ros2_cuda11.3_workspaces

Minimum environment

In this package we need this specs and packages on your host machine

• nvidia gpu
• AMD64 CPU
• nvidia drivers
• x11(X Windows System ver.11 )
• NVIDIA container toolkit
• Docker
• AMD64 CPU

How to use this package

1. clone this

git clone https://github.com/tkp153/ros2_cuda11.3_workspaces.git

2. build docker file

docker build ./ -t {containername:tag}
#example:
docker build ./ -t workspace:1.0

3. run it.

chmod +x docker_run.bash
./docker_run.bash

What modules are included

The base docker image is nvidia/cuda:11.3.0-cudnn8-devel-ubuntu20.04(cu113) and nvidia/cuda:11.6.0-cudnn8-devel-ubuntu20.04(cu116). this Doclerfile are included these packages which are shown a list below.

Main modules

• Ros2 (Foxy)
• Azure Kinect Sensor SDK(Already build)
• Intel RealSense SDK 2.0
• openpifpaf -> version 0.13.6
• Cuda 11.3
• Pytorch 1.12.1 with cuda (included torch vision and torchaudio)
• opencv
• motpy
  remove
• yolox
  remove

Sub modules

• TurtleBot3 packages (Foxy version)
  • Maybe all packages for TurtleBot3
• nano
• gazebo11
• gedbi
• pcmanfm
• terminator (v.2 add)

Docker Image information

Image Size: 22.2 GB (version:1 21.8 GB) Build Time: Approx: 30min(Depends on communication speed) X11_server: True

Version information

1.0
first release

2.0 (2022/10/27)
add terminator

2.01 (2022/11/11)
add YOLOX,motpy

2.02 (2022/11/14)
remove Yolox and MOTPY lock openpifpaf version to 0.13.6

2.03 (2022/11/14)

• add v4l2_camera_node

• reduce file size

2.04(2022/11/16)
1.add cuda 11.6 version dockerfile
2.restore Kinect Azure SDK
3. reduce image size

2.05(2023/4/7)
1. add 2023Developer_HSR Dockerfiles
2. add 2023Developer_HSR docker run bash script