Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Conversation

GdoongMathew
Copy link
Contributor

Partially fix #8762

Copy link

pytorch-bot bot commented Feb 20, 2025

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/8922

Note: Links to docs will display an error until the docs builds have been completed.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

target_transform (callable, optional): A function/transform that takes in the target and transforms it.
download (bool, optional): If True, downloads the dataset from the internet and puts it into
``root/country211/``. If dataset is already downloaded, it is not downloaded again.
loader (callable, optional): A function to load an image given its path.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the PR @GdoongMathew ,

Here and below, we should specify that the default is to use PIL, but that we encourage users to try to use decode_image()

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

added. I also updated the one in ImageNet dataset as well.

transform=transform,
target_transform=target_transform,
loader=loader,
)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

How do you feel about writing a basic test for this? Hopefully this can fit in <10 lines of code and we can re-use the same test across most datasets?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added. Hopefully, for the rest of the test classes, one could update their SUPPORT_TV_IMAGE_DECODE attribute whenever they start to support loader argument.

return image

with unittest.mock.patch("PIL.Image.open", new=new):
with unittest.mock.patch(open.__module__ + "." + open.__qualname__, new=new):
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Damn, I hope we never have to change / fix this ever lol.

Thank you so much for pushing through the test @GdoongMathew

@NicolasHug NicolasHug merged commit 86f8eb0 into pytorch:main Feb 24, 2025
1 check passed
@GdoongMathew GdoongMathew deleted the feat/loader_api branch February 24, 2025 13:24
facebook-github-bot pushed a commit that referenced this pull request Jul 28, 2025
Summary: Co-authored-by: Nicolas Hug <[email protected]>

Reviewed By: scotts

Differential Revision: D77997073

fbshipit-source-id: 8de75ef7bc07f0b665b423eae800cca679677039
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Feature Request] Datasets Should Use New torchvision.io Image Loader APIs and Return TVTensor Images by Default

3 participants