nyumedml / cnn_design_for_ad Goto Github PK

Hi, thanks for the great project. I am trying to reproduce the results. I have downloaded ADNI, and used .sh files to preprocess the dataset following ReadMe instructions. The model trained by config.yaml of the repo, I got 60% accuracy compared to 68% reported in the model evaluation notebook. Have you used another config.yaml file?

hi,
i just wanna test your best model, and where can i find "dir_to_scans = 'dir/to/ADNI_processed/subjects' "
did i need to process template data use your offered here “https://drive.google.com/file/d/1KurgyjQP-KReEO0gf31xxjwE5R-xuSRB/view” ？

2 poolings in models.py (line 92 and line 102) have same names. This means the last 5x5 pooling doesn't work at all. Please check your code.