<!-- 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_rvl_cdip_100_examples_per_class_kd_MSE
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: 1.4673
- Accuracy: 0.6425
- Brier Loss: 0.4763
- Nll: 3.0680
- F1 Micro: 0.6425
- F1 Macro: 0.6485
- Ece: 0.1946
- Aurc: 0.1381
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: 32
- eval_batch_size: 32
- 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 | 25 | 4.4851 | 0.06 | 0.9565 | 13.8276 | 0.06 | 0.0556 | 0.1688 | 0.9385 |
| No log | 2.0 | 50 | 3.5619 | 0.3775 | 0.7827 | 6.2649 | 0.3775 | 0.3611 | 0.2331 | 0.3882 |
| No log | 3.0 | 75 | 2.8990 | 0.5025 | 0.6453 | 4.7376 | 0.5025 | 0.4858 | 0.1689 | 0.2658 |
| No log | 4.0 | 100 | 2.5972 | 0.515 | 0.5980 | 4.4210 | 0.515 | 0.4895 | 0.1605 | 0.2249 |
| No log | 5.0 | 125 | 2.4353 | 0.56 | 0.5762 | 3.4885 | 0.56 | 0.5566 | 0.1548 | 0.2100 |
| No log | 6.0 | 150 | 2.4157 | 0.5475 | 0.5864 | 3.8261 | 0.5475 | 0.5323 | 0.1837 | 0.2167 |
| No log | 7.0 | 175 | 2.1786 | 0.6075 | 0.5203 | 3.4565 | 0.6075 | 0.6103 | 0.1403 | 0.1670 |
| No log | 8.0 | 200 | 2.1082 | 0.63 | 0.5040 | 3.3570 | 0.63 | 0.6246 | 0.1580 | 0.1530 |
| No log | 9.0 | 225 | 2.0472 | 0.625 | 0.5042 | 3.8572 | 0.625 | 0.6184 | 0.1552 | 0.1530 |
| No log | 10.0 | 250 | 2.0589 | 0.6025 | 0.5468 | 3.5723 | 0.6025 | 0.5982 | 0.1781 | 0.1785 |
| No log | 11.0 | 275 | 1.8965 | 0.65 | 0.4755 | 3.4466 | 0.65 | 0.6497 | 0.1605 | 0.1475 |
| No log | 12.0 | 300 | 1.9014 | 0.6325 | 0.5066 | 3.0881 | 0.6325 | 0.6359 | 0.1658 | 0.1591 |
| No log | 13.0 | 325 | 1.7904 | 0.6175 | 0.5162 | 3.4673 | 0.6175 | 0.6141 | 0.1525 | 0.1598 |
| No log | 14.0 | 350 | 1.8624 | 0.625 | 0.5173 | 3.6824 | 0.625 | 0.6179 | 0.1567 | 0.1624 |
| No log | 15.0 | 375 | 1.7083 | 0.6625 | 0.4817 | 3.1296 | 0.6625 | 0.6686 | 0.1651 | 0.1405 |
| No log | 16.0 | 400 | 1.8848 | 0.59 | 0.5478 | 4.3761 | 0.59 | 0.5913 | 0.2083 | 0.1696 |
| No log | 17.0 | 425 | 1.7238 | 0.6125 | 0.5229 | 3.1232 | 0.6125 | 0.6052 | 0.1833 | 0.1553 |
| No log | 18.0 | 450 | 1.7126 | 0.625 | 0.5152 | 2.9267 | 0.625 | 0.6284 | 0.1747 | 0.1565 |
| No log | 19.0 | 475 | 1.6459 | 0.6275 | 0.5024 | 2.9078 | 0.6275 | 0.6219 | 0.1766 | 0.1527 |
| 1.0542 | 20.0 | 500 | 1.6029 | 0.6275 | 0.4855 | 3.0931 | 0.6275 | 0.6316 | 0.1720 | 0.1414 |
| 1.0542 | 21.0 | 525 | 1.6566 | 0.6525 | 0.4847 | 3.0998 | 0.6525 | 0.6479 | 0.1558 | 0.1438 |
| 1.0542 | 22.0 | 550 | 1.6169 | 0.645 | 0.4894 | 3.0081 | 0.645 | 0.6471 | 0.1687 | 0.1400 |
| 1.0542 | 23.0 | 575 | 1.5322 | 0.6525 | 0.4557 | 3.3587 | 0.6525 | 0.6520 | 0.1428 | 0.1247 |
| 1.0542 | 24.0 | 600 | 1.5991 | 0.6475 | 0.4787 | 2.9349 | 0.6475 | 0.6444 | 0.1580 | 0.1450 |
| 1.0542 | 25.0 | 625 | 1.5625 | 0.6375 | 0.4926 | 3.0245 | 0.6375 | 0.6378 | 0.1641 | 0.1433 |
| 1.0542 | 26.0 | 650 | 1.5366 | 0.64 | 0.4884 | 3.3388 | 0.64 | 0.6461 | 0.1595 | 0.1453 |
| 1.0542 | 27.0 | 675 | 1.5686 | 0.65 | 0.4765 | 3.5120 | 0.65 | 0.6504 | 0.1625 | 0.1359 |
| 1.0542 | 28.0 | 700 | 1.5562 | 0.6475 | 0.4817 | 3.0348 | 0.6475 | 0.6488 | 0.1459 | 0.1388 |
| 1.0542 | 29.0 | 725 | 1.5213 | 0.6475 | 0.4719 | 3.2628 | 0.6475 | 0.6475 | 0.1634 | 0.1326 |
| 1.0542 | 30.0 | 750 | 1.5492 | 0.6675 | 0.4730 | 3.1693 | 0.6675 | 0.6679 | 0.1469 | 0.1415 |
| 1.0542 | 31.0 | 775 | 1.5311 | 0.65 | 0.4896 | 3.0881 | 0.65 | 0.6504 | 0.1815 | 0.1380 |
| 1.0542 | 32.0 | 800 | 1.5556 | 0.6475 | 0.4821 | 3.1829 | 0.6475 | 0.6491 | 0.1640 | 0.1405 |
| 1.0542 | 33.0 | 825 | 1.5471 | 0.6375 | 0.4846 | 3.4190 | 0.6375 | 0.6407 | 0.1628 | 0.1415 |
| 1.0542 | 34.0 | 850 | 1.4809 | 0.6575 | 0.4714 | 2.9136 | 0.6575 | 0.6612 | 0.1729 | 0.1338 |
| 1.0542 | 35.0 | 875 | 1.5256 | 0.66 | 0.4773 | 3.2303 | 0.66 | 0.6650 | 0.1746 | 0.1368 |
| 1.0542 | 36.0 | 900 | 1.4929 | 0.6675 | 0.4671 | 3.2360 | 0.6675 | 0.6698 | 0.1698 | 0.1309 |
| 1.0542 | 37.0 | 925 | 1.4923 | 0.645 | 0.4880 | 3.0567 | 0.645 | 0.6564 | 0.1764 | 0.1395 |
| 1.0542 | 38.0 | 950 | 1.5038 | 0.665 | 0.4672 | 3.2116 | 0.665 | 0.6661 | 0.1588 | 0.1343 |
| 1.0542 | 39.0 | 975 | 1.4708 | 0.6625 | 0.4669 | 3.1420 | 0.6625 | 0.6675 | 0.1683 | 0.1301 |
| 0.0522 | 40.0 | 1000 | 1.5153 | 0.6475 | 0.4865 | 3.1796 | 0.6475 | 0.6447 | 0.1639 | 0.1400 |
| 0.0522 | 41.0 | 1025 | 1.4705 | 0.6575 | 0.4642 | 3.2196 | 0.6575 | 0.6626 | 0.1440 | 0.1308 |
| 0.0522 | 42.0 | 1050 | 1.4844 | 0.6575 | 0.4722 | 3.2445 | 0.6575 | 0.6595 | 0.1746 | 0.1328 |
| 0.0522 | 43.0 | 1075 | 1.4957 | 0.6425 | 0.4828 | 3.1456 | 0.6425 | 0.6468 | 0.1499 | 0.1417 |
| 0.0522 | 44.0 | 1100 | 1.5179 | 0.645 | 0.4910 | 3.3921 | 0.645 | 0.6470 | 0.1861 | 0.1433 |
| 0.0522 | 45.0 | 1125 | 1.4878 | 0.6425 | 0.4839 | 3.2139 | 0.6425 | 0.6478 | 0.1720 | 0.1403 |
| 0.0522 | 46.0 | 1150 | 1.4666 | 0.655 | 0.4741 | 2.9333 | 0.655 | 0.6601 | 0.1813 | 0.1347 |
| 0.0522 | 47.0 | 1175 | 1.4954 | 0.6575 | 0.4776 | 3.2102 | 0.6575 | 0.6604 | 0.1842 | 0.1390 |
| 0.0522 | 48.0 | 1200 | 1.4976 | 0.645 | 0.4856 | 3.1539 | 0.645 | 0.6493 | 0.1549 | 0.1407 |
| 0.0522 | 49.0 | 1225 | 1.4772 | 0.64 | 0.4780 | 2.9845 | 0.64 | 0.6445 | 0.1826 | 0.1388 |
| 0.0522 | 50.0 | 1250 | 1.4584 | 0.65 | 0.4703 | 3.0776 | 0.65 | 0.6533 | 0.1685 | 0.1352 |
| 0.0522 | 51.0 | 1275 | 1.4828 | 0.6325 | 0.4844 | 3.1425 | 0.6325 | 0.6377 | 0.1641 | 0.1409 |
| 0.0522 | 52.0 | 1300 | 1.4676 | 0.6525 | 0.4737 | 3.1483 | 0.6525 | 0.6565 | 0.1773 | 0.1358 |
| 0.0522 | 53.0 | 1325 | 1.4675 | 0.6475 | 0.4791 | 3.1411 | 0.6475 | 0.6515 | 0.1820 | 0.1388 |
| 0.0522 | 54.0 | 1350 | 1.4724 | 0.645 | 0.4764 | 3.0744 | 0.645 | 0.6499 | 0.1847 | 0.1382 |
| 0.0522 | 55.0 | 1375 | 1.4689 | 0.6425 | 0.4769 | 3.2256 | 0.6425 | 0.6476 | 0.1839 | 0.1376 |
| 0.0522 | 56.0 | 1400 | 1.4660 | 0.6425 | 0.4760 | 2.9907 | 0.6425 | 0.6479 | 0.1906 | 0.1378 |
| 0.0522 | 57.0 | 1425 | 1.4663 | 0.645 | 0.4757 | 3.0722 | 0.645 | 0.6514 | 0.1705 | 0.1367 |
| 0.0522 | 58.0 | 1450 | 1.4678 | 0.65 | 0.4770 | 3.0710 | 0.65 | 0.6546 | 0.1794 | 0.1371 |
| 0.0522 | 59.0 | 1475 | 1.4717 | 0.64 | 0.4786 | 3.0737 | 0.64 | 0.6455 | 0.1889 | 0.1392 |
| 0.0064 | 60.0 | 1500 | 1.4691 | 0.645 | 0.4768 | 3.0688 | 0.645 | 0.6499 | 0.1815 | 0.1378 |
| 0.0064 | 61.0 | 1525 | 1.4689 | 0.64 | 0.4767 | 3.0688 | 0.64 | 0.6452 | 0.1846 | 0.1382 |
| 0.0064 | 62.0 | 1550 | 1.4689 | 0.64 | 0.4770 | 3.0674 | 0.64 | 0.6455 | 0.1937 | 0.1383 |
| 0.0064 | 63.0 | 1575 | 1.4687 | 0.6425 | 0.4767 | 3.0700 | 0.6425 | 0.6485 | 0.1897 | 0.1381 |
| 0.0064 | 64.0 | 1600 | 1.4674 | 0.6425 | 0.4764 | 3.0675 | 0.6425 | 0.6472 | 0.1855 | 0.1375 |
| 0.0064 | 65.0 | 1625 | 1.4681 | 0.6425 | 0.4766 | 3.0694 | 0.6425 | 0.6485 | 0.1917 | 0.1381 |
| 0.0064 | 66.0 | 1650 | 1.4681 | 0.6425 | 0.4766 | 3.0687 | 0.6425 | 0.6472 | 0.1905 | 0.1378 |
| 0.0064 | 67.0 | 1675 | 1.4667 | 0.645 | 0.4757 | 3.0681 | 0.645 | 0.6505 | 0.1899 | 0.1375 |
| 0.0064 | 68.0 | 1700 | 1.4683 | 0.6425 | 0.4771 | 3.0686 | 0.6425 | 0.6474 | 0.1871 | 0.1379 |
| 0.0064 | 69.0 | 1725 | 1.4672 | 0.64 | 0.4760 | 3.0679 | 0.64 | 0.6455 | 0.1932 | 0.1380 |
| 0.0064 | 70.0 | 1750 | 1.4673 | 0.6425 | 0.4763 | 3.0683 | 0.6425 | 0.6474 | 0.1955 | 0.1376 |
| 0.0064 | 71.0 | 1775 | 1.4676 | 0.645 | 0.4763 | 3.0680 | 0.645 | 0.6505 | 0.1921 | 0.1376 |
| 0.0064 | 72.0 | 1800 | 1.4674 | 0.6425 | 0.4763 | 3.0683 | 0.6425 | 0.6474 | 0.1946 | 0.1376 |
| 0.0064 | 73.0 | 1825 | 1.4675 | 0.6425 | 0.4763 | 3.0682 | 0.6425 | 0.6474 | 0.1946 | 0.1377 |
| 0.0064 | 74.0 | 1850 | 1.4674 | 0.6425 | 0.4763 | 3.0682 | 0.6425 | 0.6485 | 0.1945 | 0.1380 |
| 0.0064 | 75.0 | 1875 | 1.4674 | 0.64 | 0.4763 | 3.0680 | 0.64 | 0.6455 | 0.1960 | 0.1380 |
| 0.0064 | 76.0 | 1900 | 1.4675 | 0.64 | 0.4764 | 3.0682 | 0.64 | 0.6455 | 0.1972 | 0.1381 |
| 0.0064 | 77.0 | 1925 | 1.4675 | 0.6425 | 0.4763 | 3.0681 | 0.6425 | 0.6485 | 0.1947 | 0.1380 |
| 0.0064 | 78.0 | 1950 | 1.4674 | 0.6425 | 0.4763 | 3.0681 | 0.6425 | 0.6485 | 0.1958 | 0.1381 |
| 0.0064 | 79.0 | 1975 | 1.4674 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6474 | 0.1935 | 0.1376 |
| 0.0 | 80.0 | 2000 | 1.4673 | 0.6425 | 0.4763 | 3.0681 | 0.6425 | 0.6485 | 0.1958 | 0.1380 |
| 0.0 | 81.0 | 2025 | 1.4674 | 0.6425 | 0.4763 | 3.0681 | 0.6425 | 0.6485 | 0.1946 | 0.1380 |
| 0.0 | 82.0 | 2050 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1935 | 0.1380 |
| 0.0 | 83.0 | 2075 | 1.4674 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 84.0 | 2100 | 1.4674 | 0.6425 | 0.4763 | 3.0681 | 0.6425 | 0.6485 | 0.1958 | 0.1381 |
| 0.0 | 85.0 | 2125 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 86.0 | 2150 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 87.0 | 2175 | 1.4673 | 0.6425 | 0.4763 | 3.0681 | 0.6425 | 0.6485 | 0.1958 | 0.1381 |
| 0.0 | 88.0 | 2200 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 89.0 | 2225 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 90.0 | 2250 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 91.0 | 2275 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 92.0 | 2300 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 93.0 | 2325 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 94.0 | 2350 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1909 | 0.1381 |
| 0.0 | 95.0 | 2375 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 96.0 | 2400 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 97.0 | 2425 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 98.0 | 2450 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 99.0 | 2475 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
| 0.0 | 100.0 | 2500 | 1.4673 | 0.6425 | 0.4763 | 3.0680 | 0.6425 | 0.6485 | 0.1946 | 0.1381 |
Framework versions
- Transformers 4.28.0.dev0
- Pytorch 1.12.1+cu113
- Datasets 2.12.0
- Tokenizers 0.12.1
