deit_fold_1
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: 1.1170
- Accuracy: 0.9655
- F1 Score: 0.9681
- Recall: 0.9681
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 |
|---|---|---|---|---|---|---|
| 2.7052 | 1.0 | 19 | 2.6182 | 0.3897 | 0.2859 | 0.3210 |
| 2.6037 | 2.0 | 38 | 2.4936 | 0.4345 | 0.3591 | 0.3697 |
| 2.4310 | 3.0 | 57 | 2.2883 | 0.5552 | 0.5442 | 0.5290 |
| 2.1283 | 4.0 | 76 | 1.9885 | 0.7172 | 0.7322 | 0.7214 |
| 1.7269 | 5.0 | 95 | 1.6339 | 0.8172 | 0.8313 | 0.8241 |
| 1.4880 | 6.0 | 114 | 1.3974 | 0.8655 | 0.8730 | 0.8676 |
| 1.2856 | 7.0 | 133 | 1.2969 | 0.9 | 0.9069 | 0.9031 |
| 1.2589 | 8.0 | 152 | 1.2535 | 0.9069 | 0.9112 | 0.9057 |
| 1.1813 | 9.0 | 171 | 1.1953 | 0.9241 | 0.9300 | 0.9304 |
| 1.1441 | 10.0 | 190 | 1.1895 | 0.9379 | 0.9417 | 0.9389 |
| 1.1279 | 11.0 | 209 | 1.1635 | 0.9552 | 0.9589 | 0.9584 |
| 1.0956 | 12.0 | 228 | 1.1425 | 0.9483 | 0.9524 | 0.9511 |
| 1.0689 | 13.0 | 247 | 1.1503 | 0.9483 | 0.9511 | 0.9521 |
| 1.0613 | 14.0 | 266 | 1.1557 | 0.9552 | 0.9574 | 0.9571 |
| 1.0226 | 15.0 | 285 | 1.1592 | 0.9448 | 0.9478 | 0.9449 |
| 1.0242 | 16.0 | 304 | 1.1320 | 0.9483 | 0.9511 | 0.9510 |
| 1.0160 | 17.0 | 323 | 1.1371 | 0.9552 | 0.9584 | 0.9571 |
| 0.9991 | 18.0 | 342 | 1.1267 | 0.9621 | 0.9650 | 0.9657 |
| 1.0108 | 19.0 | 361 | 1.1350 | 0.9517 | 0.9551 | 0.9534 |
| 1.0175 | 20.0 | 380 | 1.1404 | 0.9586 | 0.9617 | 0.9608 |
| 0.9991 | 21.0 | 399 | 1.1422 | 0.9586 | 0.9614 | 0.9596 |
| 0.9951 | 22.0 | 418 | 1.1277 | 0.9552 | 0.9590 | 0.9608 |
| 0.9753 | 23.0 | 437 | 1.1113 | 0.9586 | 0.9620 | 0.9632 |
| 0.9748 | 24.0 | 456 | 1.1117 | 0.9552 | 0.9590 | 0.9608 |
| 0.9748 | 25.0 | 475 | 1.1086 | 0.9621 | 0.9648 | 0.9645 |
| 0.9704 | 26.0 | 494 | 1.1182 | 0.9586 | 0.9619 | 0.9620 |
| 0.9748 | 27.0 | 513 | 1.1061 | 0.9586 | 0.9620 | 0.9632 |
| 0.9637 | 28.0 | 532 | 1.0796 | 0.9621 | 0.9652 | 0.9669 |
| 0.9696 | 29.0 | 551 | 1.1170 | 0.9655 | 0.9681 | 0.9681 |
| 0.9656 | 30.0 | 570 | 1.0977 | 0.9621 | 0.9652 | 0.9669 |
| 0.9679 | 31.0 | 589 | 1.1275 | 0.9586 | 0.9621 | 0.9632 |
| 0.9699 | 32.0 | 608 | 1.0992 | 0.9586 | 0.9622 | 0.9645 |
| 0.9760 | 33.0 | 627 | 1.1147 | 0.9621 | 0.9647 | 0.9632 |
| 0.9595 | 34.0 | 646 | 1.0978 | 0.9621 | 0.9653 | 0.9681 |
| 0.9619 | 35.0 | 665 | 1.1000 | 0.9621 | 0.9648 | 0.9645 |
| 0.9634 | 36.0 | 684 | 1.1140 | 0.9655 | 0.9679 | 0.9669 |
| 0.9554 | 37.0 | 703 | 1.1082 | 0.9655 | 0.9679 | 0.9669 |
| 0.9644 | 38.0 | 722 | 1.1166 | 0.9655 | 0.9681 | 0.9681 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2