fix: compute auxiliary losses when denoising is disabled in D-FINE

[Abineshabee]

Fixes #45593

Problem

When num_denoising=0, denoising_meta_values is None, causing DFineForObjectDetectionLoss to skip auxiliary loss computation entirely since it is gated under the same condition.

However, auxiliary losses (deep supervision across decoder layers) are independent of denoising and should still be computed when config.auxiliary_loss=True. This leads to degraded training behavior when denoising is disabled.

Fix

Decoupled denoising logic from auxiliary loss computation:

• Tensor splitting is performed only when denoising_meta_values is not None
• Auxiliary losses are always computed when config.auxiliary_loss=True
• Denoising-specific auxiliary losses are computed only when denoising is active

This follows the same pattern used in loss_rt_detr.py.

Testing

• Verified that auxiliary loss keys (e.g., loss_vfl_aux_*, loss_bbox_aux_*) are present when num_denoising=0
• Confirmed no change in behavior when denoising is enabled

[Rocketknight1]

cc @molbap @yonigozlan @NielsRogge maybe

[vasqu]

LGTM, but can we add a small test to confirm that behavior under dfine? Would like a second opinion from @yonigozlan tho if he agrees to this as I'm not 100% familiar with this

[Abineshabee]

Thanks for the review! That makes sense — I’ll add a small test to verify auxiliary losses are computed when num_denoising=0.

[Abineshabee]

CI failed due to a CircleCI authentication error (not related to the code change). Could someone please re-run the workflow?

[yonigozlan]

Looks good to me thanks for fixing @Abineshabee !

[github-actions]

[For maintainers] Suggested jobs to run (before merge)

run-slow: d_fine

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.