vit-base-patch16-224-U6-10
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: 0.5473
- Accuracy: 0.8333
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: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.3653 | 1.0 | 16 | 1.2199 | 0.6333 |
| 1.0932 | 2.0 | 32 | 1.0086 | 0.65 |
| 0.9284 | 3.0 | 48 | 0.8466 | 0.6667 |
| 0.6745 | 4.0 | 64 | 0.8237 | 0.7 |
| 0.4775 | 5.0 | 80 | 0.7473 | 0.7667 |
| 0.4194 | 6.0 | 96 | 0.6148 | 0.7833 |
| 0.3043 | 7.0 | 112 | 0.6221 | 0.8167 |
| 0.2947 | 8.0 | 128 | 0.6156 | 0.7667 |
| 0.269 | 9.0 | 144 | 0.5700 | 0.8167 |
| 0.2261 | 10.0 | 160 | 0.5473 | 0.8333 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2+cu118
- Datasets 2.16.1
- Tokenizers 0.15.0