Brax - A Differentiable Physics Engine for Large Scale Rigid Body Simulation

C. Daniel Freeman, Erik Frey, Anton Raichuk, Sertan Girgin, Igor Mordatch, Olivier Bachem

Keywords: reinforcement learning, physics, differentiable, jax, robotics

TL;DR: We present a new differentiable rigid body physics engine in JAX and several reimplementations of common RL algorithms that can compile to and run on the same accelerator.

Abstract: We present Brax, an open source library for \textbf{r}igid \textbf{b}ody simulation with a focus on performance and parallelism on accelerators, written in JAX. We present results on a suite of tasks inspired by the existing reinforcement learning literature, but remade in our engine. Additionally, we provide reimplementations of PPO, SAC, ES, and direct policy optimization in JAX that compile alongside our environments, allowing the learning algorithm and the environment processing to occur on the same device, and to scale seamlessly on accelerators. Finally, we include notebooks that facilitate training of performant policies on common MuJoCo-like tasks in minutes.

URL: https://github.com/google/brax