Skip to content
RobothecaBeta
Log in

unitree_rl_mjlab

  • Software
  • Humanoids and robot learning
  • Framework

A reinforcement learning project built on mjlab and the MuJoCo physics engine for training and simulating locomotion and motion-imitation policies for Unitree robots, then deploying them to hardware.

  • Closed core

Made by Unitree Robotics.

Source on GitHub: github.com/unitreerobotics/unitree_rl_mjlab (opens in a new tab). Facts read at commit 1425b15 (opens in a new tab).

Licenses

Code
Apache-2.0 (opens in a new tab)

Read by Robotheca from LICENCE (opens in a new tab) (Apache-2.0) at commit 1425b15 (opens in a new tab), on 23 Oct 2026.

Commercial use
Unknown

It needs closed parts under their own terms: read them before any commercial use.

One part it needs is closed

  • Needs libunitree_sdk2 (prebuilt): Deploying to a real robot goes through unitree_sdk2, whose core ships as a prebuilt library without source.

Labels

  • Closed coreIt needs closed parts: libunitree_sdk2 (prebuilt).

What it does

  • Trains velocity-tracking policies for Unitree Go2, A2, G1, R1, H1_2 and H2 robots in MuJoCo simulation.
  • Trains motion-imitation policies that make a Unitree G1 mimic reference motion sequences converted from CSV to NPZ.
  • Replays trained policies in simulation to check their behavior before deployment.
  • Exports policy.onnx files during training for deployment on physical robots.
  • Deploys policies to real Unitree robots over Ethernet, with a MuJoCo-based simulation step beforehand.
  • Builds on mjlab, rsl_rl, whole_body_tracking and unitree_sdk2 for training and robot communication.

Written by Robotheca from the project's README (opens in a new tab) at commit 1425b15.

Getting started

Difficulty
Advanced

Maintenance

Status
Maintained: a commit in the last 12 months
Newest commit
13 Apr 2026 on main
Latest release
None published on GitHub

Checked on GitHub on 23 Oct 2026.

Links

Robots that can run it

No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.

Robotheca lists software as a reference: nothing is hosted or run here, and every fact links to where it was read.