Skip to content
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

torchvision 0.12.0 #45

Merged
merged 9 commits into from
Mar 21, 2022
Merged

torchvision 0.12.0 #45

merged 9 commits into from
Mar 21, 2022

Conversation

h-vetinari
Copy link
Member

@h-vetinari h-vetinari commented Mar 11, 2022

The update bot choked on something, and indeed, my local attempts to rerender hang indefinitely, even though I tried a couple tweaks in an attempt at mitigation.

Closes #51

@conda-forge-linter
Copy link

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

@h-vetinari
Copy link
Member Author

@conda-forge-admin, please rerender

@h-vetinari
Copy link
Member Author

h-vetinari commented Mar 12, 2022

Fails with a pytorch-related import:

  File "`[...]/lib/python3.8/site-packages/torchvision/models/quantization/mobilenetv2.py", line 5, in <module>
    from torch.ao.quantization import QuantStub, DeQuantStub
ImportError: cannot import name 'QuantStub' from 'torch.ao.quantization' 

@hmaarrfk
Copy link
Contributor

This likely needs pytorch 1.11.0

@h-vetinari
Copy link
Member Author

@conda-forge-admin, please rerender

@h-vetinari
Copy link
Member Author

Rebased on #49, WIP until that one is merged

@h-vetinari
Copy link
Member Author

AFAICT, all the errors are due to not raising an error upon use of what's supposed to be a private implementation:

        if "16" in img_path:
            # 16 bits image decoding is supported, but only as a private API
            # FIXME: see https://github.com/pytorch/vision/issues/4731 for potential solutions to making it public
>           with pytest.raises(RuntimeError, match="At most 8-bit PNG images are supported"):
E           Failed: DID NOT RAISE <class 'RuntimeError'>

@hmaarrfk
Copy link
Contributor

Thanks for reaching out upstream. Curious they didn't hit the bug.

@h-vetinari
Copy link
Member Author

Well, unless you put a hash in the path, it's not that likely. In our case though, since torchvision-split_1647862604097 seems to be using an unix epoch integer, we're guaranteed to hit it, as it's in the two most significant bits...

>>> pd.Series([pd.Timestamp.now()]).astype(np.int64)[0] // 10 ** 6
1647903504531

@hmaarrfk
Copy link
Contributor

I guess we need to wait 2000 years to release it! 🤣

@h-vetinari h-vetinari mentioned this pull request Mar 21, 2022
@h-vetinari
Copy link
Member Author

@conda-forge/torchvision
This is ready, PTAL. Since there are no controversial changes (to my mind) I'll probably merge in 10-12h if there are no comments.

@h-vetinari h-vetinari merged commit da433bf into conda-forge:main Mar 21, 2022
@h-vetinari h-vetinari deleted the bump branch March 21, 2022 23:08
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants