smids_10x_deit_tiny_adamax_001_fold3
This model is a fine-tuned version of facebook/deit-tiny-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 1.0576
- Accuracy: 0.9133
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: 0.001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.366 | 1.0 | 750 | 0.3478 | 0.86 |
| 0.3154 | 2.0 | 1500 | 0.3552 | 0.8633 |
| 0.2007 | 3.0 | 2250 | 0.4696 | 0.845 |
| 0.2296 | 4.0 | 3000 | 0.3387 | 0.8783 |
| 0.2607 | 5.0 | 3750 | 0.4239 | 0.865 |
| 0.2365 | 6.0 | 4500 | 0.3514 | 0.895 |
| 0.1244 | 7.0 | 5250 | 0.3538 | 0.8917 |
| 0.1356 | 8.0 | 6000 | 0.3984 | 0.895 |
| 0.0526 | 9.0 | 6750 | 0.4940 | 0.875 |
| 0.1195 | 10.0 | 7500 | 0.5287 | 0.8783 |
| 0.0559 | 11.0 | 8250 | 0.6054 | 0.8817 |
| 0.0512 | 12.0 | 9000 | 0.6374 | 0.8717 |
| 0.0498 | 13.0 | 9750 | 0.6405 | 0.8817 |
| 0.0296 | 14.0 | 10500 | 0.6601 | 0.895 |
| 0.0626 | 15.0 | 11250 | 0.7807 | 0.89 |
| 0.055 | 16.0 | 12000 | 0.7694 | 0.905 |
| 0.0444 | 17.0 | 12750 | 0.6413 | 0.905 |
| 0.0317 | 18.0 | 13500 | 0.7330 | 0.9033 |
| 0.0108 | 19.0 | 14250 | 0.7464 | 0.8917 |
| 0.0236 | 20.0 | 15000 | 0.7591 | 0.885 |
| 0.01 | 21.0 | 15750 | 0.8264 | 0.9067 |
| 0.0226 | 22.0 | 16500 | 0.7921 | 0.8933 |
| 0.0127 | 23.0 | 17250 | 0.7486 | 0.9033 |
| 0.0025 | 24.0 | 18000 | 0.8018 | 0.8983 |
| 0.0004 | 25.0 | 18750 | 0.7411 | 0.9083 |
| 0.0 | 26.0 | 19500 | 0.8554 | 0.895 |
| 0.0 | 27.0 | 20250 | 0.9122 | 0.9017 |
| 0.0 | 28.0 | 21000 | 0.8611 | 0.9067 |
| 0.0041 | 29.0 | 21750 | 0.8741 | 0.9033 |
| 0.0 | 30.0 | 22500 | 0.7969 | 0.9167 |
| 0.012 | 31.0 | 23250 | 0.8521 | 0.91 |
| 0.0058 | 32.0 | 24000 | 0.9974 | 0.8983 |
| 0.0 | 33.0 | 24750 | 0.9864 | 0.9 |
| 0.0 | 34.0 | 25500 | 0.8709 | 0.91 |
| 0.0 | 35.0 | 26250 | 0.9411 | 0.9117 |
| 0.0 | 36.0 | 27000 | 1.0050 | 0.9033 |
| 0.0 | 37.0 | 27750 | 0.9456 | 0.905 |
| 0.0 | 38.0 | 28500 | 0.9323 | 0.9083 |
| 0.0 | 39.0 | 29250 | 0.9349 | 0.9117 |
| 0.0 | 40.0 | 30000 | 0.9420 | 0.9117 |
| 0.0 | 41.0 | 30750 | 0.9601 | 0.9133 |
| 0.0 | 42.0 | 31500 | 0.9780 | 0.9133 |
| 0.0 | 43.0 | 32250 | 0.9953 | 0.9133 |
| 0.0 | 44.0 | 33000 | 1.0029 | 0.915 |
| 0.0 | 45.0 | 33750 | 1.0208 | 0.9133 |
| 0.0 | 46.0 | 34500 | 1.0335 | 0.9133 |
| 0.0 | 47.0 | 35250 | 1.0420 | 0.9133 |
| 0.0 | 48.0 | 36000 | 1.0522 | 0.9133 |
| 0.0 | 49.0 | 36750 | 1.0572 | 0.9117 |
| 0.0 | 50.0 | 37500 | 1.0576 | 0.9133 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2