ed-cho

Probably too late but that's because BasicLayer does SwinTransformerBlock and then PatchMerging (which downsamples), unlike the diagram in the original paper. In other words, the paper suggests stage 1 (linear...

It would be really nice to be able to use it on Mac M1. @tilmantroester Would it be possible for you to create a PR please when you are free?

same here... any updates on this?