Data augmentation

Hello! Would it be possible for you to share more details on the data augmentation performed on the training data?
Thank you

I updated #10 with a README FAQ answer for this. The repo's prediction path slices preprocessed NIfTI inputs into 80 JPG files and runs inference with the .h5 model; it does not ship a training or data-augmentation pipeline.