**Is your feature request related to a problem? Please describe.** clean up the research code and port into MONAI - identify generic image-based attention modules - port full network implementations - add workflow tutorial https://arxiv.org/abs/2103.10504