google-vit-base-patch16-224-face
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: 1.4531
- Accuracy: 0.7249
- Precision: 0.7172
- Recall: 0.7249
- F1: 0.7196
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: 0.00012
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 8
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| 0.8514 | 1.0 | 290 | 0.8464 | 0.7048 | 0.7035 | 0.7048 | 0.6909 |
| 0.7202 | 2.0 | 580 | 0.7791 | 0.7283 | 0.7297 | 0.7283 | 0.7111 |
| 0.5455 | 3.0 | 870 | 0.7950 | 0.7285 | 0.7174 | 0.7285 | 0.7171 |
| 0.334 | 4.0 | 1160 | 0.8948 | 0.7155 | 0.7152 | 0.7155 | 0.7145 |
| 0.1644 | 5.0 | 1450 | 1.0820 | 0.7239 | 0.7189 | 0.7239 | 0.7194 |
| 0.0482 | 6.0 | 1740 | 1.2792 | 0.7204 | 0.7144 | 0.7204 | 0.7160 |
| 0.0236 | 7.0 | 2030 | 1.4162 | 0.7279 | 0.7195 | 0.7279 | 0.7209 |
| 0.0049 | 8.0 | 2320 | 1.4531 | 0.7249 | 0.7172 | 0.7249 | 0.7196 |
Framework versions
- Transformers 4.24.0.dev0
- Pytorch 1.11.0+cu102
- Datasets 2.6.1
- Tokenizers 0.13.1