vit-base-25ep
This model is a fine-tuned version of google/vit-base-patch16-224 on the vuongnhathien/30VNFoods dataset. It achieves the following results on the evaluation set:
- Loss: 0.5506
- Accuracy: 0.8486
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: 0.0003
- train_batch_size: 64
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 25
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.6167 | 1.0 | 275 | 0.5712 | 0.8354 |
| 0.3183 | 2.0 | 550 | 0.5564 | 0.8406 |
| 0.1729 | 3.0 | 825 | 0.5955 | 0.8433 |
| 0.139 | 4.0 | 1100 | 0.6453 | 0.8406 |
| 0.0775 | 5.0 | 1375 | 0.6044 | 0.8517 |
| 0.0784 | 6.0 | 1650 | 0.7265 | 0.8414 |
| 0.0502 | 7.0 | 1925 | 0.6977 | 0.8533 |
| 0.0525 | 8.0 | 2200 | 0.7100 | 0.8549 |
| 0.0311 | 9.0 | 2475 | 0.7423 | 0.8525 |
| 0.026 | 10.0 | 2750 | 0.7901 | 0.8461 |
| 0.0183 | 11.0 | 3025 | 0.7261 | 0.8592 |
| 0.0218 | 12.0 | 3300 | 0.8014 | 0.8485 |
| 0.0135 | 13.0 | 3575 | 0.7391 | 0.8584 |
| 0.0066 | 14.0 | 3850 | 0.6938 | 0.8740 |
| 0.0047 | 15.0 | 4125 | 0.6765 | 0.8815 |
| 0.0052 | 16.0 | 4400 | 0.6611 | 0.8839 |
| 0.0033 | 17.0 | 4675 | 0.6794 | 0.8803 |
| 0.0037 | 18.0 | 4950 | 0.6724 | 0.8811 |
| 0.0026 | 19.0 | 5225 | 0.6759 | 0.8875 |
| 0.0031 | 20.0 | 5500 | 0.6699 | 0.8855 |
| 0.0028 | 21.0 | 5775 | 0.6720 | 0.8847 |
| 0.0029 | 22.0 | 6050 | 0.6746 | 0.8843 |
| 0.0016 | 23.0 | 6325 | 0.6731 | 0.8859 |
| 0.0016 | 24.0 | 6600 | 0.6759 | 0.8859 |
| 0.0019 | 25.0 | 6875 | 0.6767 | 0.8847 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2