vit_fold_3
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: 1.0572
- Accuracy: 0.9828
- F1 Score: 0.9840
- Recall: 0.9841
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 |
|---|---|---|---|---|---|---|
| 2.9536 | 1.0 | 19 | 2.9528 | 0.3 | 0.2473 | 0.3174 |
| 2.7882 | 2.0 | 38 | 2.6926 | 0.3690 | 0.3570 | 0.3867 |
| 2.4219 | 3.0 | 57 | 2.3561 | 0.5793 | 0.5998 | 0.5921 |
| 2.0092 | 4.0 | 76 | 1.9862 | 0.7586 | 0.7827 | 0.7875 |
| 1.6129 | 5.0 | 95 | 1.6200 | 0.8724 | 0.8863 | 0.8884 |
| 1.3723 | 6.0 | 114 | 1.3893 | 0.9138 | 0.9220 | 0.9240 |
| 1.2462 | 7.0 | 133 | 1.2759 | 0.9276 | 0.9335 | 0.9325 |
| 1.1739 | 8.0 | 152 | 1.2234 | 0.9379 | 0.9427 | 0.9411 |
| 1.1189 | 9.0 | 171 | 1.1838 | 0.9448 | 0.9486 | 0.9460 |
| 1.1044 | 10.0 | 190 | 1.1554 | 0.9517 | 0.9557 | 0.9545 |
| 1.0877 | 11.0 | 209 | 1.1483 | 0.9483 | 0.9529 | 0.9533 |
| 1.0598 | 12.0 | 228 | 1.1355 | 0.9552 | 0.9591 | 0.9595 |
| 1.0458 | 13.0 | 247 | 1.1253 | 0.9586 | 0.9624 | 0.9645 |
| 1.0242 | 14.0 | 266 | 1.0983 | 0.9724 | 0.9742 | 0.9730 |
| 1.0151 | 15.0 | 285 | 1.0982 | 0.9621 | 0.9644 | 0.9606 |
| 1.0158 | 16.0 | 304 | 1.0982 | 0.9621 | 0.9656 | 0.9682 |
| 1.0131 | 17.0 | 323 | 1.0913 | 0.9655 | 0.9685 | 0.9707 |
| 1.0065 | 18.0 | 342 | 1.0878 | 0.9655 | 0.9685 | 0.9707 |
| 0.9920 | 19.0 | 361 | 1.0783 | 0.9690 | 0.9708 | 0.9680 |
| 0.9799 | 20.0 | 380 | 1.0697 | 0.9793 | 0.9808 | 0.9791 |
| 0.9830 | 21.0 | 399 | 1.0664 | 0.9793 | 0.9808 | 0.9791 |
| 0.9798 | 22.0 | 418 | 1.0630 | 0.9793 | 0.9810 | 0.9816 |
| 0.9896 | 23.0 | 437 | 1.0678 | 0.9724 | 0.9750 | 0.9767 |
| 0.9769 | 24.0 | 456 | 1.0585 | 0.9793 | 0.9809 | 0.9803 |
| 0.9736 | 25.0 | 475 | 1.0572 | 0.9828 | 0.9840 | 0.9841 |
| 0.9697 | 26.0 | 494 | 1.0557 | 0.9793 | 0.9809 | 0.9803 |
| 0.9685 | 27.0 | 513 | 1.0585 | 0.9793 | 0.9808 | 0.9791 |
| 0.9706 | 28.0 | 532 | 1.0609 | 0.9724 | 0.9749 | 0.9754 |
| 0.9833 | 29.0 | 551 | 1.0702 | 0.9724 | 0.9750 | 0.9767 |
| 0.9717 | 30.0 | 570 | 1.0728 | 0.9759 | 0.9776 | 0.9753 |
| 0.9727 | 31.0 | 589 | 1.0614 | 0.9759 | 0.9779 | 0.9779 |
| 0.9645 | 32.0 | 608 | 1.0531 | 0.9793 | 0.9809 | 0.9803 |
| 0.9730 | 33.0 | 627 | 1.0505 | 0.9828 | 0.9839 | 0.9828 |
| 0.9642 | 34.0 | 646 | 1.0454 | 0.9828 | 0.9839 | 0.9828 |
| 0.9687 | 35.0 | 665 | 1.0516 | 0.9759 | 0.9779 | 0.9779 |
| 0.9650 | 36.0 | 684 | 1.0544 | 0.9724 | 0.9748 | 0.9754 |
| 0.9651 | 37.0 | 703 | 1.0512 | 0.9793 | 0.9809 | 0.9803 |
| 0.9675 | 38.0 | 722 | 1.0516 | 0.9793 | 0.9808 | 0.9791 |
| 0.9622 | 39.0 | 741 | 1.0499 | 0.9759 | 0.9780 | 0.9792 |
| 0.9641 | 40.0 | 760 | 1.0488 | 0.9828 | 0.9839 | 0.9828 |
| 0.9640 | 41.0 | 779 | 1.0479 | 0.9828 | 0.9839 | 0.9828 |
| 0.9619 | 42.0 | 798 | 1.0464 | 0.9828 | 0.9839 | 0.9828 |
| 0.9602 | 43.0 | 817 | 1.0481 | 0.9793 | 0.9809 | 0.9803 |
| 0.9670 | 44.0 | 836 | 1.0482 | 0.9828 | 0.9839 | 0.9828 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2