-
-
Notifications
You must be signed in to change notification settings - Fork 2k
More flexible dependencies, please #181
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Comments
Let me I add my experience to this issue. I'm on Manjaro and installed python-pytorch 1.10.1-1. When I try to
Please advice on how to proceed. Thanks, and many kudos on this project! |
torch dependency was removed in the next release. |
I still have errors. Just FYI. Troubleshooting (I'm no expert in Python), I first installed numpy using pip. No errors. I'm sorry but I cannot explain what's wrong. Then, I did two kind of installations. One as a regular user and the other as root. Hope this helps. As regular user:
As root:
|
The last issue is coming from the fact that Manjaro uses python 3.10 and there is currently no onnxruntime package available. And for more information on the advancement you can check that pr: microsoft/onnxruntime#10467 Well, since there is no wheel for python 3.10 you can either wait for it, downgrade python to 3.9 or use a virtualenv to get it (this is a way better option, playing with systemwide python is not a good idea). Or you can build onnxruntime 1.10.0 wheel from source (what i did). Note that there is a package for it on AUR but i didin't even bother trying it since installing systemwide python deps from AUR is one of the few things that can break a linux install... |
@Andrflor Thank you very much for a thorough explanation. I was not aware that Manjaro uses a different python version from other distros. I will follow your kind advise and example! Thank you again! |
I'm trying to utilize this library in a script with urllib3 version that is not compatible with rembg version because the compatibility is too limited. Besides that, I'm also not able to run it on a RTX 3060 because the accepted torch version doesn't support newer gpus.
Theres a way to flexibilize the requirements so it can run in newer versions of at least torch? As at least in theory torch versions are backward compatible, I think that it would not be a problem.
The text was updated successfully, but these errors were encountered: