vit-Covid
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.0805
- Accuracy: 0.9847
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: 10
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
0.1283 | 0.38 | 100 | 0.1878 | 0.9484 |
0.0312 | 0.76 | 200 | 0.1484 | 0.9560 |
0.0655 | 1.15 | 300 | 0.0976 | 0.9713 |
0.0587 | 1.53 | 400 | 0.0887 | 0.9713 |
0.0106 | 1.91 | 500 | 0.0980 | 0.9732 |
0.0137 | 2.29 | 600 | 0.1479 | 0.9618 |
0.07 | 2.67 | 700 | 0.0882 | 0.9751 |
0.0068 | 3.05 | 800 | 0.1160 | 0.9675 |
0.0321 | 3.44 | 900 | 0.0872 | 0.9694 |
0.0027 | 3.82 | 1000 | 0.0790 | 0.9809 |
0.0041 | 4.2 | 1100 | 0.1029 | 0.9713 |
0.0014 | 4.58 | 1200 | 0.0947 | 0.9809 |
0.0018 | 4.96 | 1300 | 0.1399 | 0.9713 |
0.001 | 5.34 | 1400 | 0.0689 | 0.9847 |
0.001 | 5.73 | 1500 | 0.0852 | 0.9790 |
0.0008 | 6.11 | 1600 | 0.1111 | 0.9790 |
0.0013 | 6.49 | 1700 | 0.0695 | 0.9866 |
0.0049 | 6.87 | 1800 | 0.0728 | 0.9885 |
0.0007 | 7.25 | 1900 | 0.0963 | 0.9790 |
0.0012 | 7.63 | 2000 | 0.0886 | 0.9847 |
0.0006 | 8.02 | 2100 | 0.0811 | 0.9847 |
0.0015 | 8.4 | 2200 | 0.0796 | 0.9847 |
0.0143 | 8.78 | 2300 | 0.0804 | 0.9847 |
0.0005 | 9.16 | 2400 | 0.0816 | 0.9847 |
0.0006 | 9.54 | 2500 | 0.0811 | 0.9847 |
0.0005 | 9.92 | 2600 | 0.0805 | 0.9847 |
Framework versions
- Transformers 4.38.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1
- Downloads last month
- 19