fix resize_mask ignoring NEAREST_EXACT interpolation

[Mr-Neutr0n]

Fixes #9188

resize_mask() was hardcoding InterpolationMode.NEAREST and ignoring the user-specified interpolation mode entirely. This meant NEAREST_EXACT (which is the more correct interpolation — matching PIL and scikit-image behavior) got silently dropped when resizing a tv_tensors.Mask.

The fix adds an interpolation parameter to resize_mask() and passes it through from the dispatch function. Only NEAREST and NEAREST_EXACT are accepted — anything else (bilinear, bicubic, etc.) falls back to NEAREST since those don't make sense for masks.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/9384

• 📄 Preview Python docs built from this PR

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

❌ 6 New Failures

As of commit 177f66e with merge base 0f6d91d ():

NEW FAILURES - The following jobs have failed:

• Tests / unittests-linux (3.10, linux.12xlarge, cpu) / linux-job (gh)
  test/test_transforms_v2.py::TestResizedCrop::test_kernel[resized_crop_mask-make_detection_masks]
• Tests / unittests-linux (3.14, linux.12xlarge, cpu) / linux-job (gh)
  test/test_transforms_v2.py::TestResizedCrop::test_kernel[resized_crop_mask-make_detection_masks]
• Tests / unittests-macos (3.10, macos-m1-stable) / macos-job (gh)
  test/test_transforms_v2.py::TestResizedCrop::test_kernel[resized_crop_mask-make_detection_masks]
• Tests / unittests-macos (3.14, macos-m1-stable) / macos-job (gh)
  test/test_transforms_v2.py::TestResizedCrop::test_kernel[resized_crop_mask-make_detection_masks]
• Tests / unittests-windows (3.10, windows.4xlarge, cpu) / windows-job (gh)
  test/test_transforms_v2.py::TestResizedCrop::test_kernel[resized_crop_mask-make_detection_masks]
• Tests / unittests-windows (3.14, windows.4xlarge, cpu) / windows-job (gh)
  test/test_transforms_v2.py::TestResizedCrop::test_kernel[resized_crop_mask-make_detection_masks]

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

[zy1git]

@Mr-Neutr0n Hi,

Thanks a lot for creating the PR. I am reviewing it. At the same time, could you please take a look at the test errors and address those?

[zy1git]

I took a look at the test errors and wanted to share a few observations:

a, the signature order does not match F.resize: resize_mask(mask, size, max_size, interpolation) should be resize_mask(mask, size, interpolation, max_size).
b, set literal {NEAREST, NEAREST_EXACT} is not compatible with TorchScript.
c, The default value (NEAREST) differs from F.resize (BILINEAR) in the test file, which causes test_functional_signature to fail.

If you have any questions while addressing the CI test failures, please feel free to reach out.