Multi-Agent-Unity

Distributed RL algorithms developed on Multi-Agent gym wrapper for unity environment

Environment

You can use either conda or docker

cd docker
bash build.sh
docker run -it -v ${path to the code in your computer}:/unity unity:1.0 bash
cd /unity
python training.py

conda env create -f environment.yml
conda activate unity
python training.py