Open bwasti opened 2 years ago
@bwasti where did you encounter that ? Code for main branch is a bit different: https://github.com/pytorch/functorch/blob/76178743084277cf6d7cac752279f905cdd60e13/functorch/_src/compilers.py#L24-L62
Yeah it's removed on main I believe - this was a hack we used to have to work around nvfuser limitations. But that's been fixed now.
Hey folks, stumbled into a CUDA assumption (on my non-CUDA machine)
Here's the fix for me, but it's obviously not very general