Details of DETR-DC5+ - Githubissues

fundamentalvision / Deformable-DETR

Deformable DETR: Deformable Transformers for End-to-End Object Detection.

Apache License 2.0

3.15k stars 513 forks source link

Open chufengt opened 2 years ago

chufengt commented 2 years ago

Hi, could you please share the details about how to reproduce the results of DETR-DC5+.

In the paper, you mentioned that it is DETR-DC5 with Focal Loss and 300 object queries.

I was wondering that are there any other differences, such as GroupNorm, sigmoid+topk PostProcess in Deformable-DETR, etc.