<!-- 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_a0.9
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: 1.6228
- Accuracy: 0.5375
- Brier Loss: 0.6009
- Nll: 2.3362
- F1 Micro: 0.5375
- F1 Macro: 0.5332
- Ece: 0.1886
- Aurc: 0.2188
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 | 4.0859 | 0.0425 | 1.0718 | 7.3724 | 0.0425 | 0.0340 | 0.2811 | 0.9554 |
No log | 2.0 | 14 | 3.1575 | 0.0925 | 0.9517 | 5.8523 | 0.0925 | 0.0709 | 0.1660 | 0.8903 |
No log | 3.0 | 21 | 2.8291 | 0.2325 | 0.8866 | 5.4080 | 0.2325 | 0.1824 | 0.1601 | 0.6363 |
No log | 4.0 | 28 | 2.4189 | 0.2875 | 0.7987 | 3.5373 | 0.2875 | 0.2714 | 0.1822 | 0.4858 |
No log | 5.0 | 35 | 2.1338 | 0.3775 | 0.7203 | 2.9804 | 0.3775 | 0.3569 | 0.1731 | 0.3581 |
No log | 6.0 | 42 | 2.0168 | 0.44 | 0.6910 | 2.9903 | 0.44 | 0.4255 | 0.1854 | 0.3060 |
No log | 7.0 | 49 | 2.0458 | 0.4575 | 0.6988 | 3.1988 | 0.4575 | 0.4286 | 0.2078 | 0.3000 |
No log | 8.0 | 56 | 1.9688 | 0.4575 | 0.6928 | 2.8691 | 0.4575 | 0.4542 | 0.2122 | 0.2986 |
No log | 9.0 | 63 | 2.1143 | 0.4625 | 0.7081 | 2.8507 | 0.4625 | 0.4496 | 0.2232 | 0.2940 |
No log | 10.0 | 70 | 2.1352 | 0.455 | 0.7242 | 2.9232 | 0.455 | 0.4317 | 0.2581 | 0.2923 |
No log | 11.0 | 77 | 2.2760 | 0.44 | 0.7545 | 2.9083 | 0.44 | 0.4144 | 0.2835 | 0.3144 |
No log | 12.0 | 84 | 2.1320 | 0.4675 | 0.7389 | 2.6277 | 0.4675 | 0.4607 | 0.2715 | 0.3142 |
No log | 13.0 | 91 | 2.0889 | 0.4825 | 0.7170 | 2.6786 | 0.4825 | 0.4755 | 0.2522 | 0.2933 |
No log | 14.0 | 98 | 2.0173 | 0.475 | 0.7020 | 2.7765 | 0.4750 | 0.4555 | 0.2593 | 0.2791 |
No log | 15.0 | 105 | 2.1265 | 0.4875 | 0.7124 | 3.0127 | 0.4875 | 0.4753 | 0.2458 | 0.3069 |
No log | 16.0 | 112 | 2.1517 | 0.4825 | 0.7184 | 2.7981 | 0.4825 | 0.4719 | 0.2505 | 0.2964 |
No log | 17.0 | 119 | 1.9485 | 0.49 | 0.6828 | 2.6670 | 0.49 | 0.4773 | 0.2430 | 0.2649 |
No log | 18.0 | 126 | 1.9434 | 0.5 | 0.6802 | 2.6131 | 0.5 | 0.4941 | 0.2267 | 0.2741 |
No log | 19.0 | 133 | 1.9055 | 0.4925 | 0.6741 | 2.7035 | 0.4925 | 0.4953 | 0.2201 | 0.2700 |
No log | 20.0 | 140 | 1.8119 | 0.5 | 0.6520 | 2.4537 | 0.5 | 0.4937 | 0.2016 | 0.2449 |
No log | 21.0 | 147 | 1.8420 | 0.49 | 0.6593 | 2.6244 | 0.49 | 0.4798 | 0.2100 | 0.2668 |
No log | 22.0 | 154 | 1.9282 | 0.5 | 0.6739 | 2.6650 | 0.5 | 0.4837 | 0.2166 | 0.2652 |
No log | 23.0 | 161 | 1.8139 | 0.5125 | 0.6543 | 2.4043 | 0.5125 | 0.5009 | 0.2275 | 0.2604 |
No log | 24.0 | 168 | 1.6997 | 0.5325 | 0.6124 | 2.3950 | 0.5325 | 0.5271 | 0.1873 | 0.2248 |
No log | 25.0 | 175 | 1.8370 | 0.5025 | 0.6526 | 2.4967 | 0.5025 | 0.4926 | 0.2249 | 0.2513 |
No log | 26.0 | 182 | 1.7508 | 0.5025 | 0.6388 | 2.4321 | 0.5025 | 0.4947 | 0.2135 | 0.2416 |
No log | 27.0 | 189 | 1.7354 | 0.52 | 0.6368 | 2.3706 | 0.52 | 0.5149 | 0.2101 | 0.2440 |
No log | 28.0 | 196 | 1.7809 | 0.52 | 0.6421 | 2.4895 | 0.52 | 0.5076 | 0.2033 | 0.2446 |
No log | 29.0 | 203 | 1.6770 | 0.55 | 0.6046 | 2.3913 | 0.55 | 0.5409 | 0.1849 | 0.2189 |
No log | 30.0 | 210 | 1.6794 | 0.5325 | 0.6127 | 2.5677 | 0.5325 | 0.5263 | 0.2095 | 0.2247 |
No log | 31.0 | 217 | 1.7352 | 0.525 | 0.6325 | 2.5068 | 0.525 | 0.5062 | 0.1865 | 0.2320 |
No log | 32.0 | 224 | 1.7396 | 0.5225 | 0.6379 | 2.4798 | 0.5225 | 0.5169 | 0.1829 | 0.2402 |
No log | 33.0 | 231 | 1.7013 | 0.52 | 0.6188 | 2.5032 | 0.52 | 0.5029 | 0.1986 | 0.2233 |
No log | 34.0 | 238 | 1.7121 | 0.5075 | 0.6313 | 2.5680 | 0.5075 | 0.5074 | 0.2249 | 0.2482 |
No log | 35.0 | 245 | 1.7053 | 0.5175 | 0.6368 | 2.3967 | 0.5175 | 0.5064 | 0.2088 | 0.2389 |
No log | 36.0 | 252 | 1.6616 | 0.55 | 0.6058 | 2.5726 | 0.55 | 0.5357 | 0.1762 | 0.2174 |
No log | 37.0 | 259 | 1.7302 | 0.5075 | 0.6350 | 2.3745 | 0.5075 | 0.4993 | 0.1939 | 0.2481 |
No log | 38.0 | 266 | 1.6741 | 0.5225 | 0.6152 | 2.5862 | 0.5225 | 0.5211 | 0.2042 | 0.2234 |
No log | 39.0 | 273 | 1.6960 | 0.515 | 0.6216 | 2.4210 | 0.515 | 0.5049 | 0.1776 | 0.2270 |
No log | 40.0 | 280 | 1.6512 | 0.54 | 0.6075 | 2.4495 | 0.54 | 0.5377 | 0.1848 | 0.2217 |
No log | 41.0 | 287 | 1.6477 | 0.52 | 0.6086 | 2.5354 | 0.52 | 0.5177 | 0.1976 | 0.2219 |
No log | 42.0 | 294 | 1.6680 | 0.52 | 0.6198 | 2.2815 | 0.52 | 0.5215 | 0.1951 | 0.2314 |
No log | 43.0 | 301 | 1.6153 | 0.545 | 0.5935 | 2.3952 | 0.545 | 0.5376 | 0.1721 | 0.2088 |
No log | 44.0 | 308 | 1.6347 | 0.5375 | 0.6023 | 2.3985 | 0.5375 | 0.5340 | 0.1778 | 0.2178 |
No log | 45.0 | 315 | 1.6132 | 0.5375 | 0.6001 | 2.3149 | 0.5375 | 0.5375 | 0.1902 | 0.2163 |
No log | 46.0 | 322 | 1.6366 | 0.5275 | 0.6102 | 2.3820 | 0.5275 | 0.5210 | 0.2057 | 0.2239 |
No log | 47.0 | 329 | 1.6118 | 0.55 | 0.5963 | 2.3834 | 0.55 | 0.5463 | 0.1748 | 0.2137 |
No log | 48.0 | 336 | 1.6208 | 0.535 | 0.6038 | 2.2855 | 0.535 | 0.5340 | 0.1791 | 0.2201 |
No log | 49.0 | 343 | 1.6006 | 0.545 | 0.5960 | 2.3253 | 0.545 | 0.5440 | 0.2007 | 0.2158 |
No log | 50.0 | 350 | 1.6184 | 0.5425 | 0.6005 | 2.3181 | 0.5425 | 0.5432 | 0.1846 | 0.2176 |
No log | 51.0 | 357 | 1.6215 | 0.5425 | 0.6022 | 2.3219 | 0.5425 | 0.5371 | 0.1776 | 0.2163 |
No log | 52.0 | 364 | 1.6068 | 0.54 | 0.5978 | 2.2928 | 0.54 | 0.5383 | 0.1865 | 0.2164 |
No log | 53.0 | 371 | 1.6128 | 0.535 | 0.5979 | 2.3243 | 0.535 | 0.5335 | 0.1913 | 0.2189 |
No log | 54.0 | 378 | 1.6182 | 0.545 | 0.6002 | 2.3604 | 0.545 | 0.5432 | 0.1740 | 0.2157 |
No log | 55.0 | 385 | 1.6143 | 0.54 | 0.5981 | 2.3632 | 0.54 | 0.5348 | 0.1911 | 0.2165 |
No log | 56.0 | 392 | 1.6163 | 0.5375 | 0.6010 | 2.3551 | 0.5375 | 0.5338 | 0.1950 | 0.2191 |
No log | 57.0 | 399 | 1.6125 | 0.54 | 0.5984 | 2.3597 | 0.54 | 0.5378 | 0.1680 | 0.2159 |
No log | 58.0 | 406 | 1.6159 | 0.5425 | 0.5997 | 2.3564 | 0.5425 | 0.5388 | 0.1796 | 0.2175 |
No log | 59.0 | 413 | 1.6102 | 0.5475 | 0.5967 | 2.3879 | 0.5475 | 0.5448 | 0.1863 | 0.2152 |
No log | 60.0 | 420 | 1.6171 | 0.535 | 0.6006 | 2.3594 | 0.535 | 0.5298 | 0.1987 | 0.2183 |
No log | 61.0 | 427 | 1.6162 | 0.5375 | 0.5987 | 2.3631 | 0.5375 | 0.5341 | 0.1836 | 0.2172 |
No log | 62.0 | 434 | 1.6167 | 0.5425 | 0.5992 | 2.3344 | 0.5425 | 0.5395 | 0.2011 | 0.2170 |
No log | 63.0 | 441 | 1.6123 | 0.545 | 0.5984 | 2.3575 | 0.545 | 0.5423 | 0.1946 | 0.2174 |
No log | 64.0 | 448 | 1.6209 | 0.535 | 0.6007 | 2.3333 | 0.535 | 0.5299 | 0.1973 | 0.2180 |
No log | 65.0 | 455 | 1.6184 | 0.5475 | 0.6003 | 2.3620 | 0.5475 | 0.5442 | 0.1911 | 0.2167 |
No log | 66.0 | 462 | 1.6179 | 0.5375 | 0.5992 | 2.3321 | 0.5375 | 0.5333 | 0.2019 | 0.2182 |
No log | 67.0 | 469 | 1.6179 | 0.5425 | 0.5999 | 2.3309 | 0.5425 | 0.5397 | 0.1780 | 0.2176 |
No log | 68.0 | 476 | 1.6204 | 0.545 | 0.6004 | 2.3282 | 0.545 | 0.5408 | 0.1740 | 0.2167 |
No log | 69.0 | 483 | 1.6195 | 0.54 | 0.6005 | 2.3372 | 0.54 | 0.5370 | 0.1898 | 0.2187 |
No log | 70.0 | 490 | 1.6189 | 0.5425 | 0.5996 | 2.3328 | 0.5425 | 0.5398 | 0.1754 | 0.2169 |
No log | 71.0 | 497 | 1.6202 | 0.5425 | 0.6007 | 2.3352 | 0.5425 | 0.5394 | 0.1973 | 0.2180 |
0.3297 | 72.0 | 504 | 1.6175 | 0.5375 | 0.5997 | 2.3297 | 0.5375 | 0.5348 | 0.1721 | 0.2184 |
0.3297 | 73.0 | 511 | 1.6185 | 0.545 | 0.6001 | 2.3339 | 0.545 | 0.5414 | 0.1780 | 0.2175 |
0.3297 | 74.0 | 518 | 1.6203 | 0.5425 | 0.6003 | 2.3315 | 0.5425 | 0.5391 | 0.1855 | 0.2175 |
0.3297 | 75.0 | 525 | 1.6199 | 0.54 | 0.6002 | 2.3366 | 0.54 | 0.5363 | 0.1841 | 0.2180 |
0.3297 | 76.0 | 532 | 1.6203 | 0.5425 | 0.6002 | 2.3332 | 0.5425 | 0.5395 | 0.1736 | 0.2176 |
0.3297 | 77.0 | 539 | 1.6191 | 0.5425 | 0.6000 | 2.3328 | 0.5425 | 0.5395 | 0.1722 | 0.2176 |
0.3297 | 78.0 | 546 | 1.6202 | 0.54 | 0.6001 | 2.3317 | 0.54 | 0.5362 | 0.1909 | 0.2177 |
0.3297 | 79.0 | 553 | 1.6199 | 0.54 | 0.6002 | 2.3335 | 0.54 | 0.5370 | 0.1982 | 0.2181 |
0.3297 | 80.0 | 560 | 1.6207 | 0.54 | 0.6006 | 2.3336 | 0.54 | 0.5363 | 0.1983 | 0.2185 |
0.3297 | 81.0 | 567 | 1.6210 | 0.5425 | 0.6003 | 2.3353 | 0.5425 | 0.5389 | 0.1875 | 0.2174 |
0.3297 | 82.0 | 574 | 1.6216 | 0.5425 | 0.6005 | 2.3340 | 0.5425 | 0.5386 | 0.1792 | 0.2178 |
0.3297 | 83.0 | 581 | 1.6213 | 0.5375 | 0.6007 | 2.3360 | 0.5375 | 0.5340 | 0.1797 | 0.2193 |
0.3297 | 84.0 | 588 | 1.6208 | 0.54 | 0.6007 | 2.3345 | 0.54 | 0.5363 | 0.1848 | 0.2189 |
0.3297 | 85.0 | 595 | 1.6221 | 0.54 | 0.6005 | 2.3360 | 0.54 | 0.5369 | 0.1836 | 0.2184 |
0.3297 | 86.0 | 602 | 1.6218 | 0.5425 | 0.6005 | 2.3342 | 0.5425 | 0.5391 | 0.1922 | 0.2178 |
0.3297 | 87.0 | 609 | 1.6224 | 0.54 | 0.6006 | 2.3365 | 0.54 | 0.5363 | 0.1849 | 0.2185 |
0.3297 | 88.0 | 616 | 1.6216 | 0.54 | 0.6005 | 2.3355 | 0.54 | 0.5363 | 0.1912 | 0.2188 |
0.3297 | 89.0 | 623 | 1.6222 | 0.535 | 0.6008 | 2.3359 | 0.535 | 0.5310 | 0.1890 | 0.2195 |
0.3297 | 90.0 | 630 | 1.6218 | 0.5425 | 0.6007 | 2.3356 | 0.5425 | 0.5386 | 0.1930 | 0.2179 |
0.3297 | 91.0 | 637 | 1.6226 | 0.54 | 0.6009 | 2.3363 | 0.54 | 0.5363 | 0.1941 | 0.2188 |
0.3297 | 92.0 | 644 | 1.6224 | 0.54 | 0.6009 | 2.3360 | 0.54 | 0.5363 | 0.1890 | 0.2188 |
0.3297 | 93.0 | 651 | 1.6226 | 0.5425 | 0.6009 | 2.3359 | 0.5425 | 0.5386 | 0.1964 | 0.2179 |
0.3297 | 94.0 | 658 | 1.6227 | 0.54 | 0.6009 | 2.3363 | 0.54 | 0.5363 | 0.1877 | 0.2187 |
0.3297 | 95.0 | 665 | 1.6226 | 0.54 | 0.6008 | 2.3360 | 0.54 | 0.5363 | 0.1916 | 0.2186 |
0.3297 | 96.0 | 672 | 1.6225 | 0.54 | 0.6008 | 2.3362 | 0.54 | 0.5363 | 0.1972 | 0.2186 |
0.3297 | 97.0 | 679 | 1.6227 | 0.5375 | 0.6009 | 2.3362 | 0.5375 | 0.5332 | 0.1949 | 0.2188 |
0.3297 | 98.0 | 686 | 1.6228 | 0.5375 | 0.6009 | 2.3362 | 0.5375 | 0.5332 | 0.1924 | 0.2188 |
0.3297 | 99.0 | 693 | 1.6228 | 0.5375 | 0.6009 | 2.3363 | 0.5375 | 0.5332 | 0.1924 | 0.2189 |
0.3297 | 100.0 | 700 | 1.6228 | 0.5375 | 0.6009 | 2.3362 | 0.5375 | 0.5332 | 0.1886 | 0.2188 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2