Skip to content

Installation Guideยค

clone the repo
git clone
create and activate env
cd pdearena
conda env create --file docker/environment.yml
conda activate pdearena
install this package
# install so the project is in PYTHONPATH
pip install -e .
additionally installing Clifford Neural Layers
pip install "cliffordlayers @ git+"

If you also want to do data generation:

pip install -e ".[datagen]"
build docker container
cd docker
docker build -t pdearena .
run docker container
cd pdearena
docker run --gpus all -it --rm --user $(id -u):$(id -g) \
    -v $(pwd):/code -v /mnt/data:/data --workdir /code -e PYTHONPATH=/code \


  • --gpus all -it --rm --user $(id -u):$(id -g): enables using all GPUs and runs an interactive session with current user's UID/GUID to avoid docker writing files as root.
  • -v $(pwd):/code -v /mnt/data:/data --workdir /code: mounts current directory and data directory (i.e. the cloned git repo) to /code and /data respectively, and use the code directory as the current working directory.