Skip to content

Unitree RL GYM

A reinforcement learning platform for Unitree robots

Visit official site

What Unitree RL GYM is

The platform is described as a reinforcement learning environment built around Unitree robots.

What you can do with it

  • Train walking and balance policies in simulation
  • Test a new RL algorithm on a real robot
  • Demonstrate robot learning in a classroom

Getting started

  • Create a Python 3.8 virtual environment
  • Install PyTorch 2.3.1 with a matching CUDA version
  • Install Isaac Gym and set up the PPO implementation

Pros & cons

✓ What we like

  • Custom tasks and experiment parameters can be configured
  • Version control of code and algorithms aids reproducibility

! What to watch out for

  • Named robot models are Go2, H1, H1_2 and G1
  • Setup names Python 3.8 and PyTorch 2.3.1

FAQ

Which robots are supported?

Go2, H1, H1_2 and G1 are listed.

What simulation environments are named?

Isaac Gym and Mujoco are listed.

Last reviewed: 2026-11-15

More AI coding tools tools

View all →

How we review