6028 | 2.4786 | 0.54 | 0.5323 | 0.1902 | 0.2194 |
No log | 37.0 | 259 | 1.5730 | 0.5425 | 0.5983 | 2.4877 | 0.5425 | 0.5368 | 0.1799 | 0.2177 |
No log | 38.0 | 266 | 1.5800 | 0.535 | 0.6029 | 2.4736 | 0.535 | 0.5282 | 0.1761 | 0.2196 |
No log | 39.0 | 273 | 1.5594 | 0.54 | 0.5955 | 2.5093 | 0.54 | 0.5327 | 0.1900 | 0.2126 |
No log | 40.0 | 280 | 1.5685 | 0.53 | 0.5979 | 2.6068 | 0.53 | 0.5208 | 0.1893 | 0.2173 |
No log | 41.0 | 287 | 1.5757 | 0.53 | 0.5995 | 2.5655 | 0.53 | 0.5218 | 0.1862 | 0.2164 |
No log | 42.0 | 294 | 1.5797 | 0.535 | 0.6039 | 2.5445 | 0.535 | 0.5273 | 0.1834 | 0.2182 |
No log | 43.0 | 301 | 1.5900 | 0.53 | 0.6074 | 2.5201 | 0.53 | 0.5189 | 0.1747 | 0.2206 |
No log | 44.0 | 308 | 1.5760 | 0.5325 | 0.5986 | 2.4974 | 0.5325 | 0.5225 | 0.1870 | 0.2148 |
No log | 45.0 | 315 | 1.5768 | 0.53 | 0.6013 | 2.5174 | 0.53 | 0.5204 | 0.1979 | 0.2158 |
No log | 46.0 | 322 | 1.5774 | 0.53 | 0.6011 | 2.5199 | 0.53 | 0.5206 | 0.1882 | 0.2165 |
No log | 47.0 | 329 | 1.5714 | 0.54 | 0.5983 | 2.5329 | 0.54 | 0.5303 | 0.1884 | 0.2135 |
No log | 48.0 | 336 | 1.5834 | 0.5325 | 0.6026 | 2.5253 | 0.5325 | 0.5238 | 0.1658 | 0.2190 |
No log | 49.0 | 343 | 1.5724 | 0.5375 | 0.5979 | 2.5569 | 0.5375 | 0.5299 | 0.1617 | 0.2151 |
No log | 50.0 | 350 | 1.5685 | 0.5375 | 0.5985 | 2.5189 | 0.5375 | 0.5285 | 0.1919 | 0.2151 |
No log | 51.0 | 357 | 1.5708 | 0.54 | 0.5986 | 2.5002 | 0.54 | 0.5305 | 0.1755 | 0.2149 |
No log | 52.0 | 364 | 1.5665 | 0.535 | 0.5977 | 2.5224 | 0.535 | 0.5267 | 0.1842 | 0.2160 |
No log | 53.0 | 371 | 1.5713 | 0.5325 | 0.5993 | 2.5515 | 0.5325 | 0.5250 | 0.1753 | 0.2160 |
No log | 54.0 | 378 | 1.5693 | 0.535 | 0.5986 | 2.5516 | 0.535 | 0.5276 | 0.1841 | 0.2158 |
No log | 55.0 | 385 | 1.5693 | 0.5375 | 0.5984 | 2.5190 | 0.5375 | 0.5285 | 0.1842 | 0.2144 |
No log | 56.0 | 392 | 1.5725 | 0.535 | 0.5992 | 2.5527 | 0.535 | 0.5262 | 0.1776 | 0.2150 |
No log | 57.0 | 399 | 1.5674 | 0.5425 | 0.5976 | 2.5502 | 0.5425 | 0.5326 | 0.1902 | 0.2137 |
No log | 58.0 | 406 | 1.5675 | 0.5375 | 0.5974 | 2.5517 | 0.5375 | 0.5288 | 0.1794 | 0.2139 |
No log | 59.0 | 413 | 1.5713 | 0.535 | 0.5988 | 2.5515 | 0.535 | 0.5257 | 0.1791 | 0.2147 |
No log | 60.0 | 420 | 1.5729 | 0.535 | 0.5988 | 2.5512 | 0.535 | 0.5262 | 0.1796 | 0.2148 |
No log | 61.0 | 427 | 1.5702 | 0.5375 | 0.5976 | 2.5521 | 0.5375 | 0.5281 | 0.1817 | 0.2139 |
No log | 62.0 | 434 | 1.5728 | 0.535 | 0.5988 | 2.5514 | 0.535 | 0.5266 | 0.1722 | 0.2149 |
No log | 63.0 | 441 | 1.5720 | 0.5325 | 0.5985 | 2.5206 | 0.5325 | 0.5231 | 0.1790 | 0.2149 |
No log | 64.0 | 448 | 1.5704 | 0.5325 | 0.5975 | 2.5510 | 0.5325 | 0.5236 | 0.1706 | 0.2139 |
No log | 65.0 | 455 | 1.5724 | 0.5325 | 0.5986 | 2.5225 | 0.5325 | 0.5236 | 0.1557 | 0.2148 |
No log | 66.0 | 462 | 1.5718 | 0.5325 | 0.5985 | 2.5246 | 0.5325 | 0.5241 | 0.1772 | 0.2148 |
No log | 67.0 | 469 | 1.5710 | 0.5325 | 0.5981 | 2.5511 | 0.5325 | 0.5237 | 0.1625 | 0.2146 |
No log | 68.0 | 476 | 1.5716 | 0.54 | 0.5981 | 2.5001 | 0.54 | 0.5304 | 0.1622 | 0.2141 |
No log | 69.0 | 483 | 1.5732 | 0.5325 | 0.5988 | 2.5517 | 0.5325 | 0.5232 | 0.1641 | 0.2150 |
No log | 70.0 | 490 | 1.5733 | 0.5325 | 0.5987 | 2.5522 | 0.5325 | 0.5237 | 0.1715 | 0.2149 |
No log | 71.0 | 497 | 1.5729 | 0.5325 | 0.5985 | 2.5523 | 0.5325 | 0.5241 | 0.1670 | 0.2147 |
0.3153 | 72.0 | 504 | 1.5730 | 0.5325 | 0.5987 | 2.5236 | 0.5325 | 0.5237 | 0.1656 | 0.2149 |
0.3153 | 73.0 | 511 | 1.5723 | 0.5325 | 0.5985 | 2.5212 | 0.5325 | 0.5238 | 0.1893 | 0.2145 |
0.3153 | 74.0 | 518 | 1.5738 | 0.5325 | 0.5989 | 2.5515 | 0.5325 | 0.5238 | 0.1744 | 0.2147 |
0.3153 | 75.0 | 525 | 1.5740 | 0.5325 | 0.5988 | 2.5318 | 0.5325 | 0.5237 | 0.1683 | 0.2150 |
0.3153 | 76.0 | 532 | 1.5734 | 0.535 | 0.5985 | 2.5525 | 0.535 | 0.5261 | 0.1763 | 0.2145 |
0.3153 | 77.0 | 539 | 1.5740 | 0.5325 | 0.5989 | 2.5516 | 0.5325 | 0.5243 | 0.1726 | 0.2149 |
0.3153 | 78.0 | 546 | 1.5738 | 0.5325 | 0.5987 | 2.5289 | 0.5325 | 0.5241 | 0.1692 | 0.2148 |
0.3153 | 79.0 | 553 | 1.5736 | 0.5325 | 0.5987 | 2.5255 | 0.5325 | 0.5242 | 0.1807 | 0.2147 |
0.3153 | 80.0 | 560 | 1.5739 | 0.5325 | 0.5988 | 2.5522 | 0.5325 | 0.5237 | 0.1769 | 0.2150 |
0.3153 | 81.0 | 567 | 1.5743 | 0.5325 | 0.5989 | 2.5519 | 0.5325 | 0.5238 | 0.1837 | 0.2151 |
0.3153 | 82.0 | 574 | 1.5742 | 0.5325 | 0.5989 | 2.5232 | 0.5325 | 0.5240 | 0.1712 | 0.2149 |
0.3153 | 83.0 | 581 | 1.5744 | 0.5325 | 0.5989 | 2.5256 | 0.5325 | 0.5239 | 0.1803 | 0.2151 |
0.3153 | 84.0 | 588 | 1.5741 | 0.5325 | 0.5988 | 2.5233 | 0.5325 | 0.5233 | 0.1655 | 0.2147 |
0.3153 | 85.0 | 595 | 1.5747 | 0.5325 | 0.5990 | 2.5274 | 0.5325 | 0.5237 | 0.1696 | 0.2152 |
0.3153 | 86.0 | 602 | 1.5747 | 0.5325 | 0.5989 | 2.5263 | 0.5325 | 0.5238 | 0.1689 | 0.2150 |
0.3153 | 87.0 | 609 | 1.5745 | 0.5325 | 0.5989 | 2.5251 | 0.5325 | 0.5237 | 0.1654 | 0.2149 |
0.3153 | 88.0 | 616 | 1.5747 | 0.5325 | 0.5989 | 2.5283 | 0.5325 | 0.5241 | 0.1693 | 0.2151 |
0.3153 | 89.0 | 623 | 1.5748 | 0.5325 | 0.5990 | 2.5275 | 0.5325 | 0.5239 | 0.1596 | 0.2152 |
0.3153 | 90.0 | 630 | 1.5749 | 0.5325 | 0.5990 | 2.5278 | 0.5325 | 0.5240 | 0.1602 | 0.2151 |
0.3153 | 91.0 | 637 | 1.5750 | 0.5325 | 0.5990 | 2.5337 | 0.5325 | 0.5239 | 0.1623 | 0.2152 |
0.3153 | 92.0 | 644 | 1.5749 | 0.5325 | 0.5990 | 2.5272 | 0.5325 | 0.5238 | 0.1653 | 0.2151 |
0.3153 | 93.0 | 651 | 1.5751 | 0.5325 | 0.5990 | 2.5281 | 0.5325 | 0.5240 | 0.1663 | 0.2149 |
0.3153 | 94.0 | 658 | 1.5750 | 0.5325 | 0.5990 | 2.5249 | 0.5325 | 0.5239 | 0.1715 | 0.2152 |
0.3153 | 95.0 | 665 | 1.5749 | 0.535 | 0.5990 | 2.5257 | 0.535 | 0.5263 | 0.1625 | 0.2149 |
0.3153 | 96.0 | 672 | 1.5750 | 0.5325 | 0.5990 | 2.5266 | 0.5325 | 0.5239 | 0.1655 | 0.2151 |
0.3153 | 97.0 | 679 | 1.5750 | 0.5325 | 0.5990 | 2.5268 | 0.5325 | 0.5239 | 0.1686 | 0.2152 |
0.3153 | 98.0 | 686 | 1.5750 | 0.5325 | 0.5990 | 2.5275 | 0.5325 | 0.5240 | 0.1664 | 0.2152 |
0.3153 | 99.0 | 693 | 1.5750 | 0.5325 | 0.5990 | 2.5269 | 0.5325 | 0.5240 | 0.1678 | 0.2152 |
0.3153 | 100.0 | 700 | 1.5750 | 0.5325 | 0.5990 | 2.5263 | 0.5325 | 0.5240 | 0.1659 | 0.2152 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2