PyTorch implementation of "Interpretable Medical Deep Framework by Logits-constraint Attention Guiding Graph-based Multi-scale Fusion for Alzheimer's Disease Analysis"

• CUDA Version 12.2    
• torch 1.13.1+cu116
• torch-cluster 1.6.1+pt113cu116
• torch-geometric 2.1.0
• torch-scatter 2.1.1+pt113cu116
• torch-sparse 0.6.17+pt113cu116
• torch-spline-conv 1.2.2+pt113cu116
• torchaudio 0.13.1+cu116
• torchvision 0.14.1+cu116
• SimpleITK 2.2.1
• numpy 1.25.2
• Python 3.9.17
• pandas 2.0.3
• scikit-learn 1.3.0

🐣For the hyperparameter λ(τ), we recommend delaying its assignment when using a smaller training set, and advancing its assignment when using a larger training set. For example, when only using ADNI1 for training, you can set λ=0(τ≤30)/λ=0.2(τ>30). When using ADNI1+ADNI2+ADNI3 for training, you can set λ=0(τ≤10)/λ=0.2(τ>10).

python main.py --device [GPU-id]