smids_3x_deit_small_adamax_00001_fold2
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: 0.9125
- Accuracy: 0.8735
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: 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 |
|---|---|---|---|---|
| 0.3392 | 1.0 | 225 | 0.3664 | 0.8502 |
| 0.2605 | 2.0 | 450 | 0.3323 | 0.8719 |
| 0.212 | 3.0 | 675 | 0.3215 | 0.8686 |
| 0.2229 | 4.0 | 900 | 0.3309 | 0.8652 |
| 0.106 | 5.0 | 1125 | 0.3345 | 0.8802 |
| 0.0845 | 6.0 | 1350 | 0.3616 | 0.8719 |
| 0.0626 | 7.0 | 1575 | 0.3907 | 0.8686 |
| 0.0326 | 8.0 | 1800 | 0.4483 | 0.8669 |
| 0.0372 | 9.0 | 2025 | 0.4833 | 0.8652 |
| 0.0087 | 10.0 | 2250 | 0.5521 | 0.8735 |
| 0.0217 | 11.0 | 2475 | 0.5679 | 0.8752 |
| 0.0111 | 12.0 | 2700 | 0.6269 | 0.8702 |
| 0.011 | 13.0 | 2925 | 0.6480 | 0.8702 |
| 0.0061 | 14.0 | 3150 | 0.6728 | 0.8686 |
| 0.0004 | 15.0 | 3375 | 0.7336 | 0.8669 |
| 0.0093 | 16.0 | 3600 | 0.7662 | 0.8702 |
| 0.0044 | 17.0 | 3825 | 0.7704 | 0.8752 |
| 0.0001 | 18.0 | 4050 | 0.7907 | 0.8735 |
| 0.0005 | 19.0 | 4275 | 0.7929 | 0.8669 |
| 0.0001 | 20.0 | 4500 | 0.8179 | 0.8669 |
| 0.0001 | 21.0 | 4725 | 0.8135 | 0.8785 |
| 0.0001 | 22.0 | 4950 | 0.8581 | 0.8702 |
| 0.0037 | 23.0 | 5175 | 0.8366 | 0.8719 |
| 0.0001 | 24.0 | 5400 | 0.8672 | 0.8686 |
| 0.0168 | 25.0 | 5625 | 0.8621 | 0.8686 |
| 0.0001 | 26.0 | 5850 | 0.8633 | 0.8702 |
| 0.0 | 27.0 | 6075 | 0.8679 | 0.8669 |
| 0.0001 | 28.0 | 6300 | 0.8863 | 0.8735 |
| 0.0001 | 29.0 | 6525 | 0.8794 | 0.8686 |
| 0.0145 | 30.0 | 6750 | 0.8923 | 0.8686 |
| 0.0 | 31.0 | 6975 | 0.8806 | 0.8719 |
| 0.0 | 32.0 | 7200 | 0.8844 | 0.8686 |
| 0.0001 | 33.0 | 7425 | 0.8917 | 0.8669 |
| 0.0 | 34.0 | 7650 | 0.8891 | 0.8719 |
| 0.0 | 35.0 | 7875 | 0.8984 | 0.8735 |
| 0.0077 | 36.0 | 8100 | 0.8879 | 0.8752 |
| 0.0 | 37.0 | 8325 | 0.9058 | 0.8702 |
| 0.0 | 38.0 | 8550 | 0.9002 | 0.8686 |
| 0.0096 | 39.0 | 8775 | 0.9018 | 0.8752 |
| 0.0 | 40.0 | 9000 | 0.9051 | 0.8752 |
| 0.0 | 41.0 | 9225 | 0.9023 | 0.8702 |
| 0.0 | 42.0 | 9450 | 0.9103 | 0.8752 |
| 0.0 | 43.0 | 9675 | 0.9151 | 0.8735 |
| 0.0 | 44.0 | 9900 | 0.9097 | 0.8735 |
| 0.0 | 45.0 | 10125 | 0.9063 | 0.8702 |
| 0.0 | 46.0 | 10350 | 0.9129 | 0.8735 |
| 0.0 | 47.0 | 10575 | 0.9170 | 0.8735 |
| 0.0 | 48.0 | 10800 | 0.9138 | 0.8735 |
| 0.0048 | 49.0 | 11025 | 0.9128 | 0.8735 |
| 0.0048 | 50.0 | 11250 | 0.9125 | 0.8735 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2