First clone the Git repo
git clone --branch release_18 https://github.com/Unity-Technologies/ml-agents.git
conda create -n unity python=3.8.6
conda activate unity
Install Python 3.8 and set to primary Path Torch Download
pip install ~/Downloads/torch-1.8.0a0-cp38-cp38-macosx_11_0_arm64.whl
pip install torchvision
pip install torchaudio
conda install grpcio h5py
In the Git Repo
pip install -e ./ml-agents-envs
pip install -e ./ml-agents
Install the Unity-Package: Start Unity-Hub Open Project from Downloaded GitHub-Repo: /Project/Assets/ML-Agents/Examples/ „Window“ -> Package Manager Activate ML-Agengs and ML-Agents-Environment
mlagents-learn config/ppo/3DBall.yaml --run-id=3DBall01
Hit Play-Button in Unity View the Results:
tensorboard --logdir results
