This project implements an hearing loss classification model using the Vision Transformer (ViT) architecture. The model is trained and evaluated on our custom dataset,
dataset
: Directory containing training and testing data.models
: Directory where trained model checkpoints are saved.main.py
: Full Fine Tuning the ViT baseline model(google/vit-base-patch16-224
)
- Training loss for each epoch
- Test accuracy for each epoch
- Accuracy graph
- Confusion matrix
- torch
- torchvision
- transformers
- scikit-learn
- matplotlib