5 ms·
Getting the right version of PyTorch installed to have the correct kind of acceleration on each different platform you support has been a long-standing headache
by shawnz 2y ago
Getting the right version of PyTorch installed to have the correct kind of acceleration on each different platform you support has been a long-standing headache across many Python dependency management tools, not just uv. For example, here's the bug in poetry regarding this issue: https://github.com/python-poetry/poetry/issues/6409 https://github.com/python-poetry/poetry/issues/6409
As I understand it, recent versions of PyTorch have made this process somewhat easier, so maybe it's worth another try.
- amelius 2y agoOn nvidia jetson systems, I always end up compiling torchvision, while torch always comes as a wheel. It seems so random.
- tmaly 2y agoI end up going to the torch website and they have a nice little UI I can click what I have and it gives me the pip line to use.
- shawnz 2y agoThat's fine if you are just trying to get it running on your machine specifically, but the problems come in when you want to support multiple different combinations of OS and compute platform in your project.
- tmaly 2y agoI could see this information on the website being encoded in some form in pypi such that it could be updated to support various platforms.
- fluidcruft 2y agouv actually handles thr issues described there very well (uv docs have have a page showing a few ways to do it). The issue for me is uv has massive amnesia about which one was selected and you end up trashing packages because of that. uv is very fast at thrashing though so it's not as bad as if poetry were thrashing.