add image_id to the default return for wrapped CocoDetection

[pmeier]

This is useful for debugging and is also used by our evaluation logic in the references:

vision/references/detection/engine.py

Line 100 in 5b07d6c

res = {target["image_id"].item(): output for target, output in zip(targets, outputs)}

Going from 3 to 4 elements only has a negligible effect on the performance.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/7511

• 📄 Preview Python docs built from this PR

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

❌ 2 Failures

As of commit 71d7e81:

NEW FAILURES - The following jobs have failed:

• unittests (3.8, linux.g5.4xlarge.nvidia.gpu, cuda, 11.7) / linux-job (gh)
• unittests-prototype (3.8, linux.g5.4xlarge.nvidia.gpu, cuda, 11.7) / linux-job (gh)

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

[NicolasHug]

Thanks Philip, LGTM if/when green

[github-actions]

Hey @pmeier!

You merged this PR, but no labels were added. The list of valid labels is available at https://github.com/pytorch/vision/blob/main/.github/process_commit.py