SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2201.12329 · ICLR · 2022

DAB-DETR: Dynamic Anchor Boxes are Better Queries for DETR

Hang Su, Jun Zhu, Xiao Yang, Lei Zhang, Hao Zhang, Shilong Liu, Feng Li, Xianbiao Qi

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 11 functions out of this paper's own repositories and ran 6 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
IDEA-Research/DAB-DETR canonical 1 of 1
helq2612/biadt — 5 of 8
idea-research/dn-detr — 0 of 2
FunctionStatusWhere it lives
GradientReversal Ran helq2612/biadt/models/dn_dab_deformable_detr/dab_deformable_detr.py
code served (permissive licence) · get_code("764f03e6f6c10909")
GradientReversalFunction Ran helq2612/biadt/models/dn_dab_deformable_detr/dab_deformable_detr.py
code served (permissive licence) · get_code("6e1a6cc4327acf17")
Mutual_loss Ran helq2612/biadt/models/dn_dab_deformable_detr/dab_deformable_detr.py
code served (permissive licence) · get_code("79c4cf758452c514")
NestedTensor Ran helq2612/biadt/models/dn_dab_deformable_detr/dab_deformable_detr.py
code served (permissive licence) · get_code("d336c6095b11629a")
nested_tensor_from_tensor_list Ran helq2612/biadt/models/dn_dab_deformable_detr/dab_deformable_detr.py
code served (permissive licence) · get_code("6442da4f34dd152c")
sigmoid_focal_loss Ran IDEA-Research/DAB-DETR/models/DAB_DETR/DABDETR.py
code served (permissive licence) · get_code("5c0711aada67957e")
DABDETR Not yet run idea-research/dn-detr/models/DN_DAB_DETR/DABDETR.py
code served (permissive licence) · get_code("530114cee9263694")
DABDeformableDETR Not yet run helq2612/biadt/models/dn_dab_deformable_detr/dab_deformable_detr.py
code served (permissive licence) · get_code("fe79de37d598acb3")
_onnx_nested_tensor_from_tensor_list Not yet run helq2612/biadt/models/dn_dab_deformable_detr/dab_deformable_detr.py
code served (permissive licence) · get_code("50ec3e702e11eebb")
prepare_for_dn Not yet run helq2612/biadt/models/dn_dab_deformable_detr/dab_deformable_detr.py
code served (permissive licence) · get_code("af4a2db04af02c0a")
prepare_for_dn Not yet run idea-research/dn-detr/models/DN_DAB_DETR/DABDETR.py
code served (permissive licence) · get_code("daa15d8205b6349a")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We present in this paper a novel query formulation using dynamic anchor boxes for DETR (DEtection TRansformer) and offer a deeper understanding of the role of queries in DETR. This new formulation directly uses box coordinates as queries in Transformer decoders and dynamically updates them layer-by-layer. Using box coordinates not only helps using explicit positional priors to improve the queryto-feature similarity and eliminate the slow training convergence issue in DETR, but also allows us to modulate the positional attention map using the box width and height information. Such a design makes it clear that queries in DETR can be implemented as performing soft ROI pooling layer-by-layer in a cascade manner. As a result, it leads to the best performance on MS-COCO benchmark among the DETR-like detection models under the same setting, e.g., AP 45.7% using ResNet50-DC5 as backbone trained in 50 epochs. We also conducted extensive experiments to confirm our analysis and verify the effectiveness of our methods.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2201.12329")
get_code_for_paper("2201.12329")
have("2201.12329")

Connect an agent — have() is free.