Open mathieuboudreau opened 11 months ago
@mathieuboudreau No particular reason honestly. I mentioned integrating nnUNet in ADS in the meetings but I guess I never created issues on the ADS repo. The endgoal is obviously to integrate these models in ADS but I feared it would take more time to merge into the main branch, and having a separate repo is simpler for the researchers for the moment.
Moreover, nnunet is a pretty big dependency. Considering ADS installation is already a bit long, I was not sure how we could deal with this.
Is there a reason why this need to be a completely separate repo, and not added as a feature (along with all the other ADS features) in our main software?