Add factor-based and independent per-mesh-dim ILP sharding optimizers#320

fmassa wants to merge 22 commits intomainmeta-pytorch/autoparallel:mainfrom

fmassa/factor_graphmeta-pytorch/autoparallel:fmassa/factor_graphCopy head branch name to clipboard