kayiwarahim/Retinal_LVM

Refactor disease mapping and prediction logic in app.py for improved clarity and accuracy

Update disease_classes in app.py with specific retinal disease labels for improved clarity

Refactor app.py for improved readability and structure; enhance prediction function output format

Enhance SparseTopKAttention to safely handle top-k selection and prevent out-of-range errors

Fix region projection layer and adjust visual embedding extraction in SceneGraphTransformer

Refactor MultiResolutionEncoder to enhance feature extraction and output normalization

Refactor SceneGraphTransformer and MultiResolutionEncoder for improved feature extraction and uncertainty estimation

Fix model checkpoint URL in load_model function

Enhance SceneGraphTransformer with multi-resolution encoding, sparse attention, and uncertainty estimation

Implement SceneGraphTransformer with multi-resolution feature extraction, ensemble branches, and uncertainty estimation

Initial project setup with Gradio integration for retinal disease detection model

initial commit