smids_3x_deit_tiny_adamax_0001_fold2
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: 1.0425
- Accuracy: 0.8769
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: 0.0001
- 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 |
|---|---|---|---|---|
| 0.3075 | 1.0 | 225 | 0.3475 | 0.8502 |
| 0.1874 | 2.0 | 450 | 0.4269 | 0.8519 |
| 0.1627 | 3.0 | 675 | 0.3974 | 0.8652 |
| 0.1048 | 4.0 | 900 | 0.4256 | 0.8669 |
| 0.0866 | 5.0 | 1125 | 0.5024 | 0.8686 |
| 0.0482 | 6.0 | 1350 | 0.5673 | 0.8752 |
| 0.0764 | 7.0 | 1575 | 0.8107 | 0.8652 |
| 0.0186 | 8.0 | 1800 | 0.7251 | 0.8885 |
| 0.0334 | 9.0 | 2025 | 0.7854 | 0.8686 |
| 0.0009 | 10.0 | 2250 | 0.8924 | 0.8619 |
| 0.0392 | 11.0 | 2475 | 0.8255 | 0.8686 |
| 0.0005 | 12.0 | 2700 | 0.9410 | 0.8686 |
| 0.0101 | 13.0 | 2925 | 0.9076 | 0.8702 |
| 0.0133 | 14.0 | 3150 | 1.0016 | 0.8602 |
| 0.0 | 15.0 | 3375 | 0.8996 | 0.8752 |
| 0.0001 | 16.0 | 3600 | 0.9808 | 0.8636 |
| 0.0141 | 17.0 | 3825 | 1.0333 | 0.8586 |
| 0.0001 | 18.0 | 4050 | 1.0084 | 0.8686 |
| 0.0101 | 19.0 | 4275 | 0.9333 | 0.8769 |
| 0.0 | 20.0 | 4500 | 0.9573 | 0.8752 |
| 0.0 | 21.0 | 4725 | 0.9496 | 0.8819 |
| 0.0 | 22.0 | 4950 | 0.9897 | 0.8785 |
| 0.0057 | 23.0 | 5175 | 1.0229 | 0.8719 |
| 0.0 | 24.0 | 5400 | 0.9747 | 0.8785 |
| 0.0089 | 25.0 | 5625 | 1.1564 | 0.8602 |
| 0.0065 | 26.0 | 5850 | 1.0567 | 0.8735 |
| 0.0 | 27.0 | 6075 | 1.0448 | 0.8686 |
| 0.0 | 28.0 | 6300 | 1.0218 | 0.8769 |
| 0.0 | 29.0 | 6525 | 1.0556 | 0.8735 |
| 0.007 | 30.0 | 6750 | 0.9967 | 0.8719 |
| 0.0 | 31.0 | 6975 | 1.0179 | 0.8769 |
| 0.0 | 32.0 | 7200 | 1.0202 | 0.8752 |
| 0.0 | 33.0 | 7425 | 1.0274 | 0.8735 |
| 0.0 | 34.0 | 7650 | 1.0303 | 0.8669 |
| 0.0 | 35.0 | 7875 | 1.0234 | 0.8686 |
| 0.0038 | 36.0 | 8100 | 1.0313 | 0.8702 |
| 0.0 | 37.0 | 8325 | 1.0205 | 0.8735 |
| 0.0 | 38.0 | 8550 | 1.0300 | 0.8702 |
| 0.0041 | 39.0 | 8775 | 1.0336 | 0.8752 |
| 0.0 | 40.0 | 9000 | 1.0365 | 0.8735 |
| 0.0 | 41.0 | 9225 | 1.0406 | 0.8735 |
| 0.0 | 42.0 | 9450 | 1.0330 | 0.8752 |
| 0.0 | 43.0 | 9675 | 1.0322 | 0.8752 |
| 0.0 | 44.0 | 9900 | 1.0369 | 0.8752 |
| 0.0 | 45.0 | 10125 | 1.0431 | 0.8769 |
| 0.0 | 46.0 | 10350 | 1.0388 | 0.8752 |
| 0.0 | 47.0 | 10575 | 1.0369 | 0.8785 |
| 0.0 | 48.0 | 10800 | 1.0413 | 0.8769 |
| 0.0025 | 49.0 | 11025 | 1.0423 | 0.8769 |
| 0.0025 | 50.0 | 11250 | 1.0425 | 0.8769 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2