Main thing from the feedback was the confusion about what challenges ViT would face under a FL environment:
"It wasn't clear to me from the project description what challenges specifically FL algorithms would pose to vision transformers as opposed to any other type of DNN. That's something you might want to clarify and think about. Please also make sure you look up all the relevant prior specifically in this space that try to do the same thing."
Main thing from the feedback was the confusion about what challenges ViT would face under a FL environment:
"It wasn't clear to me from the project description what challenges specifically FL algorithms would pose to vision transformers as opposed to any other type of DNN. That's something you might want to clarify and think about. Please also make sure you look up all the relevant prior specifically in this space that try to do the same thing."