<!-- 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-small_tobacco3482_kd_CEKD_t5.0_a0.7
This model is a fine-tuned version of WinKawaks/vit-small-patch16-224 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.4918
- Accuracy: 0.85
- Brier Loss: 0.2583
- Nll: 1.0894
- F1 Micro: 0.85
- F1 Macro: 0.8374
- Ece: 0.1917
- Aurc: 0.0470
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 | 1.8329 | 0.225 | 0.8761 | 5.2731 | 0.225 | 0.1384 | 0.2607 | 0.6977 |
No log | 2.0 | 14 | 1.4785 | 0.405 | 0.7460 | 3.4067 | 0.405 | 0.2289 | 0.3097 | 0.4085 |
No log | 3.0 | 21 | 1.0406 | 0.6 | 0.5725 | 1.8722 | 0.6 | 0.5345 | 0.3050 | 0.2010 |
No log | 4.0 | 28 | 0.8087 | 0.725 | 0.4192 | 1.6096 | 0.7250 | 0.6767 | 0.2345 | 0.1149 |
No log | 5.0 | 35 | 0.7666 | 0.735 | 0.3731 | 1.6189 | 0.735 | 0.7350 | 0.2377 | 0.1011 |
No log | 6.0 | 42 | 0.6960 | 0.78 | 0.3413 | 1.5230 | 0.78 | 0.7592 | 0.2295 | 0.0868 |
No log | 7.0 | 49 | 0.6490 | 0.805 | 0.3110 | 1.4861 | 0.805 | 0.7864 | 0.2138 | 0.0785 |
No log | 8.0 | 56 | 0.6238 | 0.795 | 0.3069 | 1.2098 | 0.795 | 0.7816 | 0.2065 | 0.0698 |
No log | 9.0 | 63 | 0.5755 | 0.83 | 0.2866 | 1.1943 | 0.83 | 0.8117 | 0.1937 | 0.0694 |
No log | 10.0 | 70 | 0.6360 | 0.77 | 0.3164 | 1.2608 | 0.7700 | 0.7550 | 0.1785 | 0.0677 |
No log | 11.0 | 77 | 0.6548 | 0.785 | 0.3335 | 1.4895 | 0.785 | 0.7707 | 0.2281 | 0.0885 |
No log | 12.0 | 84 | 0.5847 | 0.805 | 0.3002 | 1.4317 | 0.805 | 0.7807 | 0.2264 | 0.0756 |
No log | 13.0 | 91 | 0.5956 | 0.81 | 0.3040 | 1.2590 | 0.81 | 0.7928 | 0.2241 | 0.0556 |
No log | 14.0 | 98 | 0.5692 | 0.81 | 0.3025 | 1.2119 | 0.81 | 0.8043 | 0.2235 | 0.0665 |
No log | 15.0 | 105 | 0.5223 | 0.83 | 0.2762 | 1.1162 | 0.83 | 0.8221 | 0.1798 | 0.0552 |
No log | 16.0 | 112 | 0.4981 | 0.84 | 0.2523 | 1.0864 | 0.8400 | 0.8372 | 0.1868 | 0.0396 |
No log | 17.0 | 119 | 0.5207 | 0.805 | 0.2741 | 1.0416 | 0.805 | 0.7897 | 0.1960 | 0.0551 |
No log | 18.0 | 126 | 0.5165 | 0.84 | 0.2723 | 1.1596 | 0.8400 | 0.8325 | 0.1942 | 0.0506 |
No log | 19.0 | 133 | 0.4979 | 0.845 | 0.2573 | 1.2329 | 0.845 | 0.8297 | 0.1825 | 0.0444 |
No log | 20.0 | 140 | 0.4953 | 0.855 | 0.2565 | 1.1213 | 0.855 | 0.8442 | 0.1844 | 0.0474 |
No log | 21.0 | 147 | 0.5296 | 0.82 | 0.2792 | 1.0000 | 0.82 | 0.8218 | 0.1768 | 0.0523 |
No log | 22.0 | 154 | 0.5027 | 0.835 | 0.2625 | 0.9926 | 0.835 | 0.8238 | 0.2035 | 0.0481 |
No log | 23.0 | 161 | 0.5027 | 0.84 | 0.2642 | 1.0500 | 0.8400 | 0.8299 | 0.1616 | 0.0482 |
No log | 24.0 | 168 | 0.5017 | 0.84 | 0.2616 | 1.0560 | 0.8400 | 0.8314 | 0.1819 | 0.0497 |
No log | 25.0 | 175 | 0.4942 | 0.85 | 0.2594 | 1.1003 | 0.85 | 0.8407 | 0.1793 | 0.0483 |
No log | 26.0 | 182 | 0.4943 | 0.83 | 0.2586 | 1.0436 | 0.83 | 0.8140 | 0.1869 | 0.0518 |
No log | 27.0 | 189 | 0.4950 | 0.835 | 0.2613 | 1.0817 | 0.835 | 0.8224 | 0.2039 | 0.0504 |
No log | 28.0 | 196 | 0.4957 | 0.85 | 0.2599 | 1.1109 | 0.85 | 0.8309 | 0.2058 | 0.0485 |
No log | 29.0 | 203 | 0.4956 | 0.845 | 0.2599 | 1.0914 | 0.845 | 0.8304 | 0.1916 | 0.0492 |
No log | 30.0 | 210 | 0.4893 | 0.84 | 0.2561 | 1.0890 | 0.8400 | 0.8214 | 0.2071 | 0.0482 |
No log | 31.0 | 217 | 0.4920 | 0.835 | 0.2587 | 1.0907 | 0.835 | 0.8270 | 0.2031 | 0.0482 |
No log | 32.0 | 224 | 0.4927 | 0.83 | 0.2601 | 1.0879 | 0.83 | 0.8157 | 0.2093 | 0.0500 |
No log | 33.0 | 231 | 0.4925 | 0.835 | 0.2593 | 1.0886 | 0.835 | 0.8270 | 0.1810 | 0.0484 |
No log | 34.0 | 238 | 0.4909 | 0.845 | 0.2578 | 1.0871 | 0.845 | 0.8304 | 0.1916 | 0.0478 |
No log | 35.0 | 245 | 0.4927 | 0.845 | 0.2591 | 1.0866 | 0.845 | 0.8378 | 0.1943 | 0.0473 |
No log | 36.0 | 252 | 0.4919 | 0.85 | 0.2581 | 1.0891 | 0.85 | 0.8342 | 0.2193 | 0.0475 |
No log | 37.0 | 259 | 0.4908 | 0.845 | 0.2579 | 1.0867 | 0.845 | 0.8346 | 0.2215 | 0.0474 |
No log | 38.0 | 266 | 0.4929 | 0.85 | 0.2590 | 1.0873 | 0.85 | 0.8407 | 0.1884 | 0.0471 |
No log | 39.0 | 273 | 0.4913 | 0.85 | 0.2584 | 1.0861 | 0.85 | 0.8374 | 0.1944 | 0.0474 |
No log | 40.0 | 280 | 0.4933 | 0.835 | 0.2595 | 1.0871 | 0.835 | 0.8248 | 0.1893 | 0.0491 |
No log | 41.0 | 287 | 0.4936 | 0.84 | 0.2599 | 1.0863 | 0.8400 | 0.8276 | 0.1860 | 0.0486 |
No log | 42.0 | 294 | 0.4911 | 0.85 | 0.2580 | 1.0861 | 0.85 | 0.8374 | 0.2186 | 0.0474 |
No log | 43.0 | 301 | 0.4915 | 0.85 | 0.2581 | 1.0860 | 0.85 | 0.8374 | 0.2023 | 0.0475 |
No log | 44.0 | 308 | 0.4921 | 0.85 | 0.2586 | 1.0874 | 0.85 | 0.8374 | 0.2013 | 0.0477 |
No log | 45.0 | 315 | 0.4915 | 0.85 | 0.2583 | 1.0862 | 0.85 | 0.8374 | 0.1941 | 0.0475 |
No log | 46.0 | 322 | 0.4918 | 0.85 | 0.2584 | 1.0878 | 0.85 | 0.8374 | 0.1852 | 0.0473 |
No log | 47.0 | 329 | 0.4916 | 0.85 | 0.2583 | 1.0873 | 0.85 | 0.8374 | 0.2089 | 0.0473 |
No log | 48.0 | 336 | 0.4921 | 0.85 | 0.2586 | 1.0879 | 0.85 | 0.8374 | 0.2026 | 0.0477 |
No log | 49.0 | 343 | 0.4918 | 0.845 | 0.2584 | 1.0884 | 0.845 | 0.8282 | 0.1963 | 0.0478 |
No log | 50.0 | 350 | 0.4922 | 0.85 | 0.2587 | 1.0871 | 0.85 | 0.8374 | 0.2102 | 0.0474 |
No log | 51.0 | 357 | 0.4920 | 0.85 | 0.2585 | 1.0879 | 0.85 | 0.8374 | 0.2095 | 0.0474 |
No log | 52.0 | 364 | 0.4926 | 0.85 | 0.2589 | 1.0878 | 0.85 | 0.8374 | 0.2022 | 0.0477 |
No log | 53.0 | 371 | 0.4920 | 0.85 | 0.2586 | 1.0888 | 0.85 | 0.8374 | 0.2027 | 0.0475 |
No log | 54.0 | 378 | 0.4921 | 0.85 | 0.2586 | 1.0886 | 0.85 | 0.8374 | 0.2020 | 0.0474 |
No log | 55.0 | 385 | 0.4921 | 0.85 | 0.2587 | 1.0890 | 0.85 | 0.8374 | 0.1929 | 0.0471 |
No log | 56.0 | 392 | 0.4925 | 0.85 | 0.2589 | 1.0881 | 0.85 | 0.8374 | 0.1946 | 0.0473 |
No log | 57.0 | 399 | 0.4917 | 0.85 | 0.2583 | 1.0893 | 0.85 | 0.8374 | 0.1932 | 0.0472 |
No log | 58.0 | 406 | 0.4921 | 0.85 | 0.2586 | 1.0877 | 0.85 | 0.8374 | 0.1948 | 0.0476 |
No log | 59.0 | 413 | 0.4917 | 0.85 | 0.2583 | 1.0883 | 0.85 | 0.8374 | 0.1931 | 0.0472 |
No log | 60.0 | 420 | 0.4918 | 0.85 | 0.2583 | 1.0882 | 0.85 | 0.8374 | 0.1945 | 0.0475 |
No log | 61.0 | 427 | 0.4916 | 0.85 | 0.2582 | 1.0883 | 0.85 | 0.8374 | 0.1936 | 0.0472 |
No log | 62.0 | 434 | 0.4920 | 0.85 | 0.2586 | 1.0882 | 0.85 | 0.8374 | 0.1942 | 0.0473 |
No log | 63.0 | 441 | 0.4922 | 0.85 | 0.2587 | 1.0889 | 0.85 | 0.8374 | 0.1935 | 0.0473 |
No log | 64.0 | 448 | 0.4921 | 0.85 | 0.2586 | 1.0885 | 0.85 | 0.8374 | 0.1848 | 0.0473 |
No log | 65.0 | 455 | 0.4916 | 0.85 | 0.2582 | 1.0887 | 0.85 | 0.8374 | 0.1848 | 0.0474 |
No log | 66.0 | 462 | 0.4917 | 0.85 | 0.2583 | 1.0883 | 0.85 | 0.8374 | 0.1849 | 0.0472 |
No log | 67.0 | 469 | 0.4917 | 0.85 | 0.2584 | 1.0887 | 0.85 | 0.8374 | 0.1848 | 0.0472 |
No log | 68.0 | 476 | 0.4920 | 0.85 | 0.2585 | 1.0888 | 0.85 | 0.8374 | 0.2011 | 0.0471 |
No log | 69.0 | 483 | 0.4918 | 0.85 | 0.2584 | 1.0889 | 0.85 | 0.8374 | 0.2007 | 0.0471 |
No log | 70.0 | 490 | 0.4919 | 0.85 | 0.2584 | 1.0886 | 0.85 | 0.8374 | 0.1848 | 0.0474 |
No log | 71.0 | 497 | 0.4920 | 0.85 | 0.2585 | 1.0888 | 0.85 | 0.8374 | 0.1940 | 0.0474 |
0.1824 | 72.0 | 504 | 0.4919 | 0.85 | 0.2584 | 1.0889 | 0.85 | 0.8374 | 0.2011 | 0.0471 |
0.1824 | 73.0 | 511 | 0.4917 | 0.85 | 0.2583 | 1.0887 | 0.85 | 0.8374 | 0.1848 | 0.0472 |
0.1824 | 74.0 | 518 | 0.4920 | 0.85 | 0.2585 | 1.0890 | 0.85 | 0.8374 | 0.1848 | 0.0472 |
0.1824 | 75.0 | 525 | 0.4920 | 0.85 | 0.2585 | 1.0892 | 0.85 | 0.8374 | 0.1846 | 0.0472 |
0.1824 | 76.0 | 532 | 0.4918 | 0.85 | 0.2583 | 1.0889 | 0.85 | 0.8374 | 0.1930 | 0.0472 |
0.1824 | 77.0 | 539 | 0.4917 | 0.85 | 0.2582 | 1.0891 | 0.85 | 0.8374 | 0.2005 | 0.0472 |
0.1824 | 78.0 | 546 | 0.4919 | 0.85 | 0.2584 | 1.0892 | 0.85 | 0.8374 | 0.1928 | 0.0472 |
0.1824 | 79.0 | 553 | 0.4920 | 0.85 | 0.2585 | 1.0893 | 0.85 | 0.8374 | 0.1845 | 0.0473 |
0.1824 | 80.0 | 560 | 0.4919 | 0.85 | 0.2584 | 1.0890 | 0.85 | 0.8374 | 0.1929 | 0.0473 |
0.1824 | 81.0 | 567 | 0.4920 | 0.85 | 0.2585 | 1.0892 | 0.85 | 0.8374 | 0.1925 | 0.0471 |
0.1824 | 82.0 | 574 | 0.4920 | 0.85 | 0.2585 | 1.0895 | 0.85 | 0.8374 | 0.1844 | 0.0471 |
0.1824 | 83.0 | 581 | 0.4919 | 0.85 | 0.2584 | 1.0892 | 0.85 | 0.8374 | 0.1916 | 0.0471 |
0.1824 | 84.0 | 588 | 0.4918 | 0.85 | 0.2584 | 1.0890 | 0.85 | 0.8374 | 0.1926 | 0.0471 |
0.1824 | 85.0 | 595 | 0.4918 | 0.85 | 0.2584 | 1.0892 | 0.85 | 0.8374 | 0.1844 | 0.0471 |
0.1824 | 86.0 | 602 | 0.4918 | 0.85 | 0.2584 | 1.0893 | 0.85 | 0.8374 | 0.1927 | 0.0472 |
0.1824 | 87.0 | 609 | 0.4918 | 0.85 | 0.2584 | 1.0895 | 0.85 | 0.8374 | 0.1844 | 0.0471 |
0.1824 | 88.0 | 616 | 0.4918 | 0.85 | 0.2584 | 1.0892 | 0.85 | 0.8374 | 0.1844 | 0.0471 |
0.1824 | 89.0 | 623 | 0.4918 | 0.85 | 0.2583 | 1.0895 | 0.85 | 0.8374 | 0.1917 | 0.0471 |
0.1824 | 90.0 | 630 | 0.4919 | 0.85 | 0.2584 | 1.0892 | 0.85 | 0.8374 | 0.1998 | 0.0471 |
0.1824 | 91.0 | 637 | 0.4919 | 0.85 | 0.2584 | 1.0894 | 0.85 | 0.8374 | 0.1916 | 0.0471 |
0.1824 | 92.0 | 644 | 0.4918 | 0.85 | 0.2583 | 1.0895 | 0.85 | 0.8374 | 0.1917 | 0.0470 |
0.1824 | 93.0 | 651 | 0.4918 | 0.85 | 0.2583 | 1.0893 | 0.85 | 0.8374 | 0.1917 | 0.0471 |
0.1824 | 94.0 | 658 | 0.4918 | 0.85 | 0.2583 | 1.0894 | 0.85 | 0.8374 | 0.1844 | 0.0471 |
0.1824 | 95.0 | 665 | 0.4918 | 0.85 | 0.2583 | 1.0894 | 0.85 | 0.8374 | 0.1917 | 0.0470 |
0.1824 | 96.0 | 672 | 0.4918 | 0.85 | 0.2583 | 1.0894 | 0.85 | 0.8374 | 0.1917 | 0.0470 |
0.1824 | 97.0 | 679 | 0.4918 | 0.85 | 0.2583 | 1.0895 | 0.85 | 0.8374 | 0.1916 | 0.0471 |
0.1824 | 98.0 | 686 | 0.4918 | 0.85 | 0.2583 | 1.0895 | 0.85 | 0.8374 | 0.1917 | 0.0470 |
0.1824 | 99.0 | 693 | 0.4918 | 0.85 | 0.2583 | 1.0894 | 0.85 | 0.8374 | 0.1917 | 0.0470 |
0.1824 | 100.0 | 700 | 0.4918 | 0.85 | 0.2583 | 1.0894 | 0.85 | 0.8374 | 0.1917 | 0.0470 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2