SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2207.13080 · 2022

DETRs with Hybrid Matching

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 18 functions out of this paper's own repositories and ran 15 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
HDETR/H-Deformable-DETR canonical 7 of 8
HDETR/H-Detic-LVIS canonical 6 of 8
IDEA-Research/detrex extension 1 of 1
HDETR/H-PETR-3D extension 1 of 1
FunctionStatusWhere it lives
window_partition Ran HDETR/H-Deformable-DETR/models/swin_transformer.py
code served (permissive licence) · get_code("144d10b49baeb8a6")
convert_to_xywh Ran HDETR/H-Deformable-DETR/datasets/coco_eval.py
code served (permissive licence) · get_code("f31a58bf6457ced5")
dice_loss Ran HDETR/H-Deformable-DETR/models/segmentation.py
code served (permissive licence) · get_code("ac8fe530cdad4d8c")
evaluate Ran HDETR/H-Deformable-DETR/datasets/coco_eval.py
code served (permissive licence) · get_code("fe0ddcc2d420c9a0")
get_clip_embeddings Ran HDETR/H-Detic-LVIS/detic/predictor.py
code served (permissive licence) · get_code("1f211df98dc51cee")
get_fed_loss_inds Ran HDETR/H-Detic-LVIS/detic/modeling/utils.py
code served (permissive licence) · get_code("bb583e07209fe58b")
get_vit_lr_decay_rate Ran IDEA-Research/detrex/detrex/modeling/backbone/eva.py
code served (permissive licence) · get_code("08bf7ecdb2fffe84")
load_class_freq Ran HDETR/H-Detic-LVIS/detic/modeling/utils.py
code served (permissive licence) · get_code("1f6e4c3da7b82979")
match_name_keywords Ran HDETR/H-Detic-LVIS/detic/custom_solver.py
code served (permissive licence) · get_code("c15434906f804d7f")
pos2posemb3d Ran HDETR/H-PETR-3D/projects/mmdet3d_plugin/models/dense_heads/hybrid_petrv2_head.py
pointer only (licence: NOASSERTION) · get_code("718ca707584a0928")
sigmoid_focal_loss Ran HDETR/H-Deformable-DETR/models/segmentation.py
code served (permissive licence) · get_code("5c0711aada67957e")
train_hybrid Ran HDETR/H-Deformable-DETR/engine.py
code served (permissive licence) · get_code("8cc99e95f95cb562")
window_partition Ran HDETR/H-Detic-LVIS/detic/modeling/backbone/swintransformer.py
code served (permissive licence) · get_code("f9fd6241d935f07b")
window_reverse Ran HDETR/H-Deformable-DETR/models/swin_transformer.py
code served (permissive licence) · get_code("61bf152e6a42a184")
window_reverse Ran HDETR/H-Detic-LVIS/detic/modeling/backbone/swintransformer.py
code served (permissive licence) · get_code("fb32094c6dbece71")
compute_average_precision Not yet run HDETR/H-Detic-LVIS/detic/evaluation/oideval.py
code served (permissive licence) · get_code("2e0a799ccbd489e7")
create_timm_resnet Not yet run HDETR/H-Detic-LVIS/detic/modeling/backbone/timm.py
code served (permissive licence) · get_code("3634f2208a0255d7")
measure_average_inference_time Not yet run HDETR/H-Deformable-DETR/benchmark.py
code served (permissive licence) · get_code("3e5d34f155e97585")

Repositories linked to this paper

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

Abstract

One-to-one set matching is a key design for DETR to establish its end-to-end capability, so that object detection does not require a hand-crafted NMS (non-maximum suppression) to remove duplicate detections. This end-to-end signature is important for the versatility of DETR, and it has been generalized to broader vision tasks. However, we note that there are few queries assigned as positive samples and the one-to-one set matching significantly reduces the training efficacy of positive samples. We propose a simple yet effective method based on a hybrid matching scheme that combines the original one-to-one matching branch with an auxiliary one-to-many matching branch during training. Our hybrid strategy has been shown to significantly improve accuracy. In inference, only the original one-to-one match branch is used, thus maintaining the end-to-end merit and the same inference efficiency of DETR. The method is named H-DETR, and it shows that a wide range of representative DETR methods can be consistently improved across a wide range of visual tasks, including DeformableDETR, PETRv2, PETR, and TransTrack, among others. The code is available at: https://github.com/HDETR

For agents

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

get_harvested_code_for_paper("2207.13080")
get_code_for_paper("2207.13080")
have("2207.13080")

Connect an agent — have() is free.