foods
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the food101 dataset. It achieves the following results on the evaluation set:
- Loss: 0.4850
- Accuracy: 0.925
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 | Validation Loss | Accuracy |
|---|---|---|---|---|
| 3.215 | 0.99 | 62 | 2.9381 | 0.778 |
| 1.7683 | 2.0 | 125 | 1.6041 | 0.911 |
| 1.2081 | 2.99 | 187 | 1.1491 | 0.894 |
| 0.82 | 4.0 | 250 | 0.9028 | 0.899 |
| 0.7188 | 4.99 | 312 | 0.7217 | 0.913 |
| 0.5186 | 6.0 | 375 | 0.5988 | 0.928 |
| 0.4582 | 6.99 | 437 | 0.5468 | 0.926 |
| 0.4185 | 8.0 | 500 | 0.4943 | 0.93 |
| 0.3909 | 8.99 | 562 | 0.4865 | 0.925 |
| 0.3513 | 9.92 | 620 | 0.4850 | 0.925 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.0.1+cu118
- Datasets 2.13.0
- Tokenizers 0.13.3