vit_itri_gerd
This model is a fine-tuned version of google/vit-large-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.8160
- Accuracy: 0.8802
- Precision: 0.8811
- Recall: 0.8802
- F1: 0.8802
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: 24
- eval_batch_size: 4
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 15
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| 0.6915 | 1.0 | 63 | 0.4305 | 0.7904 | 0.7926 | 0.7904 | 0.7901 |
| 0.455 | 2.0 | 126 | 0.7307 | 0.7605 | 0.7836 | 0.7605 | 0.7552 |
| 0.372 | 3.0 | 189 | 0.4026 | 0.8024 | 0.8123 | 0.8024 | 0.8007 |
| 0.3159 | 4.0 | 252 | 0.3805 | 0.8323 | 0.8340 | 0.8323 | 0.8321 |
| 0.2906 | 5.0 | 315 | 0.4334 | 0.8323 | 0.8326 | 0.8323 | 0.8323 |
| 0.2589 | 6.0 | 378 | 0.4235 | 0.8084 | 0.8232 | 0.8084 | 0.8060 |
| 0.2024 | 7.0 | 441 | 0.4003 | 0.8503 | 0.8516 | 0.8503 | 0.8502 |
| 0.1218 | 8.0 | 504 | 0.6308 | 0.8204 | 0.8270 | 0.8204 | 0.8193 |
| 0.1226 | 9.0 | 567 | 0.5468 | 0.8323 | 0.8353 | 0.8323 | 0.8319 |
| 0.0627 | 10.0 | 630 | 0.7390 | 0.8263 | 0.8286 | 0.8263 | 0.8260 |
| 0.0374 | 11.0 | 693 | 0.8669 | 0.8503 | 0.8503 | 0.8503 | 0.8503 |
| 0.0389 | 12.0 | 756 | 0.6790 | 0.8623 | 0.8627 | 0.8623 | 0.8622 |
| 0.0122 | 13.0 | 819 | 0.8346 | 0.8683 | 0.8701 | 0.8683 | 0.8681 |
| 0.0064 | 14.0 | 882 | 0.7985 | 0.8802 | 0.8804 | 0.8802 | 0.8802 |
| 0.0071 | 15.0 | 945 | 0.8160 | 0.8802 | 0.8811 | 0.8802 | 0.8802 |
Framework versions
- Transformers 4.53.0.dev0
- Pytorch 2.7.1+cu126
- Datasets 3.6.0
- Tokenizers 0.21.1