[Feature Request]: add a FakeData set so that users can test the pipeline without depending on XNAT

**Description**
*Brief description of the new feature*
https://docs.pytorch.org/vision/main/generated/torchvision.datasets.FakeData.html, implement this such that users in debug mode can by-pass XNAT data import to test ML pipeline. Also could allow users to import data from sources other than XNAT.

**Why do you need the new feature?**
Make the template more adaptable and allow iterative testing.