histo_train_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: 0.7340
- Accuracy: 0.825
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: 64
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.0481 | 1.67 | 10 | 0.4926 | 0.825 |
| 0.3714 | 3.33 | 20 | 0.3388 | 0.9 |
| 0.0642 | 5.0 | 30 | 0.3255 | 0.875 |
| 0.0199 | 6.67 | 40 | 0.4111 | 0.875 |
| 0.0074 | 8.33 | 50 | 0.3334 | 0.925 |
| 0.0024 | 10.0 | 60 | 0.3710 | 0.9 |
| 0.0131 | 11.67 | 70 | 0.5366 | 0.85 |
| 0.0067 | 13.33 | 80 | 0.5172 | 0.875 |
| 0.0152 | 15.0 | 90 | 0.4835 | 0.9 |
| 0.0058 | 16.67 | 100 | 0.3979 | 0.875 |
| 0.0005 | 18.33 | 110 | 0.5964 | 0.825 |
| 0.0008 | 20.0 | 120 | 0.7340 | 0.825 |
Framework versions
- Transformers 4.27.3
- Pytorch 1.13.1+cu116
- Datasets 2.10.1
- Tokenizers 0.13.2