finetuned-skinpics
This model is a fine-tuned version of google/vit-base-patch16-224-in21K on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 1.2540
- Accuracy: 0.5139
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: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 7
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.121 | 0.57 | 100 | 1.1020 | 0.2569 |
| 1.0768 | 1.15 | 200 | 1.0546 | 0.4792 |
| 1.0532 | 1.72 | 300 | 1.0843 | 0.2917 |
| 1.0096 | 2.3 | 400 | 1.0693 | 0.4792 |
| 1.0716 | 2.87 | 500 | 1.0466 | 0.4931 |
| 1.0346 | 3.45 | 600 | 1.0225 | 0.5139 |
| 1.0232 | 4.02 | 700 | 1.0230 | 0.4931 |
| 0.8936 | 4.6 | 800 | 1.0582 | 0.5069 |
| 0.7125 | 5.17 | 900 | 1.0551 | 0.5139 |
| 0.6025 | 5.75 | 1000 | 1.1525 | 0.5278 |
| 0.4663 | 6.32 | 1100 | 1.2357 | 0.4653 |
| 0.5007 | 6.9 | 1200 | 1.2540 | 0.5139 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2