deit_fold_2
This model is a fine-tuned version of facebook/deit-small-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 1.1405
- Accuracy: 0.9586
- F1 Score: 0.9620
- Recall: 0.9621
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 |
|---|---|---|---|---|---|---|
| 2.8182 | 1.0 | 19 | 2.7982 | 0.2966 | 0.2243 | 0.2444 |
| 2.7152 | 2.0 | 38 | 2.6468 | 0.3793 | 0.2689 | 0.3032 |
| 2.5094 | 3.0 | 57 | 2.4150 | 0.5241 | 0.4335 | 0.4481 |
| 2.1993 | 4.0 | 76 | 2.0760 | 0.7172 | 0.7150 | 0.6964 |
| 1.7513 | 5.0 | 95 | 1.7070 | 0.8069 | 0.8177 | 0.8141 |
| 1.4469 | 6.0 | 114 | 1.4631 | 0.8586 | 0.8728 | 0.8742 |
| 1.2482 | 7.0 | 133 | 1.3515 | 0.8759 | 0.8882 | 0.8890 |
| 1.2058 | 8.0 | 152 | 1.3018 | 0.8897 | 0.8990 | 0.8998 |
| 1.1470 | 9.0 | 171 | 1.2735 | 0.9069 | 0.9141 | 0.9169 |
| 1.0887 | 10.0 | 190 | 1.2533 | 0.9207 | 0.9291 | 0.9306 |
| 1.0903 | 11.0 | 209 | 1.2413 | 0.9103 | 0.9182 | 0.9159 |
| 1.0576 | 12.0 | 228 | 1.2260 | 0.9207 | 0.9286 | 0.9317 |
| 1.0231 | 13.0 | 247 | 1.2594 | 0.9207 | 0.9278 | 0.9303 |
| 1.0264 | 14.0 | 266 | 1.2164 | 0.9276 | 0.9339 | 0.9305 |
| 1.0175 | 15.0 | 285 | 1.2294 | 0.9241 | 0.9301 | 0.9267 |
| 1.0135 | 16.0 | 304 | 1.2143 | 0.9345 | 0.9406 | 0.9402 |
| 1.0178 | 17.0 | 323 | 1.2314 | 0.9276 | 0.9351 | 0.9380 |
| 0.9823 | 18.0 | 342 | 1.2160 | 0.9310 | 0.9373 | 0.9365 |
| 0.9965 | 19.0 | 361 | 1.2154 | 0.9172 | 0.9232 | 0.9206 |
| 0.9854 | 20.0 | 380 | 1.1846 | 0.9345 | 0.9411 | 0.9415 |
| 0.9807 | 21.0 | 399 | 1.2096 | 0.9345 | 0.9414 | 0.9414 |
| 0.9860 | 22.0 | 418 | 1.1996 | 0.9276 | 0.9338 | 0.9316 |
| 0.9796 | 23.0 | 437 | 1.1967 | 0.9310 | 0.9365 | 0.9317 |
| 0.9826 | 24.0 | 456 | 1.2180 | 0.9172 | 0.9239 | 0.9183 |
| 0.9809 | 25.0 | 475 | 1.2030 | 0.9345 | 0.9405 | 0.9378 |
| 0.9803 | 26.0 | 494 | 1.1866 | 0.9345 | 0.9403 | 0.9389 |
| 0.9748 | 27.0 | 513 | 1.1626 | 0.9448 | 0.9497 | 0.9487 |
| 0.9659 | 28.0 | 532 | 1.1405 | 0.9586 | 0.9620 | 0.9621 |
| 0.9718 | 29.0 | 551 | 1.1410 | 0.9483 | 0.9525 | 0.9525 |
| 0.9668 | 30.0 | 570 | 1.1485 | 0.9552 | 0.9589 | 0.9584 |
| 0.9715 | 31.0 | 589 | 1.1423 | 0.9448 | 0.9499 | 0.9502 |
| 0.9729 | 32.0 | 608 | 1.1560 | 0.9483 | 0.9535 | 0.9550 |
| 0.9766 | 33.0 | 627 | 1.1721 | 0.9483 | 0.9541 | 0.9561 |
| 0.9687 | 34.0 | 646 | 1.1706 | 0.9448 | 0.9509 | 0.9524 |
| 0.9769 | 35.0 | 665 | 1.1539 | 0.9345 | 0.9410 | 0.9403 |
| 0.9609 | 36.0 | 684 | 1.1552 | 0.9483 | 0.9539 | 0.9549 |
| 0.9587 | 37.0 | 703 | 1.1565 | 0.9517 | 0.9569 | 0.9597 |
| 0.9673 | 38.0 | 722 | 1.1675 | 0.9483 | 0.9537 | 0.9562 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2