Fix DefaultBoxGenerator CUDA device mismatch in _grid_default_boxes (#9414)

[Nueramarcos]

Summary

Fixes #9414

DefaultBoxGenerator._grid_default_boxes created shifts_x/shifts_y via torch.arange with no device= argument, and accessed self._wh_pairs[k] (always CPU from __init__), causing a CPU vs CUDA mismatch at torch.cat when the model runs on GPU.

Changes

• Added device parameter to _grid_default_boxes (defaults to cpu for backwards compatibility)
• Pass device=device to torch.arange for the shift tensors
• Call .to(dtype=dtype, device=device) on _wh_pairs[k] before use
• Updated forward() to pass device through and removed the now-redundant default_boxes.to(device) call

Testing

Added test/test_default_box_generator.py covering:

• CPU: output shape, device, dtype, no NaNs, clip=True/False, batch size > 1, TorchScript traceability
• CUDA (skipped when unavailable): core regression test confirming torch.cat no longer raises, CPU→CUDA value equality, batch, clip=False, TorchScript on CUDA, _wh_pairs device verification

Verified CPU path produces 8732 default boxes for SSD-300 with no NaNs.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/9514

• 📄 Preview Python docs built from this PR

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

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

• [ROCm] MI350 CI jobs will have longer queue times due to CI migration

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

[meta-cla]

Hi @Nueramarcos!

Thank you for your pull request and welcome to our community.

Action Required

In order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks!