Skip to content

BEVDet online real-time inference using CUDA, TensorRT, ROS1 & C++.

License

Notifications You must be signed in to change notification settings

collector-m/BEVDet-ROS-TensorRT

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

11 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

BEVDet by TensorRT, C++, ROS

  • This repository contains source code and models for BEVDet online real-time inference using CUDA, TensorRT, ROS1 & C++.

  • Modify the record reference modify

1 ENV

  • ubuntu-20.04、CUDA-11.3、cuDNN-8.6.0、TensorRT-8.5
  • yaml-cpp、Eigen3、libjpeg

2 Build

mkdir -p bev_ws/src
cd bev_ws/src
git clone https://github.com/linClubs/BEVDet-ROS-TensorRT.git
cd ..
catkin_make
source devel/setup.bash

3 Run

  1. generate engine

Generate the onnx TensorRT engine reference to bevdet-tensorrt-cpp engine path as follows.

# engine path
BEVDet-ROS-TensorRT
    └──ckpts
        ├── bev_stage_lt_d.engine
        ├── img_stage_lt_d.engine
        └── lt_d.yaml
  1. data preparation The rosbag folder can be downloaded from Baidu Netdisk

  2. demo test

# 1. start bevdet_node
roslaunch bevdet bevdet_node.launch

# 2  play data
rosbag play nus.bag


References


About

BEVDet online real-time inference using CUDA, TensorRT, ROS1 & C++.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 51.2%
  • Cuda 31.1%
  • Python 14.1%
  • CMake 2.8%
  • C 0.8%