finetuned-indian-food
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the food_images dataset. It achieves the following results on the evaluation set:
- Loss: 0.2373
- Accuracy: 0.9394
- Precision: 0.9406
- Recall: 0.9410
- F1: 0.9401
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: 0.0002
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
---|---|---|---|---|---|---|---|
1.14 | 0.3 | 100 | 0.9144 | 0.8597 | 0.8756 | 0.8451 | 0.8349 |
0.8065 | 0.6 | 200 | 0.6736 | 0.8480 | 0.8735 | 0.8493 | 0.8473 |
0.5949 | 0.9 | 300 | 0.4965 | 0.8895 | 0.9046 | 0.8946 | 0.8941 |
0.4925 | 1.2 | 400 | 0.4049 | 0.9033 | 0.9084 | 0.9069 | 0.9051 |
0.3748 | 1.5 | 500 | 0.3852 | 0.9075 | 0.9129 | 0.9054 | 0.9065 |
0.3121 | 1.8 | 600 | 0.3422 | 0.9150 | 0.9183 | 0.9158 | 0.9148 |
0.3826 | 2.1 | 700 | 0.3406 | 0.9065 | 0.9137 | 0.9054 | 0.9076 |
0.2485 | 2.4 | 800 | 0.2915 | 0.9299 | 0.9298 | 0.9299 | 0.9282 |
0.2177 | 2.7 | 900 | 0.2520 | 0.9330 | 0.9333 | 0.9337 | 0.9327 |
0.1546 | 3.0 | 1000 | 0.2498 | 0.9384 | 0.9448 | 0.9385 | 0.9398 |
0.2157 | 3.3 | 1100 | 0.2658 | 0.9320 | 0.9329 | 0.9330 | 0.9318 |
0.0799 | 3.6 | 1200 | 0.2469 | 0.9384 | 0.9402 | 0.9394 | 0.9391 |
0.2242 | 3.9 | 1300 | 0.2373 | 0.9394 | 0.9406 | 0.9410 | 0.9401 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2
- Downloads last month
- 2