vit-base-3e-5-randaug
This model is a fine-tuned version of google/vit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.3568
- Accuracy: 0.9022
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: 3e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.6815 | 1.0 | 275 | 0.9075 | 0.7738 |
| 0.9759 | 2.0 | 550 | 0.5867 | 0.8501 |
| 0.7955 | 3.0 | 825 | 0.5191 | 0.8549 |
| 0.7056 | 4.0 | 1100 | 0.4548 | 0.8755 |
| 0.6455 | 5.0 | 1375 | 0.4256 | 0.8855 |
| 0.6249 | 6.0 | 1650 | 0.4114 | 0.8847 |
| 0.5742 | 7.0 | 1925 | 0.4026 | 0.8875 |
| 0.5782 | 8.0 | 2200 | 0.3943 | 0.8903 |
| 0.5383 | 9.0 | 2475 | 0.3929 | 0.8883 |
| 0.5495 | 10.0 | 2750 | 0.3921 | 0.8879 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2