fold_5
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: 1.1556
- Accuracy: 0.9516
- F1 Score: 0.9547
- Recall: 0.9558
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 |
|---|---|---|---|---|---|---|
| 3.0831 | 1.0 | 19 | 2.9877 | 0.1834 | 0.1506 | 0.1750 |
| 2.9160 | 2.0 | 38 | 2.7975 | 0.2734 | 0.2517 | 0.2583 |
| 2.5816 | 3.0 | 57 | 2.5431 | 0.4671 | 0.4271 | 0.4325 |
| 2.1999 | 4.0 | 76 | 2.2368 | 0.6263 | 0.6019 | 0.6120 |
| 1.8110 | 5.0 | 95 | 1.9213 | 0.7301 | 0.7278 | 0.7223 |
| 1.4822 | 6.0 | 114 | 1.6147 | 0.8304 | 0.8407 | 0.8357 |
| 1.2515 | 7.0 | 133 | 1.4442 | 0.8581 | 0.8646 | 0.8541 |
| 1.1471 | 8.0 | 152 | 1.3540 | 0.8789 | 0.8857 | 0.8784 |
| 1.0819 | 9.0 | 171 | 1.2801 | 0.8962 | 0.9021 | 0.8978 |
| 1.0613 | 10.0 | 190 | 1.2465 | 0.9100 | 0.9142 | 0.9124 |
| 1.0395 | 11.0 | 209 | 1.2235 | 0.9343 | 0.9386 | 0.9387 |
| 1.0240 | 12.0 | 228 | 1.2106 | 0.9377 | 0.9410 | 0.9397 |
| 1.0257 | 13.0 | 247 | 1.1934 | 0.9308 | 0.9358 | 0.9361 |
| 1.0162 | 14.0 | 266 | 1.1946 | 0.9377 | 0.9402 | 0.9384 |
| 0.9940 | 15.0 | 285 | 1.1879 | 0.9377 | 0.9411 | 0.9397 |
| 1.0097 | 16.0 | 304 | 1.1848 | 0.9412 | 0.9449 | 0.9423 |
| 0.9937 | 17.0 | 323 | 1.1863 | 0.9377 | 0.9408 | 0.9384 |
| 0.9864 | 18.0 | 342 | 1.1935 | 0.9446 | 0.9471 | 0.9446 |
| 0.9830 | 19.0 | 361 | 1.1802 | 0.9446 | 0.9474 | 0.9459 |
| 0.9825 | 20.0 | 380 | 1.1508 | 0.9446 | 0.9490 | 0.9485 |
| 0.9733 | 21.0 | 399 | 1.1568 | 0.9343 | 0.9385 | 0.9385 |
| 0.9778 | 22.0 | 418 | 1.1570 | 0.9412 | 0.9455 | 0.9460 |
| 0.9657 | 23.0 | 437 | 1.1588 | 0.9446 | 0.9487 | 0.9497 |
| 0.9771 | 24.0 | 456 | 1.1637 | 0.9446 | 0.9495 | 0.9510 |
| 0.9647 | 25.0 | 475 | 1.1640 | 0.9481 | 0.9509 | 0.9509 |
| 0.9649 | 26.0 | 494 | 1.1572 | 0.9516 | 0.9546 | 0.9546 |
| 0.9686 | 27.0 | 513 | 1.1667 | 0.9377 | 0.9411 | 0.9397 |
| 0.9681 | 28.0 | 532 | 1.1614 | 0.9377 | 0.9417 | 0.9423 |
| 0.9642 | 29.0 | 551 | 1.1599 | 0.9412 | 0.9444 | 0.9434 |
| 0.9677 | 30.0 | 570 | 1.1556 | 0.9516 | 0.9547 | 0.9558 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2