As an aside I suggest using dvc. It stores only metadata of large files in the got repository while maintain a separate data repository. The data repository can

As an aside I suggest using dvc. It stores only metadata of large files in the got repository while maintain a separate data repository. The data repository can also be synced with cloud storage like AWS. It really helps manage code better without raw data