search

fundamentalvision / Deformable-DETR

Deformable DETR: Deformable Transformers for End-to-End Object Detection.
Apache License 2.0
3.22k stars 520 forks source link

Minor issue in comment #169

Open seyiqi opened 2 years ago

seyiqi commented 2 years ago

Should the output bbox be in format (cx, cy, w, h) instead of what's written in this line of comment: https://github.com/fundamentalvision/Deformable-DETR/blob/11169a60c33333af00a4849f1808023eba96a931/models/deformable_detr.py#L263