smids_3x_deit_small_adamax_001_fold5
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: 0.9259
- Accuracy: 0.8867
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 |
|---|---|---|---|---|
| 0.5128 | 1.0 | 225 | 0.3960 | 0.845 |
| 0.2912 | 2.0 | 450 | 0.3545 | 0.855 |
| 0.2355 | 3.0 | 675 | 0.5138 | 0.8283 |
| 0.2322 | 4.0 | 900 | 0.3502 | 0.8567 |
| 0.1566 | 5.0 | 1125 | 0.4327 | 0.855 |
| 0.1654 | 6.0 | 1350 | 0.3396 | 0.8883 |
| 0.1711 | 7.0 | 1575 | 0.4306 | 0.8733 |
| 0.1936 | 8.0 | 1800 | 0.4649 | 0.8783 |
| 0.1252 | 9.0 | 2025 | 0.4662 | 0.875 |
| 0.0897 | 10.0 | 2250 | 0.4869 | 0.8767 |
| 0.0556 | 11.0 | 2475 | 0.4677 | 0.8883 |
| 0.117 | 12.0 | 2700 | 0.5865 | 0.8667 |
| 0.1159 | 13.0 | 2925 | 0.6366 | 0.87 |
| 0.0526 | 14.0 | 3150 | 0.5638 | 0.8917 |
| 0.0199 | 15.0 | 3375 | 0.6092 | 0.885 |
| 0.0105 | 16.0 | 3600 | 0.7275 | 0.8817 |
| 0.0173 | 17.0 | 3825 | 0.6827 | 0.8883 |
| 0.0471 | 18.0 | 4050 | 0.6845 | 0.87 |
| 0.0272 | 19.0 | 4275 | 0.6814 | 0.8933 |
| 0.0245 | 20.0 | 4500 | 0.7971 | 0.8767 |
| 0.0042 | 21.0 | 4725 | 0.7308 | 0.8867 |
| 0.0003 | 22.0 | 4950 | 0.8304 | 0.895 |
| 0.0001 | 23.0 | 5175 | 0.8296 | 0.8917 |
| 0.0235 | 24.0 | 5400 | 0.7781 | 0.88 |
| 0.0005 | 25.0 | 5625 | 0.8114 | 0.8833 |
| 0.0112 | 26.0 | 5850 | 0.8185 | 0.8833 |
| 0.0 | 27.0 | 6075 | 0.9165 | 0.8833 |
| 0.0034 | 28.0 | 6300 | 0.9111 | 0.8783 |
| 0.0037 | 29.0 | 6525 | 0.8969 | 0.8833 |
| 0.0001 | 30.0 | 6750 | 0.9635 | 0.8783 |
| 0.0 | 31.0 | 6975 | 0.8389 | 0.8933 |
| 0.0 | 32.0 | 7200 | 0.9477 | 0.8767 |
| 0.003 | 33.0 | 7425 | 0.8612 | 0.8917 |
| 0.0001 | 34.0 | 7650 | 0.8689 | 0.8983 |
| 0.0025 | 35.0 | 7875 | 0.8114 | 0.885 |
| 0.0 | 36.0 | 8100 | 0.8426 | 0.8867 |
| 0.0037 | 37.0 | 8325 | 0.8625 | 0.8833 |
| 0.0006 | 38.0 | 8550 | 0.8725 | 0.885 |
| 0.0 | 39.0 | 8775 | 0.8789 | 0.8833 |
| 0.0 | 40.0 | 9000 | 0.8405 | 0.885 |
| 0.0 | 41.0 | 9225 | 0.8737 | 0.8833 |
| 0.0 | 42.0 | 9450 | 0.9022 | 0.8867 |
| 0.003 | 43.0 | 9675 | 0.9115 | 0.8833 |
| 0.0 | 44.0 | 9900 | 0.9099 | 0.8867 |
| 0.0 | 45.0 | 10125 | 0.9104 | 0.8883 |
| 0.0 | 46.0 | 10350 | 0.9175 | 0.8867 |
| 0.0 | 47.0 | 10575 | 0.9200 | 0.8867 |
| 0.0 | 48.0 | 10800 | 0.9224 | 0.8867 |
| 0.0 | 49.0 | 11025 | 0.9241 | 0.8867 |
| 0.0 | 50.0 | 11250 | 0.9259 | 0.8867 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2