<!-- 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_NKD_t1.0_g1.5
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: 5.4262
- Accuracy: 0.535
- Brier Loss: 0.6080
- Nll: 2.4569
- F1 Micro: 0.535
- F1 Macro: 0.5345
- Ece: 0.2120
- Aurc: 0.2105
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 | 8.7328 | 0.0375 | 1.0712 | 7.1959 | 0.0375 | 0.0284 | 0.2770 | 0.9564 |
No log | 2.0 | 14 | 6.6754 | 0.0725 | 0.9650 | 6.2033 | 0.0725 | 0.0442 | 0.1890 | 0.9239 |
No log | 3.0 | 21 | 6.2399 | 0.1525 | 0.9213 | 5.6927 | 0.1525 | 0.1339 | 0.1647 | 0.8105 |
No log | 4.0 | 28 | 5.9557 | 0.2125 | 0.8771 | 4.2773 | 0.2125 | 0.1952 | 0.1637 | 0.6347 |
No log | 5.0 | 35 | 5.5957 | 0.3275 | 0.7962 | 3.1599 | 0.3275 | 0.3087 | 0.1928 | 0.4595 |
No log | 6.0 | 42 | 5.4573 | 0.4075 | 0.7283 | 3.0860 | 0.4075 | 0.3732 | 0.1803 | 0.3610 |
No log | 7.0 | 49 | 5.3445 | 0.43 | 0.7150 | 2.9840 | 0.4300 | 0.4112 | 0.2071 | 0.3498 |
No log | 8.0 | 56 | 5.2986 | 0.445 | 0.6860 | 2.9281 | 0.445 | 0.4327 | 0.1891 | 0.3159 |
No log | 9.0 | 63 | 5.2567 | 0.465 | 0.6608 | 3.0783 | 0.465 | 0.4494 | 0.1835 | 0.2885 |
No log | 10.0 | 70 | 5.4317 | 0.4425 | 0.6752 | 2.9067 | 0.4425 | 0.4477 | 0.1836 | 0.3209 |
No log | 11.0 | 77 | 5.2073 | 0.4825 | 0.6342 | 2.7960 | 0.4825 | 0.4664 | 0.1731 | 0.2708 |
No log | 12.0 | 84 | 5.2094 | 0.51 | 0.6150 | 2.8886 | 0.51 | 0.4767 | 0.1463 | 0.2459 |
No log | 13.0 | 91 | 5.2379 | 0.4975 | 0.6135 | 2.7446 | 0.4975 | 0.4819 | 0.1702 | 0.2436 |
No log | 14.0 | 98 | 5.2716 | 0.505 | 0.6123 | 2.7888 | 0.505 | 0.4962 | 0.1595 | 0.2496 |
No log | 15.0 | 105 | 5.2410 | 0.5125 | 0.6090 | 2.6474 | 0.5125 | 0.5144 | 0.1659 | 0.2451 |
No log | 16.0 | 112 | 5.2764 | 0.54 | 0.6063 | 2.7700 | 0.54 | 0.5316 | 0.1634 | 0.2383 |
No log | 17.0 | 119 | 5.2652 | 0.5275 | 0.6027 | 2.7934 | 0.5275 | 0.5171 | 0.1800 | 0.2326 |
No log | 18.0 | 126 | 5.2145 | 0.54 | 0.5944 | 2.6456 | 0.54 | 0.5350 | 0.1548 | 0.2323 |
No log | 19.0 | 133 | 5.2611 | 0.5175 | 0.6096 | 2.5302 | 0.5175 | 0.5273 | 0.1709 | 0.2435 |
No log | 20.0 | 140 | 5.3536 | 0.52 | 0.6110 | 2.6530 | 0.52 | 0.5229 | 0.1619 | 0.2359 |
No log | 21.0 | 147 | 5.3020 | 0.5125 | 0.6060 | 2.7184 | 0.5125 | 0.5070 | 0.1724 | 0.2398 |
No log | 22.0 | 154 | 5.2107 | 0.5275 | 0.5926 | 2.4436 | 0.5275 | 0.5242 | 0.1618 | 0.2255 |
No log | 23.0 | 161 | 5.2723 | 0.53 | 0.5953 | 2.7008 | 0.53 | 0.5209 | 0.1698 | 0.2253 |
No log | 24.0 | 168 | 5.1615 | 0.5325 | 0.5875 | 2.4753 | 0.5325 | 0.5254 | 0.1699 | 0.2247 |
No log | 25.0 | 175 | 5.1795 | 0.5375 | 0.5825 | 2.6856 | 0.5375 | 0.5316 | 0.1781 | 0.2144 |
No log | 26.0 | 182 | 5.2340 | 0.54 | 0.5937 | 2.6542 | 0.54 | 0.5271 | 0.1778 | 0.2215 |
No log | 27.0 | 189 | 5.2197 | 0.5375 | 0.5831 | 2.4800 | 0.5375 | 0.5366 | 0.1666 | 0.2119 |
No log | 28.0 | 196 | 5.2345 | 0.5275 | 0.6105 | 2.6475 | 0.5275 | 0.5247 | 0.1919 | 0.2338 |
No log | 29.0 | 203 | 5.2050 | 0.5475 | 0.5917 | 2.6350 | 0.5475 | 0.5531 | 0.1753 | 0.2251 |
No log | 30.0 | 210 | 5.1753 | 0.5425 | 0.5891 | 2.6472 | 0.5425 | 0.5282 | 0.1831 | 0.2215 |
No log | 31.0 | 217 | 5.2349 | 0.535 | 0.5946 | 2.5653 | 0.535 | 0.5257 | 0.1617 | 0.2186 |
No log | 32.0 | 224 | 5.1497 | 0.545 | 0.5778 | 2.6174 | 0.545 | 0.5425 | 0.1716 | 0.2138 |
No log | 33.0 | 231 | 5.1688 | 0.5175 | 0.5899 | 2.5079 | 0.5175 | 0.5149 | 0.1624 | 0.2159 |
No log | 34.0 | 238 | 5.2269 | 0.53 | 0.5961 | 2.5188 | 0.53 | 0.5326 | 0.1746 | 0.2206 |
No log | 35.0 | 245 | 5.1477 | 0.5325 | 0.5867 | 2.4762 | 0.5325 | 0.5369 | 0.1728 | 0.2176 |
No log | 36.0 | 252 | 5.2229 | 0.5375 | 0.5838 | 2.4397 | 0.5375 | 0.5386 | 0.1693 | 0.2167 |
No log | 37.0 | 259 | 5.1578 | 0.535 | 0.5802 | 2.5103 | 0.535 | 0.5286 | 0.1755 | 0.2124 |
No log | 38.0 | 266 | 5.1405 | 0.535 | 0.5979 | 2.5852 | 0.535 | 0.5346 | 0.1913 | 0.2268 |
No log | 39.0 | 273 | 5.1236 | 0.535 | 0.5844 | 2.4851 | 0.535 | 0.5378 | 0.1729 | 0.2168 |
No log | 40.0 | 280 | 5.0813 | 0.5475 | 0.5757 | 2.4305 | 0.5475 | 0.5434 | 0.1781 | 0.2091 |
No log | 41.0 | 287 | 5.1844 | 0.535 | 0.5888 | 2.4730 | 0.535 | 0.5306 | 0.1707 | 0.2159 |
No log | 42.0 | 294 | 5.1468 | 0.53 | 0.5926 | 2.4866 | 0.53 | 0.5316 | 0.1776 | 0.2200 |
No log | 43.0 | 301 | 5.1469 | 0.53 | 0.5837 | 2.5769 | 0.53 | 0.5252 | 0.1805 | 0.2168 |
No log | 44.0 | 308 | 5.2168 | 0.54 | 0.5955 | 2.5216 | 0.54 | 0.5419 | 0.1689 | 0.2226 |
No log | 45.0 | 315 | 5.1395 | 0.525 | 0.5861 | 2.4328 | 0.525 | 0.5293 | 0.2006 | 0.2180 |
No log | 46.0 | 322 | 5.1163 | 0.5425 | 0.5822 | 2.4635 | 0.5425 | 0.5416 | 0.1937 | 0.2106 |
No log | 47.0 | 329 | 5.1227 | 0.5475 | 0.5786 | 2.5198 | 0.5475 | 0.5489 | 0.1580 | 0.2111 |
No log | 48.0 | 336 | 5.1134 | 0.5375 | 0.5839 | 2.5239 | 0.5375 | 0.5318 | 0.1832 | 0.2071 |
No log | 49.0 | 343 | 5.1907 | 0.5375 | 0.5913 | 2.5012 | 0.5375 | 0.5334 | 0.1853 | 0.2145 |
No log | 50.0 | 350 | 5.1364 | 0.5375 | 0.5875 | 2.4105 | 0.5375 | 0.5415 | 0.1857 | 0.2121 |
No log | 51.0 | 357 | 5.1739 | 0.5425 | 0.5905 | 2.5208 | 0.5425 | 0.5399 | 0.1894 | 0.2112 |
No log | 52.0 | 364 | 5.1635 | 0.5325 | 0.5841 | 2.4658 | 0.5325 | 0.5300 | 0.1924 | 0.2124 |
No log | 53.0 | 371 | 5.2055 | 0.5425 | 0.5866 | 2.4800 | 0.5425 | 0.5390 | 0.1983 | 0.2135 |
No log | 54.0 | 378 | 5.1547 | 0.5375 | 0.5869 | 2.4575 | 0.5375 | 0.5340 | 0.1839 | 0.2117 |
No log | 55.0 | 385 | 5.1437 | 0.535 | 0.5838 | 2.4117 | 0.535 | 0.5366 | 0.1914 | 0.2110 |
No log | 56.0 | 392 | 5.2042 | 0.5425 | 0.5915 | 2.4286 | 0.5425 | 0.5445 | 0.1905 | 0.2124 |
No log | 57.0 | 399 | 5.2084 | 0.5625 | 0.5909 | 2.4774 | 0.5625 | 0.5646 | 0.2006 | 0.2116 |
No log | 58.0 | 406 | 5.1844 | 0.545 | 0.5895 | 2.3826 | 0.545 | 0.5466 | 0.1948 | 0.2102 |
No log | 59.0 | 413 | 5.1759 | 0.545 | 0.5892 | 2.4790 | 0.545 | 0.5498 | 0.1730 | 0.2143 |
No log | 60.0 | 420 | 5.1783 | 0.5475 | 0.5894 | 2.4294 | 0.5475 | 0.5452 | 0.2043 | 0.2087 |
No log | 61.0 | 427 | 5.1874 | 0.545 | 0.5879 | 2.4295 | 0.545 | 0.5412 | 0.1959 | 0.2080 |
No log | 62.0 | 434 | 5.1861 | 0.5475 | 0.5840 | 2.4513 | 0.5475 | 0.5430 | 0.2097 | 0.2107 |
No log | 63.0 | 441 | 5.1608 | 0.545 | 0.5818 | 2.4581 | 0.545 | 0.5450 | 0.1666 | 0.2055 |
No log | 64.0 | 448 | 5.2018 | 0.5475 | 0.5911 | 2.4537 | 0.5475 | 0.5448 | 0.1938 | 0.2113 |
No log | 65.0 | 455 | 5.2113 | 0.5375 | 0.5953 | 2.4444 | 0.5375 | 0.5360 | 0.1757 | 0.2106 |
No log | 66.0 | 462 | 5.1985 | 0.5425 | 0.5897 | 2.4287 | 0.5425 | 0.5377 | 0.1870 | 0.2095 |
No log | 67.0 | 469 | 5.2218 | 0.5325 | 0.5856 | 2.4340 | 0.5325 | 0.5320 | 0.1882 | 0.2059 |
No log | 68.0 | 476 | 5.2243 | 0.545 | 0.5931 | 2.3923 | 0.545 | 0.5447 | 0.1799 | 0.2120 |
No log | 69.0 | 483 | 5.2103 | 0.55 | 0.5881 | 2.4619 | 0.55 | 0.5486 | 0.2084 | 0.2060 |
No log | 70.0 | 490 | 5.2370 | 0.55 | 0.5933 | 2.4236 | 0.55 | 0.5521 | 0.1920 | 0.2108 |
No log | 71.0 | 497 | 5.2185 | 0.5475 | 0.5890 | 2.4137 | 0.5475 | 0.5435 | 0.2121 | 0.2076 |
3.6002 | 72.0 | 504 | 5.2460 | 0.545 | 0.5944 | 2.4704 | 0.545 | 0.5430 | 0.1922 | 0.2117 |
3.6002 | 73.0 | 511 | 5.2454 | 0.5425 | 0.5928 | 2.4750 | 0.5425 | 0.5406 | 0.1940 | 0.2080 |
3.6002 | 74.0 | 518 | 5.2307 | 0.5575 | 0.5935 | 2.4623 | 0.5575 | 0.5599 | 0.1959 | 0.2071 |
3.6002 | 75.0 | 525 | 5.2674 | 0.56 | 0.5877 | 2.4453 | 0.56 | 0.5587 | 0.1956 | 0.2033 |
3.6002 | 76.0 | 532 | 5.2263 | 0.5525 | 0.5907 | 2.5044 | 0.5525 | 0.5526 | 0.1862 | 0.2067 |
3.6002 | 77.0 | 539 | 5.2498 | 0.55 | 0.5938 | 2.4668 | 0.55 | 0.5467 | 0.2072 | 0.2059 |
3.6002 | 78.0 | 546 | 5.2671 | 0.545 | 0.5961 | 2.4394 | 0.545 | 0.5421 | 0.2056 | 0.2093 |
3.6002 | 79.0 | 553 | 5.2923 | 0.545 | 0.5950 | 2.4662 | 0.545 | 0.5455 | 0.1833 | 0.2058 |
3.6002 | 80.0 | 560 | 5.2854 | 0.555 | 0.5918 | 2.5010 | 0.555 | 0.5526 | 0.2040 | 0.2059 |
3.6002 | 81.0 | 567 | 5.3009 | 0.535 | 0.5955 | 2.4253 | 0.535 | 0.5319 | 0.1939 | 0.2101 |
3.6002 | 82.0 | 574 | 5.3016 | 0.535 | 0.5979 | 2.4528 | 0.535 | 0.5315 | 0.2020 | 0.2101 |
3.6002 | 83.0 | 581 | 5.3262 | 0.545 | 0.5990 | 2.4245 | 0.545 | 0.5422 | 0.1816 | 0.2081 |
3.6002 | 84.0 | 588 | 5.3206 | 0.535 | 0.5990 | 2.4519 | 0.535 | 0.5350 | 0.1959 | 0.2121 |
3.6002 | 85.0 | 595 | 5.3333 | 0.5375 | 0.5999 | 2.4909 | 0.5375 | 0.5352 | 0.1881 | 0.2109 |
3.6002 | 86.0 | 602 | 5.3407 | 0.535 | 0.6008 | 2.5019 | 0.535 | 0.5331 | 0.2087 | 0.2096 |
3.6002 | 87.0 | 609 | 5.3413 | 0.5425 | 0.6015 | 2.4753 | 0.5425 | 0.5402 | 0.2147 | 0.2101 |
3.6002 | 88.0 | 616 | 5.3716 | 0.5375 | 0.6041 | 2.4290 | 0.5375 | 0.5373 | 0.2234 | 0.2094 |
3.6002 | 89.0 | 623 | 5.3639 | 0.535 | 0.6010 | 2.4159 | 0.535 | 0.5319 | 0.2068 | 0.2108 |
3.6002 | 90.0 | 630 | 5.3742 | 0.5425 | 0.6030 | 2.4588 | 0.5425 | 0.5420 | 0.2021 | 0.2099 |
3.6002 | 91.0 | 637 | 5.3731 | 0.53 | 0.6046 | 2.4580 | 0.53 | 0.5284 | 0.2193 | 0.2122 |
3.6002 | 92.0 | 644 | 5.3919 | 0.54 | 0.6051 | 2.4317 | 0.54 | 0.5395 | 0.2057 | 0.2090 |
3.6002 | 93.0 | 651 | 5.3947 | 0.54 | 0.6049 | 2.4372 | 0.54 | 0.5385 | 0.2053 | 0.2092 |
3.6002 | 94.0 | 658 | 5.4070 | 0.535 | 0.6067 | 2.4600 | 0.535 | 0.5328 | 0.2297 | 0.2108 |
3.6002 | 95.0 | 665 | 5.4129 | 0.535 | 0.6071 | 2.4249 | 0.535 | 0.5345 | 0.2186 | 0.2104 |
3.6002 | 96.0 | 672 | 5.4137 | 0.535 | 0.6071 | 2.4580 | 0.535 | 0.5339 | 0.2192 | 0.2102 |
3.6002 | 97.0 | 679 | 5.4218 | 0.5325 | 0.6079 | 2.4584 | 0.5325 | 0.5316 | 0.2197 | 0.2112 |
3.6002 | 98.0 | 686 | 5.4236 | 0.5325 | 0.6083 | 2.4585 | 0.5325 | 0.5311 | 0.2154 | 0.2110 |
3.6002 | 99.0 | 693 | 5.4261 | 0.5325 | 0.6081 | 2.4569 | 0.5325 | 0.5316 | 0.2294 | 0.2114 |
3.6002 | 100.0 | 700 | 5.4262 | 0.535 | 0.6080 | 2.4569 | 0.535 | 0.5345 | 0.2120 | 0.2105 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2