Closed thesofakillers closed 1 year ago
Hi, It's not really necessary, more a reference for people to know what version was tested and can be expected to work. If you can try the code with the latest PL and Pytorch, specifically if torch.compile works, I am happy to merge a PR 🙂
On Thu, Apr 13, 2023, 17:57 Giulio Starace @.***> wrote:
Hi, thanks a lot for this work.
I was wondering if the hard constraints ("==") to pytorch 1.13.1 and pytorch lightning 1.8.6 are necessary (link https://github.com/mees/calvin/blob/4b35b135259f66a2d1b2dc8d3d50875cc5396604/calvin_models/requirements.txt#L13 ).
Particularly for PyTorch 2.0, this is fully backwards compatible with 1.13 https://pytorch.org/get-started/pytorch-2.0/#pytorch-2x-faster-more-pythonic-and-as-dynamic-as-ever, and given the performance gains, it would be nice to add support for this by dropping the hard constraint. I can open a quick PR if that makes it easier for you.
— Reply to this email directly, view it on GitHub https://github.com/mees/calvin/issues/45, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGH2Z5WPGM3CRSOGUYCGZLXBAO47ANCNFSM6AAAAAAW5JAL3U . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Actually, I just remembered now, migrating to the latest PL is contingent on solving this issue upstream https://github.com/Lightning-AI/lightning/issues/16830
Hey, thanks for responding. I'll keep these notes (particular that about PL) in mind and open a PR if indeed I find PyTorch 2.0 is supported
Hi, thanks a lot for this work.
I was wondering if the hard constraints ("==") to pytorch 1.13.1 and pytorch lightning 1.8.6 are necessary (link).
Particularly for PyTorch 2.0, this is fully backwards compatible with 1.13, and given the performance gains, it would be nice to add support for this by dropping the hard constraint. I can open a quick PR if that makes it easier for you.