Yes, it is very common. I work a lot with ML code released as part of research papers. Everyone uses a different version of TF/PyTorch or whatever framework the

Yes, it is very common. I work a lot with ML code released as part of research papers. Everyone uses a different version of TF/PyTorch or whatever framework they are using. I can't imagine working without something like virtualenv (though I use pipenv)