google/vit-base-patch16-224-in21k-v2-finetuned
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: 1.0612
- Accuracy: 0.7969
Model description
- Model type: Language model
- Language(s) (NLP): English
- License: Apache 2.0
- Related Model: google/vit-base-patch16-224-in21k
- Original Checkpoints: google/vit-base-patch16-224-in21k
- Resources for more information: Research paper
Intended uses & limitations
This model can be used to classify what type of food in the image provided.
Training and evaluation data
The model was trained on food101 dataset with 80:20 train-test-split.
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
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.9201 | 1.0 | 947 | 1.9632 | 0.7297 |
| 1.2002 | 2.0 | 1894 | 1.2327 | 0.7805 |
| 0.9561 | 3.0 | 2841 | 1.0612 | 0.7969 |
Framework versions
- Transformers 4.40.2
- Pytorch 2.4.0+cu121
- Datasets 2.21.0
- Tokenizers 0.19.1