dit-base-finetuned-rvlcdip-small_rvl_cdip-NK1000_og_simkd
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: 12194.7598
- Accuracy: 0.8558
- Brier Loss: 0.2688
- Nll: 1.9967
- F1 Micro: 0.8558
- F1 Macro: 0.8552
- Ece: 0.1305
- Aurc: 0.0490
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: 16
- eval_batch_size: 16
- 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: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|---|---|---|---|---|---|---|---|---|---|---|
| 12767.556 | 1.0 | 1000 | 12472.2930 | 0.5725 | 0.5598 | 2.8765 | 0.5725 | 0.5254 | 0.0980 | 0.1794 |
| 12750.588 | 2.0 | 2000 | 12455.9170 | 0.6683 | 0.4822 | 2.7098 | 0.6683 | 0.6534 | 0.1247 | 0.1325 |
| 12820.858 | 3.0 | 3000 | 12458.1924 | 0.7003 | 0.4459 | 2.6913 | 0.7003 | 0.7019 | 0.1303 | 0.0964 |
| 12762.296 | 4.0 | 4000 | 12445.5703 | 0.7167 | 0.4202 | 2.8379 | 0.7168 | 0.7217 | 0.1078 | 0.0925 |
| 12706.14 | 5.0 | 5000 | 12425.8330 | 0.753 | 0.3820 | 2.7925 | 0.753 | 0.7539 | 0.0997 | 0.0923 |
| 12764.822 | 6.0 | 6000 | 12427.2080 | 0.7635 | 0.3603 | 2.5903 | 0.7635 | 0.7659 | 0.0950 | 0.0823 |
| 12719.869 | 7.0 | 7000 | 12411.4668 | 0.769 | 0.3469 | 2.7566 | 0.769 | 0.7761 | 0.0895 | 0.0681 |
| 12628.481 | 8.0 | 8000 | 12412.3760 | 0.7738 | 0.3535 | 2.7667 | 0.7738 | 0.7832 | 0.1127 | 0.0699 |
| 12624.542 | 9.0 | 9000 | 12396.7773 | 0.7933 | 0.3243 | 2.4484 | 0.7932 | 0.7954 | 0.1002 | 0.0664 |
| 12681.642 | 10.0 | 10000 | 12391.2744 | 0.7943 | 0.3241 | 2.5709 | 0.7943 | 0.7979 | 0.1081 | 0.0592 |
| 12656.593 | 11.0 | 11000 | 12383.5020 | 0.8015 | 0.3190 | 2.4516 | 0.8015 | 0.8065 | 0.1064 | 0.0597 |
| 12638.155 | 12.0 | 12000 | 12372.9707 | 0.7957 | 0.3357 | 2.4891 | 0.7957 | 0.7956 | 0.1225 | 0.0679 |
| 12698.474 | 13.0 | 13000 | 12370.7217 | 0.813 | 0.2988 | 2.1414 | 0.813 | 0.8125 | 0.1030 | 0.0494 |
| 12574.549 | 14.0 | 14000 | 12361.6641 | 0.8045 | 0.3218 | 2.4610 | 0.8045 | 0.8043 | 0.1155 | 0.0560 |
| 12589.537 | 15.0 | 15000 | 12345.1123 | 0.8193 | 0.3046 | 2.2566 | 0.8193 | 0.8184 | 0.1220 | 0.0524 |
| 12592.604 | 16.0 | 16000 | 12354.9756 | 0.817 | 0.3078 | 2.3526 | 0.817 | 0.8207 | 0.1204 | 0.0527 |
| 12660.709 | 17.0 | 17000 | 12334.7686 | 0.8293 | 0.2942 | 2.2857 | 0.8293 | 0.8284 | 0.1201 | 0.0482 |
| 12591.369 | 18.0 | 18000 | 12334.4570 | 0.829 | 0.2948 | 2.1559 | 0.8290 | 0.8287 | 0.1211 | 0.0451 |
| 12598.469 | 19.0 | 19000 | 12320.7510 | 0.826 | 0.2997 | 2.2348 | 0.826 | 0.8251 | 0.1240 | 0.0473 |
| 12497.537 | 20.0 | 20000 | 12307.0811 | 0.8347 | 0.2833 | 2.2433 | 0.8347 | 0.8358 | 0.1200 | 0.0426 |
| 12537.66 | 21.0 | 21000 | 12310.8438 | 0.8323 | 0.2965 | 2.1513 | 0.8323 | 0.8321 | 0.1287 | 0.0490 |
| 12524.668 | 22.0 | 22000 | 12300.1055 | 0.8403 | 0.2776 | 2.1780 | 0.8403 | 0.8407 | 0.1207 | 0.0427 |
| 12433.952 | 23.0 | 23000 | 12288.1221 | 0.8353 | 0.2898 | 2.2189 | 0.8353 | 0.8357 | 0.1346 | 0.0439 |
| 12598.38 | 24.0 | 24000 | 12282.6680 | 0.8442 | 0.2765 | 2.1653 | 0.8443 | 0.8440 | 0.1264 | 0.0438 |
| 12474.447 | 25.0 | 25000 | 12277.6797 | 0.8363 | 0.2925 | 2.1209 | 0.8363 | 0.8350 | 0.1366 | 0.0451 |
| 12522.706 | 26.0 | 26000 | 12276.4502 | 0.8465 | 0.2764 | 2.0779 | 0.8465 | 0.8469 | 0.1291 | 0.0432 |
| 12502.289 | 27.0 | 27000 | 12268.1758 | 0.8445 | 0.2811 | 2.0839 | 0.8445 | 0.8442 | 0.1318 | 0.0465 |
| 12465.994 | 28.0 | 28000 | 12252.7266 | 0.8433 | 0.2882 | 2.1410 | 0.8433 | 0.8431 | 0.1380 | 0.0479 |
| 12467.13 | 29.0 | 29000 | 12260.4912 | 0.8442 | 0.2838 | 2.1129 | 0.8443 | 0.8430 | 0.1348 | 0.0487 |
| 12540.006 | 30.0 | 30000 | 12249.1670 | 0.846 | 0.2811 | 2.1134 | 0.8460 | 0.8458 | 0.1349 | 0.0486 |
| 12594.326 | 31.0 | 31000 | 12245.6699 | 0.8452 | 0.2850 | 2.0734 | 0.8452 | 0.8443 | 0.1363 | 0.0480 |
| 12486.203 | 32.0 | 32000 | 12240.5479 | 0.8468 | 0.2813 | 2.0757 | 0.8468 | 0.8463 | 0.1353 | 0.0484 |
| 12468.631 | 33.0 | 33000 | 12231.9600 | 0.852 | 0.2715 | 2.0178 | 0.852 | 0.8523 | 0.1309 | 0.0450 |
| 12423.715 | 34.0 | 34000 | 12215.6680 | 0.8472 | 0.2843 | 2.0927 | 0.8472 | 0.8470 | 0.1389 | 0.0491 |
| 12454.715 | 35.0 | 35000 | 12223.0361 | 0.8492 | 0.2772 | 2.0161 | 0.8492 | 0.8485 | 0.1340 | 0.0476 |
| 12466.932 | 36.0 | 36000 | 12221.3887 | 0.8495 | 0.2776 | 2.0135 | 0.8495 | 0.8488 | 0.1343 | 0.0467 |
| 12483.745 | 37.0 | 37000 | 12210.9414 | 0.8508 | 0.2748 | 2.0374 | 0.8508 | 0.8506 | 0.1350 | 0.0493 |
| 12453.102 | 38.0 | 38000 | 12224.9482 | 0.852 | 0.2737 | 1.9699 | 0.852 | 0.8517 | 0.1308 | 0.0460 |
| 12511.225 | 39.0 | 39000 | 12213.9756 | 0.8522 | 0.2763 | 1.9619 | 0.8522 | 0.8518 | 0.1342 | 0.0484 |
| 12561.782 | 40.0 | 40000 | 12213.9297 | 0.852 | 0.2736 | 2.0481 | 0.852 | 0.8516 | 0.1326 | 0.0477 |
| 12524.982 | 41.0 | 41000 | 12208.1758 | 0.8518 | 0.2745 | 1.9751 | 0.8518 | 0.8509 | 0.1346 | 0.0490 |
| 12465.351 | 42.0 | 42000 | 12215.3604 | 0.8532 | 0.2730 | 2.0037 | 0.8532 | 0.8521 | 0.1314 | 0.0474 |
| 12419.902 | 43.0 | 43000 | 12211.3701 | 0.8565 | 0.2680 | 2.0140 | 0.8565 | 0.8561 | 0.1297 | 0.0462 |
| 12493.264 | 44.0 | 44000 | 12196.7217 | 0.8532 | 0.2717 | 1.9866 | 0.8532 | 0.8524 | 0.1336 | 0.0487 |
| 12487.514 | 45.0 | 45000 | 12199.4902 | 0.8532 | 0.2700 | 1.9711 | 0.8532 | 0.8523 | 0.1309 | 0.0478 |
| 12321.575 | 46.0 | 46000 | 12189.0117 | 0.8535 | 0.2727 | 2.0220 | 0.8535 | 0.8528 | 0.1335 | 0.0492 |
| 12423.494 | 47.0 | 47000 | 12198.2002 | 0.8542 | 0.2711 | 1.9648 | 0.8542 | 0.8536 | 0.1331 | 0.0478 |
| 12535.605 | 48.0 | 48000 | 12192.7061 | 0.8565 | 0.2678 | 2.0098 | 0.8565 | 0.8560 | 0.1292 | 0.0489 |
| 12319.588 | 49.0 | 49000 | 12185.3916 | 0.856 | 0.2691 | 2.0285 | 0.856 | 0.8554 | 0.1311 | 0.0506 |
| 12470.527 | 50.0 | 50000 | 12194.7598 | 0.8558 | 0.2688 | 1.9967 | 0.8558 | 0.8552 | 0.1305 | 0.0490 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2