<!-- 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-tiny_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: 0.0667
- Accuracy: 0.5865
- Brier Loss: 0.5908
- Nll: 3.0393
- F1 Micro: 0.5865
- F1 Macro: 0.5890
- Ece: 0.1479
- Aurc: 0.2054
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 |
---|---|---|---|---|---|---|---|---|---|---|
0.0807 | 1.0 | 1000 | 0.0798 | 0.095 | 0.9362 | 7.0778 | 0.095 | 0.0517 | 0.0524 | 0.8510 |
0.0785 | 2.0 | 2000 | 0.0782 | 0.142 | 0.9268 | 6.5000 | 0.142 | 0.0892 | 0.0843 | 0.7446 |
0.0768 | 3.0 | 3000 | 0.0761 | 0.253 | 0.8945 | 4.3268 | 0.253 | 0.1827 | 0.1545 | 0.5697 |
0.0753 | 4.0 | 4000 | 0.0747 | 0.327 | 0.8672 | 3.7313 | 0.327 | 0.2733 | 0.2052 | 0.4558 |
0.074 | 5.0 | 5000 | 0.0739 | 0.359 | 0.8410 | 3.6965 | 0.359 | 0.2941 | 0.2102 | 0.4159 |
0.0729 | 6.0 | 6000 | 0.0725 | 0.3795 | 0.8104 | 3.2323 | 0.3795 | 0.3340 | 0.2147 | 0.3672 |
0.0718 | 7.0 | 7000 | 0.0717 | 0.4165 | 0.7806 | 3.1185 | 0.4165 | 0.3770 | 0.2186 | 0.3378 |
0.071 | 8.0 | 8000 | 0.0714 | 0.4175 | 0.7785 | 3.1984 | 0.4175 | 0.3999 | 0.2170 | 0.3408 |
0.0703 | 9.0 | 9000 | 0.0707 | 0.457 | 0.7563 | 2.8932 | 0.457 | 0.4310 | 0.2437 | 0.2965 |
0.0696 | 10.0 | 10000 | 0.0699 | 0.4665 | 0.7452 | 2.7889 | 0.4665 | 0.4529 | 0.2456 | 0.2828 |
0.0691 | 11.0 | 11000 | 0.0693 | 0.499 | 0.7219 | 2.7292 | 0.499 | 0.4756 | 0.2543 | 0.2579 |
0.0685 | 12.0 | 12000 | 0.0691 | 0.4955 | 0.7144 | 2.8807 | 0.4955 | 0.4734 | 0.2443 | 0.2515 |
0.068 | 13.0 | 13000 | 0.0688 | 0.5072 | 0.7096 | 2.6737 | 0.5072 | 0.4944 | 0.2525 | 0.2468 |
0.0675 | 14.0 | 14000 | 0.0685 | 0.513 | 0.6952 | 2.7492 | 0.513 | 0.5001 | 0.2404 | 0.2453 |
0.0669 | 15.0 | 15000 | 0.0682 | 0.5232 | 0.6855 | 2.7789 | 0.5232 | 0.5048 | 0.2441 | 0.2379 |
0.0664 | 16.0 | 16000 | 0.0680 | 0.529 | 0.6790 | 2.8249 | 0.529 | 0.5182 | 0.2366 | 0.2340 |
0.0658 | 17.0 | 17000 | 0.0678 | 0.5347 | 0.6668 | 2.7035 | 0.5347 | 0.5237 | 0.2338 | 0.2228 |
0.0652 | 18.0 | 18000 | 0.0676 | 0.5335 | 0.6673 | 2.8630 | 0.5335 | 0.5249 | 0.2319 | 0.2252 |
0.0651 | 19.0 | 19000 | 0.0675 | 0.5385 | 0.6524 | 2.7522 | 0.5385 | 0.5286 | 0.2172 | 0.2256 |
0.0645 | 20.0 | 20000 | 0.0671 | 0.5593 | 0.6454 | 2.7445 | 0.5593 | 0.5563 | 0.2324 | 0.2122 |
0.0639 | 21.0 | 21000 | 0.0672 | 0.5453 | 0.6541 | 2.9011 | 0.5453 | 0.5451 | 0.2236 | 0.2204 |
0.0634 | 22.0 | 22000 | 0.0668 | 0.5617 | 0.6398 | 2.8668 | 0.5617 | 0.5604 | 0.2264 | 0.2108 |
0.0629 | 23.0 | 23000 | 0.0670 | 0.5577 | 0.6295 | 2.8351 | 0.5577 | 0.5521 | 0.1984 | 0.2180 |
0.0625 | 24.0 | 24000 | 0.0666 | 0.5765 | 0.6201 | 2.7133 | 0.5765 | 0.5754 | 0.2138 | 0.2035 |
0.0618 | 25.0 | 25000 | 0.0666 | 0.565 | 0.6219 | 2.8775 | 0.565 | 0.5614 | 0.2010 | 0.2078 |
0.0613 | 26.0 | 26000 | 0.0664 | 0.5795 | 0.6121 | 2.8665 | 0.5795 | 0.5805 | 0.1996 | 0.2024 |
0.0606 | 27.0 | 27000 | 0.0667 | 0.5723 | 0.6101 | 2.9450 | 0.5723 | 0.5711 | 0.1804 | 0.2113 |
0.0603 | 28.0 | 28000 | 0.0664 | 0.583 | 0.6106 | 2.9126 | 0.583 | 0.5845 | 0.2004 | 0.2006 |
0.0597 | 29.0 | 29000 | 0.0665 | 0.5857 | 0.6050 | 2.9881 | 0.5857 | 0.5862 | 0.1912 | 0.2006 |
0.0594 | 30.0 | 30000 | 0.0665 | 0.5775 | 0.6043 | 2.9735 | 0.5775 | 0.5797 | 0.1823 | 0.2029 |
0.0589 | 31.0 | 31000 | 0.0666 | 0.5733 | 0.6080 | 2.9942 | 0.5733 | 0.5739 | 0.1721 | 0.2129 |
0.0585 | 32.0 | 32000 | 0.0667 | 0.5803 | 0.6066 | 3.0341 | 0.5803 | 0.5826 | 0.1748 | 0.2114 |
0.0583 | 33.0 | 33000 | 0.0665 | 0.5827 | 0.6033 | 3.0209 | 0.5827 | 0.5880 | 0.1799 | 0.2029 |
0.0578 | 34.0 | 34000 | 0.0667 | 0.577 | 0.6020 | 3.0483 | 0.577 | 0.5816 | 0.1636 | 0.2081 |
0.0576 | 35.0 | 35000 | 0.0667 | 0.577 | 0.6029 | 3.0263 | 0.577 | 0.5840 | 0.1573 | 0.2117 |
0.0574 | 36.0 | 36000 | 0.0667 | 0.5803 | 0.6006 | 3.0578 | 0.5803 | 0.5851 | 0.1627 | 0.2082 |
0.057 | 37.0 | 37000 | 0.0666 | 0.582 | 0.5997 | 3.1133 | 0.582 | 0.5867 | 0.1612 | 0.2094 |
0.0567 | 38.0 | 38000 | 0.0667 | 0.5817 | 0.5951 | 3.0727 | 0.5817 | 0.5836 | 0.1552 | 0.2091 |
0.0566 | 39.0 | 39000 | 0.0666 | 0.5815 | 0.5951 | 3.0308 | 0.5815 | 0.5853 | 0.1559 | 0.2049 |
0.0564 | 40.0 | 40000 | 0.0666 | 0.5853 | 0.5940 | 3.0629 | 0.5853 | 0.5880 | 0.1564 | 0.2057 |
0.0562 | 41.0 | 41000 | 0.0666 | 0.5845 | 0.5949 | 3.0956 | 0.5845 | 0.5881 | 0.1585 | 0.2055 |
0.0561 | 42.0 | 42000 | 0.0666 | 0.5827 | 0.5960 | 3.0679 | 0.5827 | 0.5876 | 0.1540 | 0.2098 |
0.0559 | 43.0 | 43000 | 0.0666 | 0.5833 | 0.5909 | 2.9904 | 0.5833 | 0.5854 | 0.1491 | 0.2049 |
0.0559 | 44.0 | 44000 | 0.0665 | 0.585 | 0.5915 | 3.0150 | 0.585 | 0.5876 | 0.1543 | 0.2032 |
0.0557 | 45.0 | 45000 | 0.0667 | 0.583 | 0.5923 | 3.0501 | 0.583 | 0.5851 | 0.1501 | 0.2056 |
0.0557 | 46.0 | 46000 | 0.0666 | 0.5905 | 0.5914 | 3.0110 | 0.5905 | 0.5940 | 0.1550 | 0.2045 |
0.0555 | 47.0 | 47000 | 0.0667 | 0.584 | 0.5922 | 3.0464 | 0.584 | 0.5872 | 0.1497 | 0.2069 |
0.0555 | 48.0 | 48000 | 0.0667 | 0.588 | 0.5917 | 3.0408 | 0.588 | 0.5919 | 0.1489 | 0.2051 |
0.0554 | 49.0 | 49000 | 0.0667 | 0.589 | 0.5908 | 3.0433 | 0.589 | 0.5923 | 0.1496 | 0.2044 |
0.0554 | 50.0 | 50000 | 0.0667 | 0.5865 | 0.5908 | 3.0393 | 0.5865 | 0.5890 | 0.1479 | 0.2054 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2