Logan Adams
Logan Adams
@jithunnair-amd - there do not appear to be any other replacement or equivalents for QTorch that you've seen?
@oraluben - sorry this PR has taken so long to be merged, I think it just needed to have master merged again to get the XPU fixes.
@crcrpar - I made one [here](https://github.com/NVIDIA/apex/pull/1680) if you could take a look.
@Colezwhy and @crcrpar - when I build with `--no-build-isolation` I don't get the packaging error, but neither the cpp nor cuda_ext are installed. And no warnings for reasons why they...
@crcrpar - even with the newer pip the issue doesn't seem to be multiple config settings either. When you tested this did you hit any issues with this?
Adding torch to the dependencies does help, but the bigger issue seems to be that no matter who I specify the build options, they're not being picked up as being...
@xwang233 - were you able to build the cpp or cuda exts with these commands at all? I've been able to build, but I'm not seeing the arguments passed to...
@xwang233 - Are you using the most up to date master branch from the repo? Since you should also need to add packaging and torch as dependencies to the pyproject.toml...
Interesting, mine is torch 1.13, but I doubt the torch version makes a difference. Also in a venv if that matters. I'm just not able to see it ever get...
@RuABraun - that's what many of us are doing, but that will prevent future changes from being taken. If you add pytorch and packaging, that will resolve those issues but...