sparseml icon indicating copy to clipboard operation
sparseml copied to clipboard

Lightning 2.0 support

Open lminer opened this issue 2 years ago • 3 comments

It would be nice to be able to use pytorch-lightning 2.0. We switch over to lightning 2.0 as soon as torch 2.0 was released, but would love to be able to try sparseml!

lminer avatar Jul 11 '23 10:07 lminer

Hi @lminer we currently support torch 2.0 installs with sparseml however the new torch.compile feature is not supported as it does not yet support hooks/modifications after compilation which is the core of what we do to sparsify torch models.

As for lighting support, an official integration is not yet on our roadmap, but if you are interested you should likely be able to integrate sparseml into pytorch lightning callbacks

bfineran avatar Jul 12 '23 20:07 bfineran

Hello @lminer Some time has passed between our last comment and hearing back. I hope the insight @bfineran has proven helpful. I'd like to go ahead and close out this thread. But if you would like to continue the conversation, please re-open this issue. Thank you so much! Jeannie / Neural Magic

jeanniefinks avatar Jul 26 '23 18:07 jeanniefinks

Could we leave it open. I'm guessing that at some point in time you'll add support for lightning 2.0, and it would be nice to have this issue as a record of when the change will be made.

lminer avatar Jul 26 '23 18:07 lminer