vit-base-patch16-224-high-vit
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: 0.6555
- Accuracy: 0.8421
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: 0.0002
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.8073 | 0.9787 | 23 | 1.4742 | 0.5211 |
| 0.9801 | 2.0 | 47 | 1.2410 | 0.5526 |
| 0.5808 | 2.9787 | 70 | 0.9728 | 0.7053 |
| 0.3797 | 4.0 | 94 | 0.7751 | 0.7632 |
| 0.2559 | 4.9787 | 117 | 0.8020 | 0.7684 |
| 0.1131 | 6.0 | 141 | 0.7116 | 0.8105 |
| 0.1207 | 6.9787 | 164 | 0.7258 | 0.8105 |
| 0.1068 | 8.0 | 188 | 0.6817 | 0.8316 |
| 0.0559 | 8.9787 | 211 | 0.6589 | 0.8368 |
| 0.0529 | 9.7872 | 230 | 0.6555 | 0.8421 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.2.1+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1