playground

An open-source library for GPU-accelerated robot learning and sim-to-real transfer.

https://github.com/google-deepmind/mujoco_playground

Science Score: 54.0%

This score indicates how likely this project is to be science-related based on various indicators:

✓
CITATION.cff file
Found CITATION.cff file
✓
codemeta.json file
Found codemeta.json file
✓
.zenodo.json file
Found .zenodo.json file
○
DOI references
○
Academic publication links
✓
Committers with academic emails
2 of 17 committers (11.8%) from academic institutions
○
Institutional organization owner
○
JOSS paper metadata
○
Scientific vocabulary similarity
Low similarity (16.1%) to scientific vocabulary

Keywords from Contributors

jax

Last synced: 6 months ago · JSON representation ·

Repository

An open-source library for GPU-accelerated robot learning and sim-to-real transfer.

Basic Info

Host: GitHub
Owner: google-deepmind
License: apache-2.0
Language: Jupyter Notebook
Default Branch: main
Homepage: https://playground.mujoco.org
Size: 322 MB

Statistics

Stars: 1,339
Watchers: 17
Forks: 182
Open Issues: 21
Releases: 4

Created about 1 year ago · Last pushed 6 months ago

Metadata Files

Readme Changelog Contributing License Citation

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.
Quadruped and bipedal locomotion environments.
Non-prehensile and dexterous manipulation environments.
Vision-based support available via Madrona-MJX.

For more details, check out the project website.

[!NOTE] We now support training with both the MuJoCo MJX JAX implementation, as well as the MuJoCo Warp implementation at HEAD. See MuJoCo 3.3.5 release notes under MJX for more details.

Installation

You can install MuJoCo Playground directly from PyPI:

sh pip install playground

[!WARNING] The playground release may depend on pre-release versions of mujoco and warp-lang, in which case you can try pip install playground --extra-index-url=https://py.mujoco.org --extra-index-url=https://pypi.nvidia.com/warp-lang/. If there are still version mismatches, please open a github issue, and install from source.

From Source

[!IMPORTANT] Requires Python 3.10 or later.

git clone git@github.com:google-deepmind/mujoco_playground.git && cd mujoco_playground
Install uv, a faster alternative to pip
Create a virtual environment: uv venv --python 3.11
Activate it: source .venv/bin/activate
Install CUDA 12 jax: uv pip install -U "jax[cuda12]"
- Verify GPU backend: python -c "import jax; print(jax.default_backend())" should print gpu
Install playground: uv pip install -e ".[all]"
Verify installation (and download Menagerie): python -c "import mujoco_playground"

Madrona-MJX (optional)

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

Getting started

Basic Tutorials

| Colab | Description | |-------|-------------| | | Introduction to the Playground with DM Control Suite | | | Locomotion Environments | | | Manipulation Environments |

Vision-Based Tutorials (GPU Colab)

| Colab | Description | |-------|-------------| | | Training CartPole from Vision (T4 Instance) |

Local Runtime Tutorials

Requires local Madrona-MJX installation

| Colab | Description | |-------|-------------| | | Training CartPole from Vision | | | Robotic Manipulation from Vision |

Running from CLI

[!IMPORTANT] Assumes installation from source.

For basic usage, navigate to the repo's directory and run: bash python learning/train_jax_ppo.py --env_name CartpoleBalance

Training Visualization

To interactively view trajectories throughout training with rscope, install it (pip install rscope) and run:

``` python learning/trainjaxppo.py --envname PandaPickCube --rscopeenvs 16 --runevals=False --deterministicrscope=True

In a separate terminal

python -m rscope ```

FAQ

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 are a developer with robotics experience, we would love your help—check out the contribution guidelines for more details.

Reproducibility / GPU Precision Issues

Users with NVIDIA Ampere architecture GPUs (e.g., RTX 30 and 40 series) may experience reproducibility issues in mujocoplayground due to JAX’s default use of TF32 for matrix multiplications. This lower precision can adversely affect RL training stability. To ensure consistent behavior with systems using full float32 precision (as on Turing GPUs), please run `export JAXDEFAULTMATMULPRECISION=highestin your terminal before starting your experiments (or add it to the end of~/.bashrc`).

To reproduce results using the same exact learning script as used in the paper, run the brax training script which is available here. There are slight differences in results when using the learning/train_jax_ppo.py script, see the issue here for more context.

Citation

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

bibtex @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.

Owner

Name: Google DeepMind
Login: google-deepmind
Kind: organization

Website: https://www.deepmind.com/
Repositories: 245
Profile: https://github.com/google-deepmind

Citation (CITATION.cff)

cff-version: 1.2.0
message: "If you use this software, please cite it using the following reference:"
title: "MuJoCo Playground: An open-source framework for GPU-accelerated robot learning and sim-to-real transfer."
authors:
  - family-names: Zakka
    given-names: Kevin
  - family-names: Tabanpour
    given-names: Baruch
  - family-names: Liao
    given-names: Qiayuan
  - family-names: Haiderbhai
    given-names: Mustafa
  - family-names: Holt
    given-names: Samuel
  - family-names: Luo
    given-names: Jing Yuan
  - family-names: Allshire
    given-names: Arthur
  - family-names: Frey
    given-names: Erik
  - family-names: Sreenath
    given-names: Koushil
  - family-names: Kahrs
    given-names: Lueder A.
  - family-names: Sferrazza
    given-names: Carlo
  - family-names: Tassa
    given-names: Yuval
  - family-names: Abbeel
    given-names: Pieter
year: 2025
repository-code: "https://github.com/google-deepmind/mujoco_playground"
url: "https://github.com/google-deepmind/mujoco_playground"
license: "Apache-2.0"

GitHub Events

Total

Fork event: 164
Create event: 10
Release event: 4
Issues event: 157
Watch event: 1,164
Delete event: 4
Issue comment event: 333
Member event: 1
Public event: 1
Push event: 107
Pull request review comment event: 35
Pull request review event: 62
Pull request event: 69

Last Year

Fork event: 164
Create event: 10
Release event: 4
Issues event: 157
Watch event: 1,164
Delete event: 4
Issue comment event: 333
Member event: 1
Public event: 1
Push event: 107
Pull request review comment event: 35
Pull request review event: 62
Pull request event: 69

Committers

Last synced: 8 months ago

All Time

Total Commits: 91
Total Committers: 17
Avg Commits per committer: 5.353
Development Distribution Score (DDS): 0.659

Past Year

Commits: 91
Committers: 17
Avg Commits per committer: 5.353
Development Distribution Score (DDS): 0.659

Top Committers

Name	Email	Commits
Kevin Zakka	z**a@g**m	31
Baruch Tabanpour	b**a@g**m	26
andrew	a**1@o**m	11
Sam Holt	s**t@g**m	7
VincentZhang	v**g@u**a	3
Zizhao Chen	c**h@g**m	2
Alessio Quaglino	q**o@g**m	1
DeepMind	n**y@g**m	1
Hana Joo	h**o@g**m	1
Huy Ha	3****a	1
Jakob Hollenstein	5****h	1
Mustafa H	3****H	1
Peter Hawkins	p**s@g**m	1
brentyi	y**h@g**m	1
cornmander	s**n@g**m	1
Carlo Sferrazza	c**a@b**u	1
Yasunori Toshimitsu	y**i@g**m	1

Committer Domains (Top 20 + Academic)

google.com: 7 berkeley.edu: 1 ualberta.ca: 1

Issues and Pull Requests

Last synced: 6 months ago

All Time

Total issues: 104
Total pull requests: 62
Average time to close issues: 13 days
Average time to close pull requests: 4 days
Total issue authors: 72
Total pull request authors: 21
Average comments per issue: 1.43
Average comments per pull request: 1.03
Merged pull requests: 28
Bot issues: 0
Bot pull requests: 0

Past Year

Issues: 104
Pull requests: 62
Average time to close issues: 13 days
Average time to close pull requests: 4 days
Issue authors: 72
Pull request authors: 21
Average comments per issue: 1.43
Average comments per pull request: 1.03
Merged pull requests: 28
Bot issues: 0
Bot pull requests: 0

View more stats

Top Authors

Issue Authors

dementrock (11)
kassasin (7)
azimi99 (4)
karthyyy (2)
Nate711 (2)
hwang-warren (2)
btaba (2)
cvoelcker (2)
vidurvij (2)
MankaranSingh (2)
Greenery-S (2)
jloganolson (2)
Yasu31 (2)
ChenDavidTimothy (2)
mpiseno (2)

Pull Request Authors

Andrew-Luo1 (14)
btaba (9)
vincentzhang (5)
jkbjh (4)
StafaH (4)
samholt (4)
carlosferrazza (2)
Yasu31 (2)
augustsmith11 (2)
huy-ha (2)
kevinzakka (2)
alik-git (2)
chenzizhao (2)
yardenas (1)
youwyu (1)

Top Labels

Issue Labels

enhancement (8) question (4) bug (4) help wanted (1) wontfix (1) documentation (1)

Pull Request Labels

Packages

Total packages: 1
Total downloads:
- pypi 4,951 last-month
Total docker downloads: 5

Total dependent packages: 0
Total dependent repositories: 2
Total versions: 6
Total maintainers: 1

pypi.org: playground

Documentation: https://playground.readthedocs.io/
License: Apache Software License
Latest release: 0.0.5
published 8 months ago

Versions: 6
Dependent Packages: 0
Dependent Repositories: 2
Downloads: 4,951 Last month
Docker Downloads: 5

Rankings

Docker downloads count: 3.8%

Dependent packages count: 10.1%

Dependent repos count: 11.6%

Downloads: 13.1%

Average: 15.5%

Forks count: 22.6%

Stargazers count: 31.9%

Maintainers (1)

kevinzakka

Last synced: 6 months ago

Dependencies

.github/workflows/ci.yml actions

actions/checkout v2 composite
actions/setup-python v1 composite

pyproject.toml pypi

playground

Science Score: 54.0%

Keywords from Contributors

Repository

Basic Info

Statistics

Metadata Files

README.md

MuJoCo Playground

Installation

From Source

Madrona-MJX (optional)

Getting started

Basic Tutorials

Vision-Based Tutorials (GPU Colab)

Local Runtime Tutorials

Running from CLI

Training Visualization

In a separate terminal

FAQ

How can I contribute?

Reproducibility / GPU Precision Issues

Citation

License and Disclaimer

Owner

Citation (CITATION.cff)

GitHub Events

Total

Last Year

Committers

All Time

Past Year

Top Committers

Committer Domains (Top 20 + Academic)

Issues and Pull Requests

All Time

Past Year

Top Authors

Issue Authors

Pull Request Authors

Top Labels

Issue Labels

Pull Request Labels

Packages

pypi.org: playground

Rankings

Maintainers (1)

Dependencies