-
-
Notifications
You must be signed in to change notification settings - Fork 5.1k
Test Python 3.11 support #1830
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
Test Python 3.11 support #1830
Conversation
| python: ['3.10'] | ||
| torch: ['1.13.0'] | ||
| torchvision: ['0.14.0'] | ||
| python: ['3.11'] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Might be better to test the full range of supported Python versions
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. |
|
Are these failures due to Python 3.11? It seems library-specific. |
|
@adamjstewart no, it's due to PyTorch 2.0 actually, and change in optimizer interface for class hierarchy. It's not hard to fix, a bit ugly, have it had coded in my cross platform win/osx test machines, but a reason I haven't switched to 2.0 in the tests for github runners yet... |
|
@adamjstewart fyi, the only compat issue is the 'Lookahead' optimizer, which is exercised in tests, but not commonly used so the surface area is minimal |
|
Thanks, I'll remove the Spack constraint I added. |
|
@adamjstewart I added the fix for optimizer on main |
|
@adamjstewart FYI, forgot to mention, there is a bug in torchscript w/ Python 3.11 ... I've reported it to torch team but not at all clear when they intend to fix it. No issues with Python 3.10 and nothing to do with timm. |
7231f6e to
1b0a5d5
Compare
|
Is Python 3.11 testing still held up by this torchscript bug? Wondering if I should just close this PR and let you handle testing once it's fixed. |
|
Going to close this PR for now but let me know when 3.11 is supported and I can reopen. |
We should probably test Python 3.11 in CI so that something like #1649 doesn't reoccur.