Original link: https://github.com/Unity-Technologies/ml-agents/blob/master/docs/Installation.md
installation and setup
To install and use ML-Agents, you need to install Unity, clone this repository and install Python along with other dependencies. In addition to the Docker setup, each subsection below outlines each step.
Install Unity 2017.1 or later
Download and install Unity. If you want to use our Docker setup (covered later), make sure to select the Linux Build Support component when installing Unity.
Clone the ml-agents repository
Once installed, you need to clone the ML-Agents GitHub repository.
git clone [email protected]:Unity-Technologies/ml-agents.git
The unity-environment directory in this repository contains Unity assets to add to the project. The python directory contains training code. Both directories are located at the root of the repository.
Install Python (and dependencies)
In order to use ML-Agents, Python 3 is required along with the dependencies listed in the requirements file . Some major dependencies include:
- TensorFlow
- Jupyter
Windows users
If you are a Windows user new to Python and TensorFlow, follow this guide to set up your Python environment.
Mac and Unix users
Download and install Python 3 if you don't already have it .
If your Python environment doesn't include pip , see instructions for installing it .
To install dependencies, go into the python subdirectory of the repository and run from the command line:
pip3 install .
Docker-based installation
If you want to use Docker for ML-Agents, follow this guide .
Unity packages
You can download the TensorFlowSharp plugin as a Unity package here .
Help
If you encounter any problems installing ML-Agents, please file an issue and be sure to include information about the operating system, Python version and exact error message (if possible).