<!-- 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. -->
300-tiny_tobacco3482_kd
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: 0.3298
- Accuracy: 0.79
- Brier Loss: 0.3334
- Nll: 1.0051
- F1 Micro: 0.79
- F1 Macro: 0.7591
- Ece: 0.2152
- Aurc: 0.0601
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: 64
- eval_batch_size: 64
- 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 | 13 | 1.7587 | 0.225 | 0.8896 | 7.6699 | 0.225 | 0.1499 | 0.2894 | 0.7588 |
No log | 2.0 | 26 | 1.1905 | 0.4 | 0.7901 | 3.7933 | 0.4000 | 0.2961 | 0.3234 | 0.4395 |
No log | 3.0 | 39 | 0.9530 | 0.53 | 0.6670 | 2.9907 | 0.53 | 0.4066 | 0.3137 | 0.2821 |
No log | 4.0 | 52 | 0.8046 | 0.615 | 0.5821 | 2.2124 | 0.615 | 0.4898 | 0.3097 | 0.1862 |
No log | 5.0 | 65 | 0.7084 | 0.685 | 0.5084 | 2.1522 | 0.685 | 0.6027 | 0.3191 | 0.1389 |
No log | 6.0 | 78 | 0.7243 | 0.68 | 0.4683 | 2.2673 | 0.68 | 0.5904 | 0.2695 | 0.1279 |
No log | 7.0 | 91 | 0.6734 | 0.67 | 0.4675 | 2.2909 | 0.67 | 0.5844 | 0.2436 | 0.1510 |
No log | 8.0 | 104 | 0.5780 | 0.7 | 0.4215 | 2.0061 | 0.7 | 0.6160 | 0.2418 | 0.1016 |
No log | 9.0 | 117 | 0.6270 | 0.71 | 0.4402 | 1.8620 | 0.7100 | 0.6574 | 0.2485 | 0.1249 |
No log | 10.0 | 130 | 0.5604 | 0.72 | 0.4074 | 1.5914 | 0.72 | 0.6430 | 0.2566 | 0.0935 |
No log | 11.0 | 143 | 0.5814 | 0.705 | 0.4079 | 1.6933 | 0.705 | 0.6190 | 0.2350 | 0.1035 |
No log | 12.0 | 156 | 0.5901 | 0.71 | 0.4176 | 1.7974 | 0.7100 | 0.6472 | 0.2225 | 0.1058 |
No log | 13.0 | 169 | 0.5041 | 0.71 | 0.3918 | 1.8429 | 0.7100 | 0.6562 | 0.2336 | 0.0958 |
No log | 14.0 | 182 | 0.5099 | 0.72 | 0.3982 | 1.6343 | 0.72 | 0.6550 | 0.2202 | 0.1021 |
No log | 15.0 | 195 | 0.4843 | 0.745 | 0.3951 | 1.3599 | 0.745 | 0.6719 | 0.2680 | 0.0884 |
No log | 16.0 | 208 | 0.4529 | 0.74 | 0.3776 | 1.3838 | 0.74 | 0.6951 | 0.2112 | 0.0839 |
No log | 17.0 | 221 | 0.4420 | 0.745 | 0.3782 | 1.4403 | 0.745 | 0.6982 | 0.2285 | 0.0800 |
No log | 18.0 | 234 | 0.4428 | 0.755 | 0.3710 | 1.3696 | 0.755 | 0.7298 | 0.2170 | 0.0825 |
No log | 19.0 | 247 | 0.4306 | 0.75 | 0.3794 | 1.4095 | 0.75 | 0.7235 | 0.2470 | 0.0862 |
No log | 20.0 | 260 | 0.4166 | 0.74 | 0.3648 | 1.2893 | 0.74 | 0.6776 | 0.2312 | 0.0835 |
No log | 21.0 | 273 | 0.3830 | 0.77 | 0.3524 | 1.1764 | 0.7700 | 0.7256 | 0.2535 | 0.0730 |
No log | 22.0 | 286 | 0.3918 | 0.77 | 0.3564 | 1.2293 | 0.7700 | 0.7067 | 0.2372 | 0.0710 |
No log | 23.0 | 299 | 0.4125 | 0.75 | 0.3656 | 1.1419 | 0.75 | 0.7109 | 0.2357 | 0.0766 |
No log | 24.0 | 312 | 0.3771 | 0.785 | 0.3543 | 1.0960 | 0.785 | 0.7583 | 0.2345 | 0.0712 |
No log | 25.0 | 325 | 0.3846 | 0.745 | 0.3613 | 1.0616 | 0.745 | 0.7061 | 0.2060 | 0.0766 |
No log | 26.0 | 338 | 0.3660 | 0.77 | 0.3547 | 1.3094 | 0.7700 | 0.7196 | 0.2515 | 0.0724 |
No log | 27.0 | 351 | 0.3634 | 0.78 | 0.3476 | 1.0645 | 0.78 | 0.7479 | 0.2401 | 0.0677 |
No log | 28.0 | 364 | 0.3715 | 0.755 | 0.3522 | 1.1981 | 0.755 | 0.6984 | 0.2257 | 0.0709 |
No log | 29.0 | 377 | 0.3701 | 0.765 | 0.3597 | 1.1645 | 0.765 | 0.7239 | 0.2631 | 0.0747 |
No log | 30.0 | 390 | 0.3562 | 0.775 | 0.3465 | 1.1094 | 0.775 | 0.7140 | 0.2428 | 0.0659 |
No log | 31.0 | 403 | 0.3811 | 0.775 | 0.3499 | 1.2515 | 0.775 | 0.7368 | 0.2214 | 0.0694 |
No log | 32.0 | 416 | 0.3555 | 0.77 | 0.3439 | 1.1715 | 0.7700 | 0.7053 | 0.2532 | 0.0705 |
No log | 33.0 | 429 | 0.3592 | 0.775 | 0.3449 | 1.1606 | 0.775 | 0.7364 | 0.2336 | 0.0729 |
No log | 34.0 | 442 | 0.3555 | 0.78 | 0.3431 | 1.1054 | 0.78 | 0.7373 | 0.2143 | 0.0653 |
No log | 35.0 | 455 | 0.3454 | 0.77 | 0.3415 | 1.0386 | 0.7700 | 0.7333 | 0.2463 | 0.0668 |
No log | 36.0 | 468 | 0.3403 | 0.8 | 0.3394 | 1.1435 | 0.8000 | 0.7664 | 0.2674 | 0.0625 |
No log | 37.0 | 481 | 0.3390 | 0.785 | 0.3379 | 1.1183 | 0.785 | 0.7552 | 0.2432 | 0.0633 |
No log | 38.0 | 494 | 0.3413 | 0.79 | 0.3347 | 1.1538 | 0.79 | 0.7406 | 0.2239 | 0.0615 |
0.2994 | 39.0 | 507 | 0.3364 | 0.795 | 0.3362 | 0.9975 | 0.795 | 0.7650 | 0.2334 | 0.0639 |
0.2994 | 40.0 | 520 | 0.3340 | 0.79 | 0.3328 | 1.0045 | 0.79 | 0.7466 | 0.2711 | 0.0580 |
0.2994 | 41.0 | 533 | 0.3381 | 0.77 | 0.3391 | 0.9829 | 0.7700 | 0.7427 | 0.2147 | 0.0675 |
0.2994 | 42.0 | 546 | 0.3297 | 0.8 | 0.3319 | 1.0739 | 0.8000 | 0.7685 | 0.2613 | 0.0585 |
0.2994 | 43.0 | 559 | 0.3338 | 0.8 | 0.3373 | 1.1507 | 0.8000 | 0.7719 | 0.2491 | 0.0637 |
0.2994 | 44.0 | 572 | 0.3316 | 0.79 | 0.3359 | 1.1274 | 0.79 | 0.7539 | 0.2469 | 0.0620 |
0.2994 | 45.0 | 585 | 0.3283 | 0.79 | 0.3336 | 1.0644 | 0.79 | 0.7531 | 0.2636 | 0.0612 |
0.2994 | 46.0 | 598 | 0.3297 | 0.8 | 0.3344 | 1.1343 | 0.8000 | 0.7670 | 0.2317 | 0.0600 |
0.2994 | 47.0 | 611 | 0.3293 | 0.79 | 0.3318 | 1.0692 | 0.79 | 0.7542 | 0.2396 | 0.0616 |
0.2994 | 48.0 | 624 | 0.3339 | 0.79 | 0.3357 | 1.1225 | 0.79 | 0.7590 | 0.2508 | 0.0617 |
0.2994 | 49.0 | 637 | 0.3290 | 0.795 | 0.3343 | 1.0692 | 0.795 | 0.7618 | 0.2529 | 0.0604 |
0.2994 | 50.0 | 650 | 0.3298 | 0.79 | 0.3348 | 1.1343 | 0.79 | 0.7591 | 0.2330 | 0.0609 |
0.2994 | 51.0 | 663 | 0.3305 | 0.795 | 0.3330 | 1.0045 | 0.795 | 0.7618 | 0.2357 | 0.0607 |
0.2994 | 52.0 | 676 | 0.3299 | 0.79 | 0.3339 | 1.0722 | 0.79 | 0.7542 | 0.2562 | 0.0614 |
0.2994 | 53.0 | 689 | 0.3280 | 0.8 | 0.3325 | 1.0688 | 0.8000 | 0.7685 | 0.2500 | 0.0593 |
0.2994 | 54.0 | 702 | 0.3284 | 0.795 | 0.3323 | 1.0175 | 0.795 | 0.7618 | 0.2436 | 0.0598 |
0.2994 | 55.0 | 715 | 0.3287 | 0.79 | 0.3331 | 1.0750 | 0.79 | 0.7591 | 0.2497 | 0.0604 |
0.2994 | 56.0 | 728 | 0.3286 | 0.795 | 0.3335 | 1.0115 | 0.795 | 0.7618 | 0.2296 | 0.0602 |
0.2994 | 57.0 | 741 | 0.3285 | 0.79 | 0.3330 | 1.0648 | 0.79 | 0.7591 | 0.2446 | 0.0602 |
0.2994 | 58.0 | 754 | 0.3299 | 0.795 | 0.3339 | 1.0193 | 0.795 | 0.7618 | 0.2345 | 0.0608 |
0.2994 | 59.0 | 767 | 0.3294 | 0.79 | 0.3329 | 1.0139 | 0.79 | 0.7591 | 0.2369 | 0.0601 |
0.2994 | 60.0 | 780 | 0.3292 | 0.795 | 0.3332 | 1.0118 | 0.795 | 0.7618 | 0.2226 | 0.0601 |
0.2994 | 61.0 | 793 | 0.3293 | 0.795 | 0.3333 | 1.0716 | 0.795 | 0.7618 | 0.2282 | 0.0602 |
0.2994 | 62.0 | 806 | 0.3294 | 0.795 | 0.3331 | 1.0107 | 0.795 | 0.7618 | 0.2224 | 0.0601 |
0.2994 | 63.0 | 819 | 0.3295 | 0.795 | 0.3336 | 1.0144 | 0.795 | 0.7618 | 0.2294 | 0.0605 |
0.2994 | 64.0 | 832 | 0.3293 | 0.795 | 0.3332 | 1.0104 | 0.795 | 0.7618 | 0.2324 | 0.0603 |
0.2994 | 65.0 | 845 | 0.3298 | 0.795 | 0.3337 | 1.0114 | 0.795 | 0.7618 | 0.2478 | 0.0606 |
0.2994 | 66.0 | 858 | 0.3297 | 0.795 | 0.3333 | 1.0076 | 0.795 | 0.7618 | 0.2366 | 0.0601 |
0.2994 | 67.0 | 871 | 0.3298 | 0.79 | 0.3338 | 1.0120 | 0.79 | 0.7591 | 0.2513 | 0.0606 |
0.2994 | 68.0 | 884 | 0.3297 | 0.795 | 0.3337 | 1.0110 | 0.795 | 0.7618 | 0.2376 | 0.0605 |
0.2994 | 69.0 | 897 | 0.3297 | 0.795 | 0.3335 | 1.0115 | 0.795 | 0.7618 | 0.2228 | 0.0602 |
0.2994 | 70.0 | 910 | 0.3292 | 0.795 | 0.3333 | 1.0089 | 0.795 | 0.7618 | 0.2215 | 0.0602 |
0.2994 | 71.0 | 923 | 0.3297 | 0.795 | 0.3334 | 1.0083 | 0.795 | 0.7618 | 0.2226 | 0.0600 |
0.2994 | 72.0 | 936 | 0.3297 | 0.79 | 0.3335 | 1.0072 | 0.79 | 0.7591 | 0.2257 | 0.0604 |
0.2994 | 73.0 | 949 | 0.3297 | 0.795 | 0.3332 | 1.0060 | 0.795 | 0.7618 | 0.2381 | 0.0600 |
0.2994 | 74.0 | 962 | 0.3295 | 0.795 | 0.3335 | 1.0082 | 0.795 | 0.7618 | 0.2366 | 0.0603 |
0.2994 | 75.0 | 975 | 0.3296 | 0.79 | 0.3334 | 1.0089 | 0.79 | 0.7591 | 0.2373 | 0.0601 |
0.2994 | 76.0 | 988 | 0.3298 | 0.795 | 0.3334 | 1.0098 | 0.795 | 0.7618 | 0.2310 | 0.0602 |
0.0006 | 77.0 | 1001 | 0.3297 | 0.79 | 0.3334 | 1.0084 | 0.79 | 0.7591 | 0.2228 | 0.0603 |
0.0006 | 78.0 | 1014 | 0.3297 | 0.79 | 0.3333 | 1.0071 | 0.79 | 0.7591 | 0.2148 | 0.0600 |
0.0006 | 79.0 | 1027 | 0.3298 | 0.795 | 0.3334 | 1.0059 | 0.795 | 0.7618 | 0.2309 | 0.0602 |
0.0006 | 80.0 | 1040 | 0.3298 | 0.795 | 0.3334 | 1.0046 | 0.795 | 0.7618 | 0.2309 | 0.0602 |
0.0006 | 81.0 | 1053 | 0.3298 | 0.79 | 0.3335 | 1.0073 | 0.79 | 0.7591 | 0.2239 | 0.0602 |
0.0006 | 82.0 | 1066 | 0.3298 | 0.795 | 0.3336 | 1.0072 | 0.795 | 0.7618 | 0.2317 | 0.0603 |
0.0006 | 83.0 | 1079 | 0.3297 | 0.795 | 0.3334 | 1.0055 | 0.795 | 0.7618 | 0.2224 | 0.0601 |
0.0006 | 84.0 | 1092 | 0.3298 | 0.79 | 0.3335 | 1.0061 | 0.79 | 0.7591 | 0.2240 | 0.0601 |
0.0006 | 85.0 | 1105 | 0.3297 | 0.79 | 0.3334 | 1.0052 | 0.79 | 0.7591 | 0.2322 | 0.0601 |
0.0006 | 86.0 | 1118 | 0.3298 | 0.79 | 0.3335 | 1.0059 | 0.79 | 0.7591 | 0.2323 | 0.0602 |
0.0006 | 87.0 | 1131 | 0.3298 | 0.79 | 0.3335 | 1.0065 | 0.79 | 0.7591 | 0.2152 | 0.0602 |
0.0006 | 88.0 | 1144 | 0.3298 | 0.79 | 0.3335 | 1.0056 | 0.79 | 0.7591 | 0.2235 | 0.0603 |
0.0006 | 89.0 | 1157 | 0.3297 | 0.79 | 0.3334 | 1.0050 | 0.79 | 0.7591 | 0.2152 | 0.0602 |
0.0006 | 90.0 | 1170 | 0.3297 | 0.79 | 0.3334 | 1.0049 | 0.79 | 0.7591 | 0.2153 | 0.0602 |
0.0006 | 91.0 | 1183 | 0.3297 | 0.79 | 0.3334 | 1.0059 | 0.79 | 0.7591 | 0.2234 | 0.0601 |
0.0006 | 92.0 | 1196 | 0.3298 | 0.79 | 0.3334 | 1.0049 | 0.79 | 0.7591 | 0.2152 | 0.0602 |
0.0006 | 93.0 | 1209 | 0.3299 | 0.79 | 0.3335 | 1.0056 | 0.79 | 0.7591 | 0.2152 | 0.0601 |
0.0006 | 94.0 | 1222 | 0.3298 | 0.79 | 0.3335 | 1.0049 | 0.79 | 0.7591 | 0.2152 | 0.0602 |
0.0006 | 95.0 | 1235 | 0.3298 | 0.79 | 0.3334 | 1.0048 | 0.79 | 0.7591 | 0.2152 | 0.0602 |
0.0006 | 96.0 | 1248 | 0.3298 | 0.79 | 0.3334 | 1.0050 | 0.79 | 0.7591 | 0.2152 | 0.0601 |
0.0006 | 97.0 | 1261 | 0.3298 | 0.79 | 0.3335 | 1.0053 | 0.79 | 0.7591 | 0.2152 | 0.0602 |
0.0006 | 98.0 | 1274 | 0.3298 | 0.79 | 0.3334 | 1.0051 | 0.79 | 0.7591 | 0.2152 | 0.0602 |
0.0006 | 99.0 | 1287 | 0.3298 | 0.79 | 0.3334 | 1.0052 | 0.79 | 0.7591 | 0.2152 | 0.0601 |
0.0006 | 100.0 | 1300 | 0.3298 | 0.79 | 0.3334 | 1.0051 | 0.79 | 0.7591 | 0.2152 | 0.0601 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2