Jesse Cai

Thanks @alexsamardzic! @cpuhrsch I just got access to the cusparselt 0.4.0 dev build and plan to run some microbenchmarks. Will update you guys when I have numbers.

closing b/c this is out of sync

Can you try updating torchao? I don't think the top level `quantize_` API is available in 0.1 But from what I understand, torch.compile() does not work on windows because triton...

Closing as we've removed all m1 specific wheels last february cc @joecummings

cc @lakshminarasimmanv just curious here - are you working on this for fun or do you have a use-case in mind that needs this? If you have a specific use...

closing as this is now part of @jainapurva work on benchmarking

cc @EnragedAntelope Thanks for flagging, could you open a PR to make those changes? Feel free to tag me for review :)

@jerryzh168 All I did was run pytest/test and run the API examples mentioned in the doc, so it's entirely possible that this slipped through if there wasn't a test. We...

cc @eqy I think we can land this instead of https://github.com/pytorch/pytorch/pull/134626, does that sound good to you?

@pytorchbot rebase