fold_4
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: 1.1328
- Accuracy: 0.9654
- F1 Score: 0.9688
- Recall: 0.9684
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 |
|---|---|---|---|---|---|---|
| 2.7279 | 1.0 | 19 | 2.6961 | 0.3287 | 0.3257 | 0.3264 |
| 2.6497 | 2.0 | 38 | 2.5197 | 0.4775 | 0.4752 | 0.4600 |
| 2.3856 | 3.0 | 57 | 2.2554 | 0.6574 | 0.6632 | 0.6472 |
| 2.0467 | 4.0 | 76 | 1.9435 | 0.8062 | 0.8186 | 0.8137 |
| 1.6696 | 5.0 | 95 | 1.6132 | 0.8720 | 0.8800 | 0.8750 |
| 1.3967 | 6.0 | 114 | 1.3746 | 0.9170 | 0.9249 | 0.9244 |
| 1.2758 | 7.0 | 133 | 1.2437 | 0.9412 | 0.9473 | 0.9422 |
| 1.1517 | 8.0 | 152 | 1.1767 | 0.9550 | 0.9593 | 0.9559 |
| 1.1069 | 9.0 | 171 | 1.1511 | 0.9516 | 0.9556 | 0.9512 |
| 1.0847 | 10.0 | 190 | 1.1386 | 0.9550 | 0.9589 | 0.9584 |
| 1.0566 | 11.0 | 209 | 1.1328 | 0.9654 | 0.9688 | 0.9684 |
| 1.0535 | 12.0 | 228 | 1.1346 | 0.9550 | 0.9597 | 0.9609 |
| 1.0154 | 13.0 | 247 | 1.1250 | 0.9550 | 0.9595 | 0.9595 |
| 1.0196 | 14.0 | 266 | 1.1252 | 0.9585 | 0.9625 | 0.9609 |
| 1.0003 | 15.0 | 285 | 1.1256 | 0.9481 | 0.9529 | 0.9524 |
| 0.9951 | 16.0 | 304 | 1.1191 | 0.9619 | 0.9653 | 0.9659 |
| 0.9877 | 17.0 | 323 | 1.1133 | 0.9619 | 0.9651 | 0.9646 |
| 0.9841 | 18.0 | 342 | 1.1038 | 0.9619 | 0.9651 | 0.9646 |
| 0.9887 | 19.0 | 361 | 1.1080 | 0.9619 | 0.9650 | 0.9635 |
| 0.9794 | 20.0 | 380 | 1.1105 | 0.9619 | 0.9651 | 0.9646 |
| 0.9786 | 21.0 | 399 | 1.1086 | 0.9654 | 0.9681 | 0.9671 |
| 0.9719 | 22.0 | 418 | 1.1084 | 0.9550 | 0.9591 | 0.9597 |
| 0.9723 | 23.0 | 437 | 1.1242 | 0.9516 | 0.9562 | 0.9584 |
| 0.9813 | 24.0 | 456 | 1.0986 | 0.9654 | 0.9681 | 0.9671 |
| 0.9682 | 25.0 | 475 | 1.0959 | 0.9654 | 0.9681 | 0.9671 |
| 0.9750 | 26.0 | 494 | 1.0990 | 0.9654 | 0.9681 | 0.9671 |
| 0.9716 | 27.0 | 513 | 1.0995 | 0.9619 | 0.9651 | 0.9646 |
| 0.9701 | 28.0 | 532 | 1.0949 | 0.9654 | 0.9681 | 0.9671 |
| 0.9694 | 29.0 | 551 | 1.0997 | 0.9654 | 0.9681 | 0.9671 |
| 0.9637 | 30.0 | 570 | 1.1002 | 0.9654 | 0.9681 | 0.9671 |
| 0.9647 | 31.0 | 589 | 1.0992 | 0.9654 | 0.9681 | 0.9671 |
| 0.9658 | 32.0 | 608 | 1.1084 | 0.9619 | 0.9650 | 0.9635 |
| 0.9686 | 33.0 | 627 | 1.1156 | 0.9619 | 0.9652 | 0.9659 |
| 0.9632 | 34.0 | 646 | 1.1068 | 0.9619 | 0.9650 | 0.9634 |
| 0.9630 | 35.0 | 665 | 1.1040 | 0.9619 | 0.9650 | 0.9634 |
| 0.9680 | 36.0 | 684 | 1.1131 | 0.9619 | 0.9650 | 0.9634 |
| 0.9579 | 37.0 | 703 | 1.1063 | 0.9619 | 0.9650 | 0.9634 |
| 0.9615 | 38.0 | 722 | 1.1063 | 0.9619 | 0.9650 | 0.9634 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2