vit_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: 1.1208
- Accuracy: 0.9654
- F1 Score: 0.9686
- Recall: 0.9680
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 |
|---|---|---|---|---|---|---|
| 2.7928 | 1.0 | 19 | 2.8746 | 0.3080 | 0.2543 | 0.2785 |
| 2.6676 | 2.0 | 38 | 2.6793 | 0.3910 | 0.3269 | 0.3445 |
| 2.3938 | 3.0 | 57 | 2.3879 | 0.5190 | 0.4584 | 0.4678 |
| 2.0704 | 4.0 | 76 | 2.0268 | 0.7128 | 0.7070 | 0.6885 |
| 1.6362 | 5.0 | 95 | 1.6508 | 0.8651 | 0.8722 | 0.8623 |
| 1.3687 | 6.0 | 114 | 1.3848 | 0.8997 | 0.9090 | 0.9068 |
| 1.2580 | 7.0 | 133 | 1.2794 | 0.9100 | 0.9172 | 0.9094 |
| 1.1350 | 8.0 | 152 | 1.2075 | 0.9308 | 0.9372 | 0.9339 |
| 1.0887 | 9.0 | 171 | 1.1870 | 0.9343 | 0.9402 | 0.9363 |
| 1.0795 | 10.0 | 190 | 1.1688 | 0.9481 | 0.9531 | 0.9497 |
| 1.0515 | 11.0 | 209 | 1.1478 | 0.9516 | 0.9564 | 0.9547 |
| 1.0498 | 12.0 | 228 | 1.1544 | 0.9481 | 0.9524 | 0.9525 |
| 1.0147 | 13.0 | 247 | 1.1355 | 0.9585 | 0.9625 | 0.9609 |
| 1.0256 | 14.0 | 266 | 1.1517 | 0.9481 | 0.9529 | 0.9484 |
| 1.0107 | 15.0 | 285 | 1.1487 | 0.9481 | 0.9527 | 0.9511 |
| 1.0068 | 16.0 | 304 | 1.1287 | 0.9550 | 0.9594 | 0.9583 |
| 0.9904 | 17.0 | 323 | 1.1279 | 0.9550 | 0.9589 | 0.9584 |
| 0.9846 | 18.0 | 342 | 1.1228 | 0.9585 | 0.9620 | 0.9620 |
| 0.9907 | 19.0 | 361 | 1.1302 | 0.9550 | 0.9588 | 0.9572 |
| 0.9867 | 20.0 | 380 | 1.1309 | 0.9585 | 0.9620 | 0.9609 |
| 0.9810 | 21.0 | 399 | 1.1288 | 0.9516 | 0.9556 | 0.9534 |
| 0.9740 | 22.0 | 418 | 1.1208 | 0.9654 | 0.9686 | 0.9680 |
| 0.9674 | 23.0 | 437 | 1.1232 | 0.9619 | 0.9656 | 0.9656 |
| 0.9809 | 24.0 | 456 | 1.1269 | 0.9585 | 0.9624 | 0.9607 |
| 0.9679 | 25.0 | 475 | 1.1233 | 0.9585 | 0.9624 | 0.9607 |
| 0.9755 | 26.0 | 494 | 1.1295 | 0.9585 | 0.9624 | 0.9607 |
| 0.9730 | 27.0 | 513 | 1.1253 | 0.9550 | 0.9589 | 0.9584 |
| 0.9661 | 28.0 | 532 | 1.1280 | 0.9550 | 0.9588 | 0.9572 |
| 0.9692 | 29.0 | 551 | 1.1154 | 0.9585 | 0.9620 | 0.9609 |
| 0.9650 | 30.0 | 570 | 1.1211 | 0.9585 | 0.9620 | 0.9609 |
| 0.9671 | 31.0 | 589 | 1.1155 | 0.9619 | 0.9657 | 0.9657 |
| 0.9640 | 32.0 | 608 | 1.1290 | 0.9585 | 0.9620 | 0.9609 |
| 0.9675 | 33.0 | 627 | 1.1330 | 0.9585 | 0.9621 | 0.9621 |
| 0.9647 | 34.0 | 646 | 1.1300 | 0.9585 | 0.9625 | 0.9620 |
| 0.9634 | 35.0 | 665 | 1.1283 | 0.9585 | 0.9621 | 0.9621 |
| 0.9696 | 36.0 | 684 | 1.1407 | 0.9481 | 0.9524 | 0.9497 |
| 0.9605 | 37.0 | 703 | 1.1275 | 0.9585 | 0.9621 | 0.9621 |
| 0.9624 | 38.0 | 722 | 1.1291 | 0.9619 | 0.9651 | 0.9646 |
| 0.9678 | 39.0 | 741 | 1.1230 | 0.9619 | 0.9651 | 0.9646 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2