vit_fold_2
This model is a fine-tuned version of google/vit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 1.1330
- Accuracy: 0.9552
- F1 Score: 0.9591
- Recall: 0.9611
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: 1e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 150
- num_epochs: 100
- label_smoothing_factor: 0.15
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 Score | Recall |
|---|---|---|---|---|---|---|
| 3.0764 | 1.0 | 19 | 3.0505 | 0.2069 | 0.1685 | 0.1989 |
| 2.8915 | 2.0 | 38 | 2.8592 | 0.3276 | 0.2738 | 0.2986 |
| 2.5571 | 3.0 | 57 | 2.5948 | 0.4724 | 0.4113 | 0.4311 |
| 2.1836 | 4.0 | 76 | 2.2633 | 0.6241 | 0.6056 | 0.6021 |
| 1.7834 | 5.0 | 95 | 1.9104 | 0.7759 | 0.7743 | 0.7669 |
| 1.4444 | 6.0 | 114 | 1.6022 | 0.8448 | 0.8531 | 0.8439 |
| 1.2156 | 7.0 | 133 | 1.4031 | 0.8724 | 0.8811 | 0.8695 |
| 1.1415 | 8.0 | 152 | 1.3280 | 0.8793 | 0.8863 | 0.8743 |
| 1.0863 | 9.0 | 171 | 1.2605 | 0.9069 | 0.9147 | 0.9123 |
| 1.0580 | 10.0 | 190 | 1.2364 | 0.9207 | 0.9275 | 0.9257 |
| 1.0411 | 11.0 | 209 | 1.2385 | 0.9172 | 0.9225 | 0.9171 |
| 1.0299 | 12.0 | 228 | 1.1877 | 0.9310 | 0.9368 | 0.9355 |
| 1.0096 | 13.0 | 247 | 1.1798 | 0.9345 | 0.9406 | 0.9428 |
| 0.9902 | 14.0 | 266 | 1.1914 | 0.9241 | 0.9291 | 0.9233 |
| 0.9951 | 15.0 | 285 | 1.1893 | 0.9241 | 0.9291 | 0.9233 |
| 1.0118 | 16.0 | 304 | 1.1582 | 0.9414 | 0.9465 | 0.9477 |
| 1.0025 | 17.0 | 323 | 1.1507 | 0.9483 | 0.9525 | 0.9539 |
| 0.9899 | 18.0 | 342 | 1.1432 | 0.9448 | 0.9490 | 0.9478 |
| 0.9766 | 19.0 | 361 | 1.1452 | 0.9448 | 0.9490 | 0.9478 |
| 0.9757 | 20.0 | 380 | 1.1502 | 0.9448 | 0.9497 | 0.9515 |
| 0.9742 | 21.0 | 399 | 1.1370 | 0.9483 | 0.9522 | 0.9515 |
| 0.9779 | 22.0 | 418 | 1.1409 | 0.9414 | 0.9459 | 0.9442 |
| 0.9699 | 23.0 | 437 | 1.1304 | 0.9448 | 0.9490 | 0.9478 |
| 0.9857 | 24.0 | 456 | 1.1345 | 0.9414 | 0.9459 | 0.9442 |
| 0.9649 | 25.0 | 475 | 1.1319 | 0.9483 | 0.9524 | 0.9539 |
| 0.9665 | 26.0 | 494 | 1.1383 | 0.9448 | 0.9495 | 0.9514 |
| 0.9725 | 27.0 | 513 | 1.1625 | 0.9414 | 0.9468 | 0.9502 |
| 0.9801 | 28.0 | 532 | 1.1431 | 0.9483 | 0.9525 | 0.9539 |
| 0.9628 | 29.0 | 551 | 1.1296 | 0.9483 | 0.9524 | 0.9527 |
| 0.9617 | 30.0 | 570 | 1.1306 | 0.9448 | 0.9495 | 0.9503 |
| 0.9642 | 31.0 | 589 | 1.1330 | 0.9552 | 0.9591 | 0.9611 |
| 0.9632 | 32.0 | 608 | 1.1253 | 0.9483 | 0.9527 | 0.9525 |
| 0.9768 | 33.0 | 627 | 1.1471 | 0.9483 | 0.9528 | 0.9551 |
| 0.9642 | 34.0 | 646 | 1.1479 | 0.9483 | 0.9528 | 0.9551 |
| 0.9688 | 35.0 | 665 | 1.1393 | 0.9483 | 0.9527 | 0.9525 |
| 0.9639 | 36.0 | 684 | 1.1510 | 0.9483 | 0.9531 | 0.9562 |
| 0.9625 | 37.0 | 703 | 1.1417 | 0.9448 | 0.9497 | 0.9489 |
| 0.9624 | 38.0 | 722 | 1.1501 | 0.9517 | 0.9561 | 0.9586 |
| 0.9635 | 39.0 | 741 | 1.1514 | 0.9448 | 0.9498 | 0.9513 |
| 0.9616 | 40.0 | 760 | 1.1471 | 0.9414 | 0.9466 | 0.9478 |
| 0.9619 | 41.0 | 779 | 1.1508 | 0.9414 | 0.9468 | 0.9491 |
| 0.9585 | 42.0 | 798 | 1.1411 | 0.9448 | 0.9499 | 0.9502 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2