Skip to content

Latest commit

 

History

History
92 lines (62 loc) · 4.39 KB

README.md

File metadata and controls

92 lines (62 loc) · 4.39 KB

MuJoCo Playground

MuJoCo Playground

A comprehensive suite of GPU-accelerated environments for robot learning research and sim-to-real, built with MuJoCo MJX.

Features include:

  • Classic control environments from dm_control reimplemented in MJX.
  • Quadruped and bipedal locomotion environments.
  • Non-prehensile and dexterous manipulation environments.
  • Vision-based support available via Madrona-MJX.

For more details, checkout the project website.

Installation

You can install MuJoCo Playground directly from PyPI:

pip install playground

From Source

Important

Requires Python 3.9 or later.

  1. pip install -U "jax[cuda12]"
    • Verify GPU backend: python -c "import jax; print(jax.default_backend())" should print gpu
  2. git clone [email protected]:google-deepmind/mujoco_playground.git
  3. cd mujoco_playground
  4. uv pip install -e ".[all]"

Madrona-MJX (optional)

For vision-based environments, please refer to the installation instructions in the Madrona-MJX repository.

Getting started

To try out MuJoCo Playground locally on a simple locomotion environment, run the following:

import jax
import jax.numpy as jp
from mujoco_playground import registry

env = registry.load('Go1JoystickFlatTerrain')
state = jax.jit(env.reset)(jax.random.PRNGKey(0))
print(state.obs)
state = jax.jit(env.step)(state, jp.zeros(env.action_size))
print(state.obs)

For detailed tutorials on using MuJoCo Playground, see:

  1. Intro. to the Playground with DM Control Suite Open In Colab
  2. Locomotion Environments Open In Colab
  3. Manipulation Environments Open In Colab

For tutorials on using MuJoCo Playground with Madrona-MJX, see:

  1. Training CartPole from Vision Open In Colab
  2. Robotic Manipulation from Vision Open In Colab

How can I contribute?

Get started by installing the library and exploring its features! Found a bug? Report it in the issue tracker. Interested in contributing? If you’re a developer with robotics experience, we’d love your help—check out the contribution guidelines for more details.

Citation

If you use Playground in your scientific works, please cite it as follows:

@misc{mujoco_playground_2025,
  title = {MuJoCo Playground: An open-source framework for GPU-accelerated robot learning and sim-to-real transfer.},
  author = {Zakka, Kevin and Tabanpour, Baruch and Liao, Qiayuan and Haiderbhai, Mustafa and Holt, Samuel and Luo, Jing Yuan and Allshire, Arthur and Frey, Erik and Sreenath, Koushil and Kahrs, Lueder A. and Sferrazza, Carlo and Tassa, Yuval and Abbeel, Pieter},
  year = {2025},
  publisher = {GitHub},
  url = {https://github.com/google-deepmind/mujoco_playground}
}

License and Disclaimer

The texture used in the rough terrain for the locomotion environments is from Polyhaven and licensed under CC0.

All other content in this repository is licensed under the Apache License, Version 2.0. A copy of this license is provided in the top-level LICENSE file in this repository. You can also obtain it from https://www.apache.org/licenses/LICENSE-2.0.

This is not an officially supported Google product.