ai_art_exp1_vit_final
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on an unknown dataset. It achieves the following results on the evaluation set:
- Accuracy: {'accuracy': 0.9946666666666667}
- Overall Accuracy: 0.9947
- Loss: 0.0231
- Human Accuracy: 0.99
- Ld Accuracy: 0.998
- Sd Accuracy: 0.996
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: 5e-05
- 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 | Accuracy | Overall Accuracy | Validation Loss | Human Accuracy | Ld Accuracy | Sd Accuracy |
|---|---|---|---|---|---|---|---|---|
| 0.198 | 0.992 | 93 | {'accuracy': 0.9506666666666667} | 0.9507 | 0.1906 | 0.8548 | 0.9981 | 0.9959 |
| 0.0647 | 1.9947 | 187 | {'accuracy': 0.9793333333333333} | 0.9793 | 0.0811 | 0.9489 | 0.9923 | 0.9959 |
| 0.0395 | 2.9973 | 281 | {'accuracy': 0.988} | 0.988 | 0.0567 | 0.9734 | 0.9904 | 1.0 |
| 0.069 | 4.0 | 375 | {'accuracy': 0.9933333333333333} | 0.9933 | 0.0399 | 0.9816 | 1.0 | 0.9980 |
| 0.0456 | 4.992 | 468 | {'accuracy': 0.9946666666666667} | 0.9947 | 0.0309 | 0.9877 | 1.0 | 0.9959 |
| 0.0324 | 5.9947 | 562 | {'accuracy': 0.9906666666666667} | 0.9907 | 0.0444 | 0.9734 | 1.0 | 0.9980 |
| 0.0136 | 6.9973 | 656 | {'accuracy': 0.996} | 0.996 | 0.0234 | 0.9939 | 1.0 | 0.9939 |
| 0.0137 | 8.0 | 750 | {'accuracy': 0.9953333333333333} | 0.9953 | 0.0218 | 0.9898 | 0.9962 | 1.0 |
| 0.0105 | 8.992 | 843 | {'accuracy': 0.9953333333333333} | 0.9953 | 0.0222 | 0.9877 | 1.0 | 0.9980 |
| 0.0111 | 9.92 | 930 | {'accuracy': 0.9986666666666667} | 0.9987 | 0.0122 | 0.9980 | 0.9981 | 1.0 |
Framework versions
- Transformers 4.41.0
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1