BEVDet

News

2022.08.26 A blog in chinese for code explanation. BEVDet系列源码解读
2022.08.15 Support FP16 training for BEVDet series that with image view backbone type of ResNet.
2022.07.29 Support BEVDepth.
2022.07.26 Add configs and pretrained models of bevdet-r50 and bevdet4d-r50.
2022.07.13 Support bev-pool proposed in BEVFusion, which will speed up the training process of bevdet-tiny by +25%.
2022.07.08 Support visualization remotely! Please refer to Get Started for usage.
2022.06.29 Support acceleration of the Lift-Splat-Shoot view transformer! Please refer to [Technical Report] for detailed introduction and Get Started for testing BEVDet with acceleration.
2022.06.01 We release the code and models of both BEVDet and BEVDet4D!
2022.04.01 We propose BEVDet4D to lift the scalable BEVDet paradigm from the spatial-only 3D space to the spatial-temporal 4D space. Technical report is released on arixv. [BEVDet4D].
2022.04.01 We upgrade the BEVDet paradigm with some modifications to improve its performance and inference speed. Thchnical report of BEVDet has been updated. [BEVDetv1].
2021.12.23 BEVDet is now on arxiv. [BEVDet].

Main Results

FP32

Method	mAP	NDS	FPS	Mem (MB)	Model	Log
BEVDet-R50	29.9	37.7	16.7	5,007	google	google
BEVDepth-R50*	33.3	40.6	15.7	5,185	google	google
BEVDet4D-R50	32.2	45.7	16.7	7,089	google	google
BEVDepth4D-R50*	36.1	48.5	15.7	7,365	google	google
-	-	-	-	-	-	-
BEVDet-Tiny	30.8	40.4	15.6	6,187	google / baidu	google /baidu
BEVDet4D-Tiny	33.8	47.6	15.5	9,255	google / baidu	google /baidu

*Thirdparty implementation, please refer to Megvii for official implementation.
Memory is tested in the training process with batch 1 and without using torch.checkpoint.

FP16

Method	mAP	NDS	FPS	Mem (MB)	Model	Log
BEVDet-R50-FP16	29.6	37.4	16.7	3,931	google	google
BEVDet4D-R50-FP16	32.7	46.0	16.7	5,145	google	google
BEVDepth4D-R50-FP16*	36.4	48.4	15.7	5,361	google	google

Get Started

Installation and Data Preparation

Please see getting_started.md

Estimate the inference speed of BEVDet

# with acceleration
python tools/analysis_tools/benchmark.py configs/bevdet/bevdet-sttiny-accelerated.py $checkpoint
# without acceleration
python tools/analysis_tools/benchmark.py configs/bevdet/bevdet-sttiny.py $checkpoint

Estimate the flops of BEVDet

Note: For bevdet4d, the FLOP result involves the current frame only.

python tools/analysis_tools/get_flops.py configs/bevdet/bevdet-sttiny.py --shape 256 704
python tools/analysis_tools/get_flops.py configs/bevdet4d/bevdet4d-sttiny.py --shape 256 704

Visualize the predicted result with open3d.

Official implementation. (Visualization locally only)

python tools/test.py $config $checkpoint --show --show-dir $save-path

Private implementation. (Visualization remotely/locally)

python tools/test.py $config $checkpoint --format-only --eval-options jsonfile_prefix=$savepath
python tools/analysis_tools/vis.py $savepath/pts_bbox/results_nusc.json

Acknowledgement

This project is not possible without multiple great open-sourced code bases. We list some notable examples below.

Beside, there are some other attractive works extend the boundary of BEVDet.

BEVerse for multi-task learning.
BEVFusion for acceleration, multi-task learning, and multi-sensor fusion. (Note: The acceleration method is a concurrent work of that of BEVDet and has some superior characteristics like memory saving and completely equivalent.)

Bibtex

If this work is helpful for your research, please consider citing the following BibTeX entry.

@article{huang2022bevdet4d,
  title={BEVDet4D: Exploit Temporal Cues in Multi-camera 3D Object Detection},
  author={Huang, Junjie and Huang, Guan},
  journal={arXiv preprint arXiv:2203.17054},
  year={2022}
}

@article{huang2021bevdet,
  title={BEVDet: High-performance Multi-camera 3D Object Detection in Bird-Eye-View},
  author={Huang, Junjie and Huang, Guan and Zhu, Zheng and Yun, Ye and Du, Dalong},
  journal={arXiv preprint arXiv:2112.11790},
  year={2021}
}

klingner/BEVDet