SO-ARM Starter Workflow Docker Container#
Prerequisites#
- NVIDIA Docker Runtime (nvidia-container-toolkit)
- NVIDIA GPU with CUDA support (RTX 6000 Ada Generation or compatible)
- NVIDIA Drivers (version 535 or higher)
- X11 forwarding support (for GUI mode)
- Sufficient disk space (at least 50GB for asset caching and model)
- Memory (at least 32GB RAM recommended)
- RTI License should be placed in the
~/docker/rti/directory on your host system
Build the Docker Image#
Build the Docker image with all necessary dependencies:
# Clone the repository (if not already done)
git clone https://github.com/isaac-for-healthcare/i4h-workflows.git
cd i4h-workflows
# Build the Docker image with no cache to ensure fresh build
docker build --no-cache -f workflows/so_arm_starter/docker/Dockerfile -t so_arm_starter:latest .
Running the Container#
Enable X11 Display Access#
# Allow Docker containers to access your X11 display
xhost +local:docker
Run the Container#
Please add --privileged and mount host USB ports (e.g., -v /dev:/dev) when accessing SOARM hardware through the container, as this grants the necessary permissions for hardware communication.
docker run --name soarm -it --gpus all --privileged --rm \
--runtime=nvidia \
--entrypoint=bash \
-e "OMNI_KIT_ACCEPT_EULA=Y" \
-e "ACCEPT_EULA=Y" \
-e "PRIVACY_CONSENT=Y" \
-e "DISPLAY=$DISPLAY" \
-e "NVIDIA_VISIBLE_DEVICES=all" \
-e "NVIDIA_DRIVER_CAPABILITIES=compute,utility,graphics,display" \
-v /tmp/.X11-unix:/tmp/.X11-unix:rw \
-v ~/.Xauthority:/root/.Xauthority:rw \
-v ~/docker/isaac-sim/cache/kit:/isaac-sim/kit/cache:rw \
-v ~/docker/isaac-sim/cache/ov:/root/.cache/ov:rw \
-v ~/docker/isaac-sim/cache/pip:/root/.cache/pip:rw \
-v ~/docker/isaac-sim/cache/glcache:/root/.cache/nvidia/GLCache:rw \
-v ~/docker/isaac-sim/cache/computecache:/root/.nv/ComputeCache:rw \
-v ~/docker/isaac-sim/logs:/root/.nvidia-omniverse/logs:rw \
-v ~/docker/isaac-sim/data:/root/.local/share/ov/data:rw \
-v ~/docker/isaac-sim/documents:/root/Documents:rw \
-v ~/docker/rti:/root/rti:ro \
-v /dev:/dev \
so_arm_starter:latest
Running Workflow#
Once inside the container, you can run the SO-ARM Starter workflow. For detailed instructions on running simulation scenarios, training policies, and evaluating models, refer to the Main Workflow Guide which contains comprehensive examples and command-line options.
Troubleshooting#
- If run with error `GLIBCXX_3.4.30' not found, please run
conda install -c conda-forge libgcc-ng=12 libstdcxx-ng=12 -y
- For RTI license configuration, software installation issues, and additional troubleshooting steps, refer to the Docker Guide which contains comprehensive solutions for common container and environment setup problems.