MuJoCo Playground

TL;DR

MuJoCo Playground is an open-source framework that simplifies robot simulation and training, supporting diverse robots and enabling effective zero-shot sim-to-real transfer with minimal setup.

Contribution

It introduces a comprehensive, easy-to-use platform combining physics simulation, rendering, and training environments for rapid robot policy development.

Findings

Supports diverse robotic platforms including quadrupeds and humanoids

Enables zero-shot sim-to-real transfer from state and pixel inputs

Allows training policies in minutes on a single GPU

Abstract

We introduce MuJoCo Playground, a fully open-source framework for robot learning built with MJX, with the express goal of streamlining simulation, training, and sim-to-real transfer onto robots. With a simple "pip install playground", researchers can train policies in minutes on a single GPU. Playground 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. This is achieved through an integrated stack comprising a physics engine, batch renderer, and training environments. Along with video results, the entire framework is freely available at playground.mujoco.org