Add support for non-tensor inputs in AutoParallel#264

fmassa wants to merge 6 commits intomainmeta-pytorch/autoparallel:mainfrom

fmassa/support_non_tensor_inputsmeta-pytorch/autoparallel:fmassa/support_non_tensor_inputsCopy head branch name to clipboard