vit-base-patch16-224-ve-U11-12
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: 0.5924
- Accuracy: 0.8478
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: 5.5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.05
- num_epochs: 12
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.3668 | 0.96 | 16 | 1.2319 | 0.5652 |
| 1.1102 | 1.97 | 33 | 0.9996 | 0.6957 |
| 0.8257 | 2.99 | 50 | 0.8429 | 0.6304 |
| 0.68 | 4.0 | 67 | 0.6906 | 0.8043 |
| 0.4763 | 4.96 | 83 | 0.6871 | 0.7609 |
| 0.341 | 5.97 | 100 | 0.5924 | 0.8478 |
| 0.2956 | 6.99 | 117 | 0.4863 | 0.8478 |
| 0.2376 | 8.0 | 134 | 0.5947 | 0.7826 |
| 0.2098 | 8.96 | 150 | 0.5579 | 0.8043 |
| 0.2213 | 9.97 | 167 | 0.6474 | 0.7609 |
| 0.1767 | 10.99 | 184 | 0.6015 | 0.7826 |
| 0.1757 | 11.46 | 192 | 0.5928 | 0.7609 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2+cu118
- Datasets 2.16.1
- Tokenizers 0.15.0