Docker Tutorial for DIAL-MPC
This repository contains the Dockerfile of the DIAL-MPC. DIAL-MPC is an impressive framework for legged robot full-order torque-level control with both precision and agility in a training-free manner. You can find the DIAL-MPC in here.
Caution
You need a Linux or Ubuntu system with Nvidia-driver (>=525.60.13) to support CUDA 12.*
curl -fsSL https://get.docker.com -o get-docker.sh
sudo bash get-docker.sh
Step 2. Docker Postinstall(Follow here)
Step 3. Docker Nvidia Toolkit(Follow here)
sudo apt install x11-xserver-utils
git clone https://github.com/XuXinhangNTU/dial-mpc-docker
cd dial-mpc-docker
docker build -t dial-mpc .
xhost +
docker run -it -v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=unix$DISPLAY --gpus all --name <Your Name> dial-mpc
xhost +
docker start <Your Name> && docker exec -it <Your Name> bash
The vscode plugin "Dev Containers" is very easy to use when you do docker development.
Caution
If the demo looks odd, check your GPU usage! RTX 3090 is tested.