<!-- 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. -->
dit-base-finetuned-rvlcdip-tiny_rvl_cdip-NK1000_og_simkd
This model is a fine-tuned version of WinKawaks/vit-tiny-patch16-224 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 12216.4121
- Accuracy: 0.8337
- Brier Loss: 0.3073
- Nll: 2.1945
- F1 Micro: 0.8337
- F1 Macro: 0.8337
- Ece: 0.1506
- Aurc: 0.0535
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 |
---|---|---|---|---|---|---|---|---|---|---|
12731.392 | 1.0 | 1000 | 12479.0312 | 0.5323 | 0.5967 | 3.1288 | 0.5323 | 0.4700 | 0.1047 | 0.2100 |
12715.18 | 2.0 | 2000 | 12453.9434 | 0.5787 | 0.6261 | 3.4535 | 0.5787 | 0.5518 | 0.1953 | 0.2225 |
12672.101 | 3.0 | 3000 | 12456.4629 | 0.6723 | 0.4708 | 2.9701 | 0.6723 | 0.6487 | 0.1102 | 0.1171 |
12681.216 | 4.0 | 4000 | 12448.6143 | 0.6815 | 0.4741 | 2.8370 | 0.6815 | 0.6707 | 0.1250 | 0.1245 |
12638.181 | 5.0 | 5000 | 12443.0645 | 0.716 | 0.4178 | 2.7837 | 0.7160 | 0.7273 | 0.1095 | 0.0932 |
12802.019 | 6.0 | 6000 | 12432.3438 | 0.7468 | 0.3768 | 2.6575 | 0.7468 | 0.7537 | 0.0898 | 0.0800 |
12671.194 | 7.0 | 7000 | 12420.4395 | 0.7335 | 0.4163 | 2.8018 | 0.7335 | 0.7331 | 0.1256 | 0.1139 |
12705.783 | 8.0 | 8000 | 12410.6143 | 0.7598 | 0.3739 | 2.7562 | 0.7598 | 0.7634 | 0.1157 | 0.0729 |
12559.44 | 9.0 | 9000 | 12412.9453 | 0.7612 | 0.3793 | 2.8905 | 0.7612 | 0.7658 | 0.1232 | 0.0731 |
12618.725 | 10.0 | 10000 | 12390.2139 | 0.7722 | 0.3613 | 2.6477 | 0.7722 | 0.7753 | 0.1190 | 0.0750 |
12731.292 | 11.0 | 11000 | 12387.4863 | 0.7875 | 0.3427 | 2.5376 | 0.7875 | 0.7898 | 0.1154 | 0.0689 |
12705.794 | 12.0 | 12000 | 12379.9336 | 0.7805 | 0.3555 | 2.6072 | 0.7805 | 0.7813 | 0.1284 | 0.0681 |
12550.782 | 13.0 | 13000 | 12380.7959 | 0.787 | 0.3400 | 2.5381 | 0.787 | 0.7887 | 0.1162 | 0.0662 |
12670.568 | 14.0 | 14000 | 12376.7646 | 0.7867 | 0.3423 | 2.6149 | 0.7868 | 0.7925 | 0.1186 | 0.0574 |
12580.616 | 15.0 | 15000 | 12352.3135 | 0.7953 | 0.3468 | 2.6324 | 0.7953 | 0.7969 | 0.1382 | 0.0622 |
12723.865 | 16.0 | 16000 | 12345.4600 | 0.8015 | 0.3312 | 2.4793 | 0.8015 | 0.8034 | 0.1244 | 0.0601 |
12620.305 | 17.0 | 17000 | 12343.1553 | 0.8023 | 0.3424 | 2.6488 | 0.8023 | 0.8031 | 0.1420 | 0.0644 |
12668.087 | 18.0 | 18000 | 12336.9277 | 0.8 | 0.3455 | 2.7019 | 0.8000 | 0.8023 | 0.1401 | 0.0592 |
12654.687 | 19.0 | 19000 | 12332.4404 | 0.8075 | 0.3321 | 2.5589 | 0.8075 | 0.8094 | 0.1393 | 0.0556 |
12578.655 | 20.0 | 20000 | 12321.3037 | 0.8075 | 0.3395 | 2.4255 | 0.8075 | 0.8050 | 0.1484 | 0.0638 |
12525.448 | 21.0 | 21000 | 12315.5303 | 0.8067 | 0.3328 | 2.5264 | 0.8067 | 0.8066 | 0.1440 | 0.0548 |
12610.837 | 22.0 | 22000 | 12311.0215 | 0.8105 | 0.3291 | 2.4781 | 0.8105 | 0.8112 | 0.1445 | 0.0540 |
12494.528 | 23.0 | 23000 | 12303.3623 | 0.8145 | 0.3337 | 2.5535 | 0.8145 | 0.8154 | 0.1510 | 0.0561 |
12561.799 | 24.0 | 24000 | 12296.2363 | 0.8153 | 0.3246 | 2.4243 | 0.8153 | 0.8142 | 0.1475 | 0.0513 |
12580.176 | 25.0 | 25000 | 12291.8018 | 0.8193 | 0.3262 | 2.3932 | 0.8193 | 0.8174 | 0.1484 | 0.0550 |
12455.165 | 26.0 | 26000 | 12276.9355 | 0.826 | 0.3223 | 2.4710 | 0.826 | 0.8251 | 0.1507 | 0.0597 |
12528.496 | 27.0 | 27000 | 12280.9180 | 0.8257 | 0.3154 | 2.4010 | 0.8257 | 0.8260 | 0.1462 | 0.0524 |
12521.554 | 28.0 | 28000 | 12262.9600 | 0.821 | 0.3274 | 2.4721 | 0.821 | 0.8201 | 0.1560 | 0.0595 |
12557.871 | 29.0 | 29000 | 12260.7754 | 0.823 | 0.3217 | 2.3929 | 0.823 | 0.8226 | 0.1552 | 0.0551 |
12535.524 | 30.0 | 30000 | 12271.4717 | 0.8263 | 0.3183 | 2.3249 | 0.8263 | 0.8269 | 0.1503 | 0.0502 |
12488.263 | 31.0 | 31000 | 12259.3057 | 0.823 | 0.3219 | 2.3830 | 0.823 | 0.8226 | 0.1541 | 0.0528 |
12498.048 | 32.0 | 32000 | 12253.2412 | 0.8263 | 0.3174 | 2.2771 | 0.8263 | 0.8243 | 0.1527 | 0.0541 |
12465.825 | 33.0 | 33000 | 12257.4863 | 0.8323 | 0.3088 | 2.3466 | 0.8323 | 0.8319 | 0.1454 | 0.0500 |
12439.6 | 34.0 | 34000 | 12238.5957 | 0.8323 | 0.3093 | 2.4057 | 0.8323 | 0.8329 | 0.1482 | 0.0552 |
12407.423 | 35.0 | 35000 | 12250.7178 | 0.8335 | 0.3072 | 2.2532 | 0.8335 | 0.8336 | 0.1471 | 0.0521 |
12534.711 | 36.0 | 36000 | 12231.9902 | 0.8353 | 0.3032 | 2.2711 | 0.8353 | 0.8353 | 0.1464 | 0.0548 |
12458.666 | 37.0 | 37000 | 12232.9521 | 0.835 | 0.3041 | 2.2523 | 0.835 | 0.8352 | 0.1467 | 0.0539 |
12461.748 | 38.0 | 38000 | 12230.4639 | 0.8317 | 0.3096 | 2.3052 | 0.8317 | 0.8318 | 0.1512 | 0.0539 |
12434.679 | 39.0 | 39000 | 12229.0684 | 0.8317 | 0.3081 | 2.2172 | 0.8317 | 0.8317 | 0.1497 | 0.0547 |
12468.468 | 40.0 | 40000 | 12226.4775 | 0.8323 | 0.3096 | 2.3112 | 0.8323 | 0.8324 | 0.1509 | 0.0524 |
12540.176 | 41.0 | 41000 | 12213.8359 | 0.8357 | 0.3085 | 2.2929 | 0.8357 | 0.8356 | 0.1502 | 0.0541 |
12513.896 | 42.0 | 42000 | 12216.2480 | 0.8333 | 0.3096 | 2.1638 | 0.8333 | 0.8329 | 0.1501 | 0.0559 |
12406.31 | 43.0 | 43000 | 12213.7012 | 0.8347 | 0.3078 | 2.1971 | 0.8347 | 0.8345 | 0.1504 | 0.0542 |
12350.768 | 44.0 | 44000 | 12224.6738 | 0.8323 | 0.3086 | 2.1722 | 0.8323 | 0.8320 | 0.1514 | 0.0546 |
12394.478 | 45.0 | 45000 | 12221.9336 | 0.8325 | 0.3100 | 2.2464 | 0.8325 | 0.8323 | 0.1516 | 0.0536 |
12399.318 | 46.0 | 46000 | 12207.5957 | 0.8347 | 0.3089 | 2.2193 | 0.8347 | 0.8344 | 0.1517 | 0.0553 |
12476.218 | 47.0 | 47000 | 12213.4814 | 0.8353 | 0.3055 | 2.2084 | 0.8353 | 0.8353 | 0.1488 | 0.0532 |
12448.278 | 48.0 | 48000 | 12212.2119 | 0.8347 | 0.3058 | 2.1518 | 0.8347 | 0.8345 | 0.1492 | 0.0545 |
12486.848 | 49.0 | 49000 | 12210.5742 | 0.8347 | 0.3062 | 2.2778 | 0.8347 | 0.8345 | 0.1498 | 0.0546 |
12376.327 | 50.0 | 50000 | 12216.4121 | 0.8337 | 0.3073 | 2.1945 | 0.8337 | 0.8337 | 0.1506 | 0.0535 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2