beit-base-patch16-224-pt22k-ft22k-finetuned-ind-17-imbalanced-aadhaarmask
This model is a fine-tuned version of microsoft/beit-base-patch16-224-pt22k-ft22k on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.3480
- Accuracy: 0.8450
- Recall: 0.8450
- F1: 0.8442
- Precision: 0.8494
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- 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 | Recall | F1 | Precision |
---|---|---|---|---|---|---|---|
0.5859 | 0.9974 | 293 | 0.6117 | 0.8114 | 0.8114 | 0.7891 | 0.8139 |
0.5281 | 1.9983 | 587 | 0.4362 | 0.8442 | 0.8442 | 0.8375 | 0.8484 |
0.4214 | 2.9991 | 881 | 0.4228 | 0.8438 | 0.8438 | 0.8392 | 0.8529 |
0.4221 | 4.0 | 1175 | 0.4121 | 0.8382 | 0.8382 | 0.8331 | 0.8495 |
0.4127 | 4.9974 | 1468 | 0.3692 | 0.8476 | 0.8476 | 0.8454 | 0.8511 |
0.3122 | 5.9983 | 1762 | 0.3741 | 0.8408 | 0.8408 | 0.8394 | 0.8462 |
0.3079 | 6.9991 | 2056 | 0.3628 | 0.8429 | 0.8429 | 0.8403 | 0.8445 |
0.2851 | 8.0 | 2350 | 0.3635 | 0.8412 | 0.8412 | 0.8389 | 0.8412 |
0.297 | 8.9974 | 2643 | 0.3407 | 0.8510 | 0.8510 | 0.8497 | 0.8545 |
0.2109 | 9.9745 | 2930 | 0.3566 | 0.8421 | 0.8421 | 0.8406 | 0.8418 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.2.0a0+81ea7a4
- Datasets 2.19.0
- Tokenizers 0.19.1
- Downloads last month
- 2
Inference API (serverless) is not available, repository is disabled.
Model tree for Kushagra07/beit-base-patch16-224-pt22k-ft22k-finetuned-ind-17-imbalanced-aadhaarmask
Base model
microsoft/beit-base-patch16-224-pt22k-ft22k
Finetuned
this model
Evaluation results
- Accuracy on imagefolderself-reported0.845
- Recall on imagefolderself-reported0.845
- F1 on imagefolderself-reported0.844
- Precision on imagefolderself-reported0.849