Object DGCNN: 3D Object Detection using Dynamic Graphs

Authors: Yue Wang, Justin M. Solomon

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our method achieves state-of-the-art performance on autonomous driving benchmarks. We also provide abundant analysis of the detection model and distillation framework. We present our experiments in four parts. We introduce the dataset, metrics, implementation, and optimization details in 6.1. Then, we demonstrate performance on the nu Scenes dataset [66] in 6.2. We present knowledge distillation results in 6.3. Finally, we provide ablation studies in 6.4.
Researcher Affiliation Academia Yue Wang Massachusetts Institute of Technology yuewang@csail.mit.edu Justin Solomon Massachusetts Institute of Technology jsolomon@mit.edu
Pseudocode No The paper describes the method procedurally but does not include any labeled pseudocode or algorithm blocks.
Open Source Code Yes We release our code to promote reproducibility and future research. 1https://github.com/WangYueFt/detr3d
Open Datasets Yes We experiment on the nu Scenes dataset [66].
Dataset Splits Yes It has 1K short sequences captured in Boston and Singapore with 700, 150, 150 sequences for training, validation, and testing, respectively.
Hardware Specification Yes We train for 20 epochs on 8 RTX 3090 GPUs.
Software Dependencies No The paper mentions 'Adam W' and 'MMDetection3D' but does not specify their version numbers or versions for other key software components.
Experiment Setup Yes We use Adam W [67] to train the model. The weight decay for Adam W is 10^-2. Following a cyclic schedule [68], the learning rate is initially 10^-4 and gradually increased to 10^-3, which is finally decreased to 10^-8. The model is initialized with a pre-trained Point Pillars network on the same dataset. We train for 20 epochs on 8 RTX 3090 GPUs.