<!-- 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-base_rvl-cdip-small_rvl_cdip-NK1000_simkd_rand
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0648
- Accuracy: 0.6072
- Brier Loss: 0.5503
- Nll: 2.7228
- F1 Micro: 0.6072
- F1 Macro: 0.6102
- Ece: 0.1175
- Aurc: 0.1871
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 |
---|---|---|---|---|---|---|---|---|---|---|
0.0792 | 1.0 | 1000 | 0.0787 | 0.1103 | 0.9330 | 7.5575 | 0.1103 | 0.0531 | 0.0745 | 0.7817 |
0.0778 | 2.0 | 2000 | 0.0772 | 0.1867 | 0.9126 | 5.0754 | 0.1867 | 0.1367 | 0.1099 | 0.6583 |
0.0753 | 3.0 | 3000 | 0.0748 | 0.3185 | 0.8694 | 3.7575 | 0.3185 | 0.2711 | 0.2000 | 0.4734 |
0.0738 | 4.0 | 4000 | 0.0733 | 0.3633 | 0.8431 | 3.5732 | 0.3633 | 0.3239 | 0.2222 | 0.4034 |
0.0724 | 5.0 | 5000 | 0.0721 | 0.4083 | 0.8051 | 3.1556 | 0.4083 | 0.3747 | 0.2309 | 0.3543 |
0.0712 | 6.0 | 6000 | 0.0716 | 0.422 | 0.7734 | 3.1574 | 0.422 | 0.3887 | 0.1999 | 0.3595 |
0.07 | 7.0 | 7000 | 0.0703 | 0.4718 | 0.7584 | 2.9015 | 0.4718 | 0.4495 | 0.2548 | 0.2992 |
0.0693 | 8.0 | 8000 | 0.0696 | 0.493 | 0.7247 | 3.0406 | 0.493 | 0.4634 | 0.2430 | 0.2667 |
0.0683 | 9.0 | 9000 | 0.0693 | 0.4955 | 0.7122 | 3.1754 | 0.4955 | 0.4725 | 0.2329 | 0.2620 |
0.0675 | 10.0 | 10000 | 0.0682 | 0.528 | 0.6893 | 2.7012 | 0.528 | 0.5188 | 0.2477 | 0.2418 |
0.0668 | 11.0 | 11000 | 0.0680 | 0.5035 | 0.6961 | 2.9081 | 0.5035 | 0.4967 | 0.2243 | 0.2569 |
0.066 | 12.0 | 12000 | 0.0669 | 0.5595 | 0.6626 | 2.6583 | 0.5595 | 0.5541 | 0.2550 | 0.2169 |
0.0649 | 13.0 | 13000 | 0.0668 | 0.5493 | 0.6559 | 2.7433 | 0.5493 | 0.5525 | 0.2334 | 0.2165 |
0.0645 | 14.0 | 14000 | 0.0663 | 0.5645 | 0.6334 | 2.6699 | 0.5645 | 0.5605 | 0.2191 | 0.2099 |
0.0636 | 15.0 | 15000 | 0.0659 | 0.5765 | 0.6160 | 2.6406 | 0.5765 | 0.5664 | 0.2084 | 0.2012 |
0.0629 | 16.0 | 16000 | 0.0661 | 0.56 | 0.6311 | 2.7536 | 0.56 | 0.5686 | 0.2129 | 0.2026 |
0.0622 | 17.0 | 17000 | 0.0660 | 0.5725 | 0.6108 | 2.8055 | 0.5725 | 0.5707 | 0.1866 | 0.2046 |
0.0617 | 18.0 | 18000 | 0.0656 | 0.5697 | 0.6081 | 2.8309 | 0.5697 | 0.5730 | 0.1810 | 0.2024 |
0.0608 | 19.0 | 19000 | 0.0654 | 0.585 | 0.5982 | 2.6432 | 0.585 | 0.5867 | 0.1834 | 0.1975 |
0.06 | 20.0 | 20000 | 0.0656 | 0.584 | 0.5959 | 2.8363 | 0.584 | 0.5856 | 0.1662 | 0.2067 |
0.0592 | 21.0 | 21000 | 0.0657 | 0.5875 | 0.5896 | 2.8259 | 0.5875 | 0.5892 | 0.1575 | 0.2059 |
0.0584 | 22.0 | 22000 | 0.0655 | 0.5887 | 0.5832 | 2.8147 | 0.5887 | 0.5895 | 0.1531 | 0.1998 |
0.058 | 23.0 | 23000 | 0.0654 | 0.5945 | 0.5829 | 2.9399 | 0.5945 | 0.5955 | 0.1475 | 0.2007 |
0.0571 | 24.0 | 24000 | 0.0654 | 0.5962 | 0.5779 | 2.8266 | 0.5962 | 0.5982 | 0.1460 | 0.1996 |
0.0566 | 25.0 | 25000 | 0.0655 | 0.596 | 0.5815 | 2.9480 | 0.596 | 0.5975 | 0.1447 | 0.2099 |
0.0561 | 26.0 | 26000 | 0.0660 | 0.5883 | 0.5840 | 2.9985 | 0.5883 | 0.5903 | 0.1146 | 0.2202 |
0.0556 | 27.0 | 27000 | 0.0654 | 0.6042 | 0.5713 | 2.8775 | 0.6042 | 0.6052 | 0.1353 | 0.2020 |
0.055 | 28.0 | 28000 | 0.0655 | 0.5945 | 0.5750 | 3.0404 | 0.5945 | 0.5965 | 0.1215 | 0.2051 |
0.0546 | 29.0 | 29000 | 0.0655 | 0.5978 | 0.5740 | 2.9173 | 0.5978 | 0.6012 | 0.1226 | 0.2046 |
0.0543 | 30.0 | 30000 | 0.0657 | 0.588 | 0.5813 | 3.0493 | 0.588 | 0.5915 | 0.1210 | 0.2104 |
0.054 | 31.0 | 31000 | 0.0652 | 0.597 | 0.5715 | 2.9423 | 0.597 | 0.5989 | 0.1207 | 0.2055 |
0.0537 | 32.0 | 32000 | 0.0650 | 0.6075 | 0.5618 | 2.8731 | 0.6075 | 0.6080 | 0.1209 | 0.1987 |
0.0534 | 33.0 | 33000 | 0.0650 | 0.602 | 0.5651 | 2.7807 | 0.602 | 0.6046 | 0.1254 | 0.1988 |
0.0535 | 34.0 | 34000 | 0.0652 | 0.602 | 0.5661 | 3.0050 | 0.602 | 0.6068 | 0.1187 | 0.1977 |
0.053 | 35.0 | 35000 | 0.0649 | 0.6008 | 0.5603 | 2.8814 | 0.6008 | 0.6028 | 0.1172 | 0.1981 |
0.0527 | 36.0 | 36000 | 0.0649 | 0.5988 | 0.5575 | 2.8419 | 0.5988 | 0.5974 | 0.1156 | 0.1917 |
0.0526 | 37.0 | 37000 | 0.0649 | 0.598 | 0.5586 | 2.7982 | 0.598 | 0.5986 | 0.1173 | 0.1900 |
0.0524 | 38.0 | 38000 | 0.0646 | 0.604 | 0.5546 | 2.8202 | 0.604 | 0.6060 | 0.1244 | 0.1908 |
0.0524 | 39.0 | 39000 | 0.0651 | 0.5965 | 0.5627 | 2.8458 | 0.5965 | 0.6010 | 0.1125 | 0.1949 |
0.0522 | 40.0 | 40000 | 0.0649 | 0.6072 | 0.5515 | 2.7872 | 0.6072 | 0.6100 | 0.1211 | 0.1881 |
0.0521 | 41.0 | 41000 | 0.0648 | 0.6078 | 0.5542 | 2.7802 | 0.6078 | 0.6108 | 0.1199 | 0.1868 |
0.052 | 42.0 | 42000 | 0.0648 | 0.6 | 0.5557 | 2.7968 | 0.6 | 0.6029 | 0.1190 | 0.1940 |
0.0519 | 43.0 | 43000 | 0.0647 | 0.604 | 0.5503 | 2.7110 | 0.604 | 0.6060 | 0.1178 | 0.1896 |
0.0516 | 44.0 | 44000 | 0.0647 | 0.6065 | 0.5515 | 2.7595 | 0.6065 | 0.6089 | 0.1170 | 0.1870 |
0.0516 | 45.0 | 45000 | 0.0646 | 0.611 | 0.5496 | 2.7426 | 0.611 | 0.6129 | 0.1212 | 0.1873 |
0.0515 | 46.0 | 46000 | 0.0648 | 0.6082 | 0.5510 | 2.7436 | 0.6082 | 0.6120 | 0.1227 | 0.1876 |
0.0514 | 47.0 | 47000 | 0.0647 | 0.6088 | 0.5511 | 2.7379 | 0.6088 | 0.6115 | 0.1240 | 0.1874 |
0.0514 | 48.0 | 48000 | 0.0647 | 0.6095 | 0.5501 | 2.7369 | 0.6095 | 0.6122 | 0.1193 | 0.1868 |
0.0513 | 49.0 | 49000 | 0.0647 | 0.6095 | 0.5508 | 2.7295 | 0.6095 | 0.6122 | 0.1218 | 0.1870 |
0.0513 | 50.0 | 50000 | 0.0648 | 0.6072 | 0.5503 | 2.7228 | 0.6072 | 0.6102 | 0.1175 | 0.1871 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2