Implementation of the MuZero algorithm
This repo contains:
- a simple but working implementation of the MuZero algorithm
- an agent trained using the MuZero algorithm to play an openAI gym game (CartPole-v1)
The code is an implementation of the official MuZero pseudocode.
This is an implementation of an agent that uses MuZero in order to play the openAI gym game of CartPole.
Execute the code in the notebook to train the agent!
To set up your python environment to run the code in this repository, follow the instructions below.
-
Create (and activate) a new environment with Python 3.6.
- Linux or Mac:
conda create --name MuZero python=3.6 source activate MuZero
- Windows:
conda create --name MuZero python=3.6 activate MuZero
-
Clone the repository, and then, install the required packages (see requirements).
git clone https://github.com/ciamic/MuZero.git
- Create an IPython kernel for the
MuZero
environment.
python -m ipykernel install --user --name MuZero --display-name "MuZero"
- Before running code in a notebook, change the kernel to match the
MuZero
environment by using the drop-down contextualKernel
menu.
Python 3
numpy
matplotlib
gym
Tensorflow