Sparse DETR: Efficient End-to-End Object Detection with Learnable Sparsity

Authors: Byungseok Roh, JaeWoong Shin, Wuhyun Shin, Saehoon Kim

ICLR 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on the COCO 2017 benchmark (Lin et al., 2014) demonstrate that Sparse DETR effectively reduces computational cost while achieving better detection performance.
Researcher Affiliation Industry Byungseok Roh1 , Jae Woong Shin2 , Wuhyun Shin1 , Saehoon Kim1 1Kakao Brain 2Lunit {peter.roh,aiden.hsin,sam.kim}@kakaobrain.com jwoong.shin@lunit.io
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code Yes Code is available at https://github.com/kakaobrain/sparse-detr.
Open Datasets Yes Extensive experiments on the COCO 2017 benchmark (Lin et al., 2014)
Dataset Splits Yes Extensive experiments on the COCO 2017 benchmark (Lin et al., 2014) demonstrate that Sparse DETR effectively reduces computational cost while achieving better detection performance.
Hardware Specification Yes We train the model on a 4 V100 GPU machine with a total batch size of 16
Software Dependencies No The paper does not provide specific version numbers for software dependencies.
Experiment Setup Yes We train the model on a 4 V100 GPU machine with a total batch size of 16, for 50 epochs, where the initial learning rate is 0.0002 and decayed by 1/10 at the 40 epoch.