Skip to content

[Simulator] MuJoCo Playground #43

Description

@skt0725

ID (slug)

mujoco-playground

Name

MuJoCo Playground

Organization

Google DeepMind / UC Berkeley

Year

2025

Description (English)

MuJoCo Playground is a fully open-source framework for GPU-accelerated robot learning and sim-to-real transfer, built on MuJoCo MJX. With a simple pip install playground, researchers can train policies in minutes on a single GPU. It supports diverse robotic platforms including quadrupeds, humanoids, dexterous hands, and robotic arms, enabling zero-shot sim-to-real transfer from both state and pixel inputs through an integrated stack of physics engine, batch renderer and training environments.

Description (Korean)

MuJoCo Playground는 MuJoCo MJX 기반의 GPU 가속 로봇 학습 및 sim-to-real 전이를 위한 완전 오픈소스 프레임워크입니다. pip install playground 한 줄로 설치할 수 있으며, 단일 GPU에서 수 분 내에 정책을 학습할 수 있습니다. 사족보행 로봇, 휴머노이드, 로봇 손, 로봇 팔 등 다양한 로봇 플랫폼을 지원하며, 물리 엔진·배치 렌더러·학습 환경의 통합 스택을 통해 상태 및 픽셀 입력 모두에서 제로샷 sim-to-real 전이를 가능하게 합니다.

GitHub URL

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

Paper URL (arXiv)

https://arxiv.org/abs/2502.08844

Project / Docs URL

https://playground.mujoco.org/

Type

rl_framework — RL 학습 프레임워크 (Isaac Lab 등)

Features

  • GPU-accelerated (병렬 시뮬레이션 지원)
  • ROS2 support

Primary Language(s)

Python (JAX, Warp)

Tags (optional)

MJX, JAX, sim-to-real, locomotion, manipulation, zero-shot-transfer, reinforcement-learning

Checklist

  • The tool is open-source (publicly available on GitHub)
  • At least one URL (GitHub, paper, or project page) is provided
  • I have read the contribution guidelines

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Fields

    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions