<!-- 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. -->
6_e_200-tiny_tobacco3482_kd_CEKD_t2.5_a0.9
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: 0.5528
- Accuracy: 0.84
- Brier Loss: 0.2493
- Nll: 1.6062
- F1 Micro: 0.8400
- F1 Macro: 0.8256
- Ece: 0.1626
- Aurc: 0.0556
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: 32
- eval_batch_size: 32
- 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 | 25 | 1.9962 | 0.24 | 0.8634 | 4.6099 | 0.24 | 0.1987 | 0.2924 | 0.7628 |
No log | 2.0 | 50 | 1.2785 | 0.545 | 0.5960 | 2.6707 | 0.545 | 0.4456 | 0.2808 | 0.2569 |
No log | 3.0 | 75 | 0.9740 | 0.685 | 0.4455 | 2.0688 | 0.685 | 0.5937 | 0.2231 | 0.1314 |
No log | 4.0 | 100 | 0.8052 | 0.75 | 0.3628 | 1.5271 | 0.75 | 0.7144 | 0.2031 | 0.0915 |
No log | 5.0 | 125 | 0.7531 | 0.77 | 0.3324 | 1.5448 | 0.7700 | 0.7348 | 0.1868 | 0.0829 |
No log | 6.0 | 150 | 0.9730 | 0.735 | 0.4050 | 1.5875 | 0.735 | 0.7229 | 0.1899 | 0.1118 |
No log | 7.0 | 175 | 0.6437 | 0.83 | 0.2790 | 1.3915 | 0.83 | 0.7996 | 0.1967 | 0.0612 |
No log | 8.0 | 200 | 0.6670 | 0.78 | 0.2984 | 1.2128 | 0.78 | 0.7429 | 0.1701 | 0.0716 |
No log | 9.0 | 225 | 0.6012 | 0.845 | 0.2521 | 1.4007 | 0.845 | 0.8208 | 0.1573 | 0.0581 |
No log | 10.0 | 250 | 0.6754 | 0.795 | 0.3063 | 1.4194 | 0.795 | 0.7638 | 0.2036 | 0.0759 |
No log | 11.0 | 275 | 0.5950 | 0.83 | 0.2554 | 1.1558 | 0.83 | 0.8053 | 0.1877 | 0.0529 |
No log | 12.0 | 300 | 0.7061 | 0.79 | 0.3153 | 1.6781 | 0.79 | 0.7676 | 0.1879 | 0.0848 |
No log | 13.0 | 325 | 0.6349 | 0.805 | 0.2806 | 1.3102 | 0.805 | 0.7767 | 0.1523 | 0.0667 |
No log | 14.0 | 350 | 0.5973 | 0.82 | 0.2677 | 1.5498 | 0.82 | 0.8020 | 0.1734 | 0.0567 |
No log | 15.0 | 375 | 0.6186 | 0.825 | 0.2792 | 1.3816 | 0.825 | 0.8170 | 0.1558 | 0.0672 |
No log | 16.0 | 400 | 0.5694 | 0.815 | 0.2662 | 1.1759 | 0.815 | 0.7962 | 0.1675 | 0.0559 |
No log | 17.0 | 425 | 0.5993 | 0.825 | 0.2793 | 1.2547 | 0.825 | 0.8112 | 0.1822 | 0.0647 |
No log | 18.0 | 450 | 0.6333 | 0.815 | 0.2844 | 1.6540 | 0.815 | 0.8024 | 0.1562 | 0.0622 |
No log | 19.0 | 475 | 0.5208 | 0.845 | 0.2349 | 1.2879 | 0.845 | 0.8155 | 0.1553 | 0.0494 |
0.4544 | 20.0 | 500 | 0.5412 | 0.86 | 0.2438 | 1.6726 | 0.8600 | 0.8465 | 0.1531 | 0.0485 |
0.4544 | 21.0 | 525 | 0.6171 | 0.825 | 0.2775 | 1.9997 | 0.825 | 0.8183 | 0.1464 | 0.0549 |
0.4544 | 22.0 | 550 | 0.5479 | 0.84 | 0.2447 | 1.5015 | 0.8400 | 0.8263 | 0.1481 | 0.0680 |
0.4544 | 23.0 | 575 | 0.5508 | 0.835 | 0.2491 | 1.8095 | 0.835 | 0.8209 | 0.1616 | 0.0469 |
0.4544 | 24.0 | 600 | 0.5597 | 0.825 | 0.2577 | 1.6676 | 0.825 | 0.8077 | 0.1572 | 0.0486 |
0.4544 | 25.0 | 625 | 0.5505 | 0.835 | 0.2535 | 1.6085 | 0.835 | 0.8166 | 0.1664 | 0.0524 |
0.4544 | 26.0 | 650 | 0.5347 | 0.84 | 0.2442 | 1.4694 | 0.8400 | 0.8288 | 0.1825 | 0.0505 |
0.4544 | 27.0 | 675 | 0.5333 | 0.84 | 0.2418 | 1.5809 | 0.8400 | 0.8280 | 0.1634 | 0.0521 |
0.4544 | 28.0 | 700 | 0.5417 | 0.84 | 0.2471 | 1.5289 | 0.8400 | 0.8231 | 0.1500 | 0.0503 |
0.4544 | 29.0 | 725 | 0.5369 | 0.845 | 0.2434 | 1.5333 | 0.845 | 0.8318 | 0.1690 | 0.0523 |
0.4544 | 30.0 | 750 | 0.5396 | 0.84 | 0.2448 | 1.5269 | 0.8400 | 0.8260 | 0.1689 | 0.0534 |
0.4544 | 31.0 | 775 | 0.5411 | 0.845 | 0.2459 | 1.5325 | 0.845 | 0.8289 | 0.1524 | 0.0514 |
0.4544 | 32.0 | 800 | 0.5429 | 0.845 | 0.2456 | 1.5239 | 0.845 | 0.8318 | 0.1550 | 0.0527 |
0.4544 | 33.0 | 825 | 0.5445 | 0.84 | 0.2468 | 1.5275 | 0.8400 | 0.8231 | 0.1626 | 0.0535 |
0.4544 | 34.0 | 850 | 0.5432 | 0.845 | 0.2461 | 1.5210 | 0.845 | 0.8289 | 0.1557 | 0.0533 |
0.4544 | 35.0 | 875 | 0.5438 | 0.845 | 0.2459 | 1.5269 | 0.845 | 0.8318 | 0.1564 | 0.0533 |
0.4544 | 36.0 | 900 | 0.5451 | 0.845 | 0.2466 | 1.5262 | 0.845 | 0.8289 | 0.1610 | 0.0541 |
0.4544 | 37.0 | 925 | 0.5415 | 0.85 | 0.2448 | 1.5254 | 0.85 | 0.8348 | 0.1667 | 0.0528 |
0.4544 | 38.0 | 950 | 0.5447 | 0.845 | 0.2461 | 1.5367 | 0.845 | 0.8318 | 0.1519 | 0.0535 |
0.4544 | 39.0 | 975 | 0.5437 | 0.85 | 0.2454 | 1.5223 | 0.85 | 0.8348 | 0.1605 | 0.0536 |
0.0607 | 40.0 | 1000 | 0.5445 | 0.845 | 0.2460 | 1.5252 | 0.845 | 0.8318 | 0.1610 | 0.0539 |
0.0607 | 41.0 | 1025 | 0.5460 | 0.845 | 0.2465 | 1.5925 | 0.845 | 0.8318 | 0.1416 | 0.0541 |
0.0607 | 42.0 | 1050 | 0.5466 | 0.84 | 0.2467 | 1.5304 | 0.8400 | 0.8260 | 0.1555 | 0.0542 |
0.0607 | 43.0 | 1075 | 0.5458 | 0.84 | 0.2464 | 1.5272 | 0.8400 | 0.8231 | 0.1633 | 0.0539 |
0.0607 | 44.0 | 1100 | 0.5460 | 0.85 | 0.2464 | 1.5459 | 0.85 | 0.8377 | 0.1534 | 0.0550 |
0.0607 | 45.0 | 1125 | 0.5464 | 0.85 | 0.2465 | 1.5390 | 0.85 | 0.8377 | 0.1471 | 0.0544 |
0.0607 | 46.0 | 1150 | 0.5462 | 0.85 | 0.2465 | 1.5972 | 0.85 | 0.8377 | 0.1549 | 0.0540 |
0.0607 | 47.0 | 1175 | 0.5475 | 0.85 | 0.2472 | 1.5910 | 0.85 | 0.8377 | 0.1592 | 0.0546 |
0.0607 | 48.0 | 1200 | 0.5482 | 0.845 | 0.2475 | 1.5943 | 0.845 | 0.8294 | 0.1548 | 0.0545 |
0.0607 | 49.0 | 1225 | 0.5475 | 0.845 | 0.2471 | 1.5922 | 0.845 | 0.8294 | 0.1534 | 0.0545 |
0.0607 | 50.0 | 1250 | 0.5476 | 0.85 | 0.2470 | 1.5908 | 0.85 | 0.8377 | 0.1539 | 0.0545 |
0.0607 | 51.0 | 1275 | 0.5480 | 0.845 | 0.2471 | 1.5990 | 0.845 | 0.8322 | 0.1545 | 0.0547 |
0.0607 | 52.0 | 1300 | 0.5479 | 0.85 | 0.2469 | 1.5917 | 0.85 | 0.8348 | 0.1688 | 0.0547 |
0.0607 | 53.0 | 1325 | 0.5479 | 0.845 | 0.2472 | 1.6052 | 0.845 | 0.8322 | 0.1545 | 0.0543 |
0.0607 | 54.0 | 1350 | 0.5490 | 0.85 | 0.2477 | 1.5948 | 0.85 | 0.8348 | 0.1610 | 0.0545 |
0.0607 | 55.0 | 1375 | 0.5489 | 0.85 | 0.2474 | 1.5967 | 0.85 | 0.8377 | 0.1543 | 0.0560 |
0.0607 | 56.0 | 1400 | 0.5499 | 0.845 | 0.2480 | 1.5939 | 0.845 | 0.8294 | 0.1561 | 0.0549 |
0.0607 | 57.0 | 1425 | 0.5492 | 0.845 | 0.2476 | 1.6048 | 0.845 | 0.8322 | 0.1570 | 0.0549 |
0.0607 | 58.0 | 1450 | 0.5497 | 0.845 | 0.2478 | 1.6004 | 0.845 | 0.8322 | 0.1724 | 0.0548 |
0.0607 | 59.0 | 1475 | 0.5496 | 0.85 | 0.2477 | 1.5982 | 0.85 | 0.8377 | 0.1634 | 0.0546 |
0.0589 | 60.0 | 1500 | 0.5497 | 0.845 | 0.2478 | 1.5969 | 0.845 | 0.8322 | 0.1592 | 0.0545 |
0.0589 | 61.0 | 1525 | 0.5492 | 0.85 | 0.2476 | 1.6095 | 0.85 | 0.8377 | 0.1630 | 0.0547 |
0.0589 | 62.0 | 1550 | 0.5507 | 0.845 | 0.2483 | 1.6060 | 0.845 | 0.8322 | 0.1649 | 0.0554 |
0.0589 | 63.0 | 1575 | 0.5490 | 0.845 | 0.2474 | 1.6021 | 0.845 | 0.8322 | 0.1635 | 0.0546 |
0.0589 | 64.0 | 1600 | 0.5508 | 0.845 | 0.2483 | 1.5970 | 0.845 | 0.8294 | 0.1697 | 0.0552 |
0.0589 | 65.0 | 1625 | 0.5505 | 0.84 | 0.2483 | 1.6023 | 0.8400 | 0.8256 | 0.1658 | 0.0553 |
0.0589 | 66.0 | 1650 | 0.5503 | 0.845 | 0.2481 | 1.6032 | 0.845 | 0.8322 | 0.1637 | 0.0546 |
0.0589 | 67.0 | 1675 | 0.5514 | 0.84 | 0.2486 | 1.6000 | 0.8400 | 0.8227 | 0.1649 | 0.0559 |
0.0589 | 68.0 | 1700 | 0.5516 | 0.84 | 0.2487 | 1.5979 | 0.8400 | 0.8227 | 0.1649 | 0.0550 |
0.0589 | 69.0 | 1725 | 0.5510 | 0.84 | 0.2485 | 1.6005 | 0.8400 | 0.8256 | 0.1639 | 0.0548 |
0.0589 | 70.0 | 1750 | 0.5510 | 0.84 | 0.2484 | 1.5990 | 0.8400 | 0.8256 | 0.1653 | 0.0549 |
0.0589 | 71.0 | 1775 | 0.5517 | 0.84 | 0.2487 | 1.6080 | 0.8400 | 0.8256 | 0.1640 | 0.0558 |
0.0589 | 72.0 | 1800 | 0.5525 | 0.84 | 0.2491 | 1.6069 | 0.8400 | 0.8227 | 0.1669 | 0.0558 |
0.0589 | 73.0 | 1825 | 0.5519 | 0.84 | 0.2488 | 1.6147 | 0.8400 | 0.8256 | 0.1638 | 0.0554 |
0.0589 | 74.0 | 1850 | 0.5519 | 0.84 | 0.2487 | 1.6027 | 0.8400 | 0.8256 | 0.1657 | 0.0558 |
0.0589 | 75.0 | 1875 | 0.5522 | 0.84 | 0.2490 | 1.6082 | 0.8400 | 0.8256 | 0.1717 | 0.0556 |
0.0589 | 76.0 | 1900 | 0.5523 | 0.84 | 0.2489 | 1.6022 | 0.8400 | 0.8256 | 0.1645 | 0.0553 |
0.0589 | 77.0 | 1925 | 0.5514 | 0.84 | 0.2486 | 1.6027 | 0.8400 | 0.8256 | 0.1635 | 0.0551 |
0.0589 | 78.0 | 1950 | 0.5518 | 0.84 | 0.2488 | 1.6007 | 0.8400 | 0.8256 | 0.1641 | 0.0556 |
0.0589 | 79.0 | 1975 | 0.5522 | 0.84 | 0.2490 | 1.6057 | 0.8400 | 0.8256 | 0.1637 | 0.0556 |
0.0588 | 80.0 | 2000 | 0.5520 | 0.84 | 0.2489 | 1.6110 | 0.8400 | 0.8256 | 0.1658 | 0.0552 |
0.0588 | 81.0 | 2025 | 0.5521 | 0.84 | 0.2489 | 1.6047 | 0.8400 | 0.8256 | 0.1659 | 0.0555 |
0.0588 | 82.0 | 2050 | 0.5521 | 0.84 | 0.2490 | 1.6015 | 0.8400 | 0.8256 | 0.1635 | 0.0551 |
0.0588 | 83.0 | 2075 | 0.5521 | 0.84 | 0.2489 | 1.6115 | 0.8400 | 0.8256 | 0.1637 | 0.0553 |
0.0588 | 84.0 | 2100 | 0.5523 | 0.84 | 0.2490 | 1.6033 | 0.8400 | 0.8256 | 0.1738 | 0.0553 |
0.0588 | 85.0 | 2125 | 0.5525 | 0.84 | 0.2491 | 1.6072 | 0.8400 | 0.8256 | 0.1658 | 0.0555 |
0.0588 | 86.0 | 2150 | 0.5521 | 0.84 | 0.2489 | 1.6057 | 0.8400 | 0.8256 | 0.1574 | 0.0553 |
0.0588 | 87.0 | 2175 | 0.5527 | 0.84 | 0.2492 | 1.6605 | 0.8400 | 0.8256 | 0.1610 | 0.0555 |
0.0588 | 88.0 | 2200 | 0.5526 | 0.84 | 0.2491 | 1.6056 | 0.8400 | 0.8256 | 0.1544 | 0.0556 |
0.0588 | 89.0 | 2225 | 0.5527 | 0.84 | 0.2492 | 1.6126 | 0.8400 | 0.8256 | 0.1547 | 0.0556 |
0.0588 | 90.0 | 2250 | 0.5525 | 0.84 | 0.2491 | 1.6059 | 0.8400 | 0.8256 | 0.1525 | 0.0556 |
0.0588 | 91.0 | 2275 | 0.5528 | 0.84 | 0.2492 | 1.6060 | 0.8400 | 0.8256 | 0.1604 | 0.0556 |
0.0588 | 92.0 | 2300 | 0.5526 | 0.84 | 0.2491 | 1.6080 | 0.8400 | 0.8256 | 0.1525 | 0.0555 |
0.0588 | 93.0 | 2325 | 0.5527 | 0.84 | 0.2492 | 1.6034 | 0.8400 | 0.8256 | 0.1547 | 0.0556 |
0.0588 | 94.0 | 2350 | 0.5526 | 0.84 | 0.2492 | 1.6040 | 0.8400 | 0.8256 | 0.1673 | 0.0555 |
0.0588 | 95.0 | 2375 | 0.5529 | 0.84 | 0.2493 | 1.6053 | 0.8400 | 0.8256 | 0.1545 | 0.0556 |
0.0588 | 96.0 | 2400 | 0.5526 | 0.84 | 0.2492 | 1.6050 | 0.8400 | 0.8256 | 0.1626 | 0.0555 |
0.0588 | 97.0 | 2425 | 0.5528 | 0.84 | 0.2492 | 1.6040 | 0.8400 | 0.8256 | 0.1686 | 0.0557 |
0.0588 | 98.0 | 2450 | 0.5528 | 0.84 | 0.2492 | 1.6068 | 0.8400 | 0.8256 | 0.1626 | 0.0555 |
0.0588 | 99.0 | 2475 | 0.5528 | 0.84 | 0.2492 | 1.6065 | 0.8400 | 0.8256 | 0.1626 | 0.0556 |
0.0588 | 100.0 | 2500 | 0.5528 | 0.84 | 0.2493 | 1.6062 | 0.8400 | 0.8256 | 0.1626 | 0.0556 |
Framework versions
- Transformers 4.30.2
- Pytorch 1.13.1
- Datasets 2.13.1
- Tokenizers 0.13.3