results
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: 1.5746
- Accuracy: 0.894
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: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 2.6455 | 0.99 | 62 | 2.4448 | 0.841 |
| 1.751 | 2.0 | 125 | 1.7416 | 0.879 |
| 1.5345 | 2.98 | 186 | 1.5746 | 0.894 |
Framework versions
- Transformers 4.28.1
- Pytorch 1.12.1+cu116
- Datasets 2.4.0
- Tokenizers 0.12.1