smids_3x_deit_small_adamax_001_fold3
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: 0.9786
- Accuracy: 0.8933
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 |
|---|---|---|---|---|
| 0.4746 | 1.0 | 225 | 0.4469 | 0.7917 |
| 0.3712 | 2.0 | 450 | 0.4363 | 0.8183 |
| 0.2858 | 3.0 | 675 | 0.3457 | 0.8733 |
| 0.1805 | 4.0 | 900 | 0.4133 | 0.865 |
| 0.1699 | 5.0 | 1125 | 0.3336 | 0.89 |
| 0.1967 | 6.0 | 1350 | 0.3958 | 0.88 |
| 0.1372 | 7.0 | 1575 | 0.4702 | 0.87 |
| 0.1698 | 8.0 | 1800 | 0.3428 | 0.8867 |
| 0.1529 | 9.0 | 2025 | 0.5351 | 0.8583 |
| 0.0621 | 10.0 | 2250 | 0.5492 | 0.8733 |
| 0.1311 | 11.0 | 2475 | 0.5821 | 0.8767 |
| 0.1035 | 12.0 | 2700 | 0.6050 | 0.8667 |
| 0.0328 | 13.0 | 2925 | 0.6304 | 0.875 |
| 0.036 | 14.0 | 3150 | 0.5969 | 0.8983 |
| 0.0422 | 15.0 | 3375 | 0.6326 | 0.8817 |
| 0.0088 | 16.0 | 3600 | 0.6452 | 0.8733 |
| 0.0148 | 17.0 | 3825 | 0.5210 | 0.8883 |
| 0.0267 | 18.0 | 4050 | 0.6134 | 0.88 |
| 0.0523 | 19.0 | 4275 | 0.6448 | 0.8933 |
| 0.0182 | 20.0 | 4500 | 0.7298 | 0.8783 |
| 0.0004 | 21.0 | 4725 | 0.7641 | 0.8683 |
| 0.0037 | 22.0 | 4950 | 0.8419 | 0.87 |
| 0.0038 | 23.0 | 5175 | 0.7762 | 0.8817 |
| 0.0002 | 24.0 | 5400 | 0.7488 | 0.8883 |
| 0.0367 | 25.0 | 5625 | 0.6912 | 0.8767 |
| 0.0002 | 26.0 | 5850 | 0.6890 | 0.895 |
| 0.0002 | 27.0 | 6075 | 0.6160 | 0.9017 |
| 0.0001 | 28.0 | 6300 | 0.6922 | 0.8967 |
| 0.007 | 29.0 | 6525 | 0.8317 | 0.885 |
| 0.0083 | 30.0 | 6750 | 0.6909 | 0.8983 |
| 0.0048 | 31.0 | 6975 | 0.7613 | 0.8967 |
| 0.0 | 32.0 | 7200 | 0.8055 | 0.895 |
| 0.0 | 33.0 | 7425 | 0.8267 | 0.8917 |
| 0.0 | 34.0 | 7650 | 0.8303 | 0.8917 |
| 0.0 | 35.0 | 7875 | 0.8741 | 0.8967 |
| 0.0 | 36.0 | 8100 | 0.8381 | 0.8967 |
| 0.0024 | 37.0 | 8325 | 0.8583 | 0.8983 |
| 0.0 | 38.0 | 8550 | 0.9234 | 0.8917 |
| 0.0 | 39.0 | 8775 | 0.8565 | 0.8967 |
| 0.0 | 40.0 | 9000 | 0.8898 | 0.8917 |
| 0.0 | 41.0 | 9225 | 0.9312 | 0.8917 |
| 0.0 | 42.0 | 9450 | 0.9410 | 0.8933 |
| 0.0 | 43.0 | 9675 | 0.9514 | 0.8933 |
| 0.0 | 44.0 | 9900 | 0.9553 | 0.8933 |
| 0.0 | 45.0 | 10125 | 0.9599 | 0.8933 |
| 0.0 | 46.0 | 10350 | 0.9610 | 0.8933 |
| 0.0027 | 47.0 | 10575 | 0.9689 | 0.8933 |
| 0.0 | 48.0 | 10800 | 0.9724 | 0.8933 |
| 0.0 | 49.0 | 11025 | 0.9761 | 0.8933 |
| 0.0 | 50.0 | 11250 | 0.9786 | 0.8933 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2