smids_3x_deit_tiny_adamax_00001_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: 0.9651
- Accuracy: 0.8719
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 |
|---|---|---|---|---|
| 0.4234 | 1.0 | 225 | 0.4415 | 0.8186 |
| 0.3257 | 2.0 | 450 | 0.3946 | 0.8353 |
| 0.2484 | 3.0 | 675 | 0.3209 | 0.8719 |
| 0.2642 | 4.0 | 900 | 0.3296 | 0.8702 |
| 0.1771 | 5.0 | 1125 | 0.3172 | 0.8702 |
| 0.173 | 6.0 | 1350 | 0.3544 | 0.8586 |
| 0.1484 | 7.0 | 1575 | 0.3447 | 0.8702 |
| 0.0971 | 8.0 | 1800 | 0.3412 | 0.8785 |
| 0.1373 | 9.0 | 2025 | 0.3898 | 0.8735 |
| 0.0316 | 10.0 | 2250 | 0.4151 | 0.8702 |
| 0.056 | 11.0 | 2475 | 0.4368 | 0.8669 |
| 0.0751 | 12.0 | 2700 | 0.5052 | 0.8752 |
| 0.0299 | 13.0 | 2925 | 0.5143 | 0.8769 |
| 0.0139 | 14.0 | 3150 | 0.5498 | 0.8802 |
| 0.0304 | 15.0 | 3375 | 0.6069 | 0.8752 |
| 0.0339 | 16.0 | 3600 | 0.6246 | 0.8785 |
| 0.0114 | 17.0 | 3825 | 0.6695 | 0.8735 |
| 0.0005 | 18.0 | 4050 | 0.7207 | 0.8702 |
| 0.0014 | 19.0 | 4275 | 0.7338 | 0.8669 |
| 0.0073 | 20.0 | 4500 | 0.7686 | 0.8669 |
| 0.0003 | 21.0 | 4725 | 0.8099 | 0.8669 |
| 0.0003 | 22.0 | 4950 | 0.8291 | 0.8686 |
| 0.002 | 23.0 | 5175 | 0.8321 | 0.8669 |
| 0.0005 | 24.0 | 5400 | 0.8587 | 0.8669 |
| 0.0136 | 25.0 | 5625 | 0.8652 | 0.8702 |
| 0.0002 | 26.0 | 5850 | 0.8692 | 0.8702 |
| 0.0001 | 27.0 | 6075 | 0.8870 | 0.8702 |
| 0.0001 | 28.0 | 6300 | 0.8976 | 0.8686 |
| 0.0057 | 29.0 | 6525 | 0.9057 | 0.8719 |
| 0.0181 | 30.0 | 6750 | 0.9160 | 0.8719 |
| 0.0001 | 31.0 | 6975 | 0.9098 | 0.8702 |
| 0.0001 | 32.0 | 7200 | 0.9072 | 0.8719 |
| 0.0001 | 33.0 | 7425 | 0.9201 | 0.8686 |
| 0.0001 | 34.0 | 7650 | 0.9257 | 0.8752 |
| 0.0 | 35.0 | 7875 | 0.9399 | 0.8702 |
| 0.0059 | 36.0 | 8100 | 0.9356 | 0.8686 |
| 0.0 | 37.0 | 8325 | 0.9495 | 0.8719 |
| 0.0 | 38.0 | 8550 | 0.9431 | 0.8702 |
| 0.0039 | 39.0 | 8775 | 0.9538 | 0.8702 |
| 0.0 | 40.0 | 9000 | 0.9587 | 0.8686 |
| 0.0 | 41.0 | 9225 | 0.9526 | 0.8702 |
| 0.0 | 42.0 | 9450 | 0.9618 | 0.8686 |
| 0.0 | 43.0 | 9675 | 0.9656 | 0.8719 |
| 0.0 | 44.0 | 9900 | 0.9609 | 0.8719 |
| 0.0001 | 45.0 | 10125 | 0.9585 | 0.8719 |
| 0.0 | 46.0 | 10350 | 0.9641 | 0.8702 |
| 0.0 | 47.0 | 10575 | 0.9675 | 0.8686 |
| 0.0 | 48.0 | 10800 | 0.9660 | 0.8719 |
| 0.0045 | 49.0 | 11025 | 0.9652 | 0.8719 |
| 0.0044 | 50.0 | 11250 | 0.9651 | 0.8719 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2