Skip to content
RobothecaBeta
Log in

robot_lab

  • Software
  • Humanoids and robot learning
  • Framework

An extension library for Isaac Lab that trains reinforcement learning locomotion policies for many legged, wheeled and humanoid robots in an isolated repository.

  • Closed core

Source on GitHub: github.com/fan-ziqi/robot_lab (opens in a new tab). Facts read at commit 500399e (opens in a new tab).

Licenses

Code
Apache-2.0 (opens in a new tab)

Read by Robotheca from LICENSE (opens in a new tab) (Apache-2.0) at commit 500399e (opens in a new tab), on 16 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 NVIDIA Isaac Sim: robot_lab, an extension of Isaac Lab, runs on NVIDIA Isaac Sim, whose Omniverse components are under NVIDIA's own terms.

Labels

  • Closed coreIt needs closed parts: NVIDIA Isaac Sim.

What it does

  • Provides ready-made Isaac Lab environments for quadruped, wheeled and humanoid robots, such as Anymal D, Unitree Go2 and G1.
  • Installs as a Python package on top of an existing Isaac Lab installation and can be listed with a helper script.
  • Supports training and playback with the RSL-RL library and an experimental CusRL backend.
  • Includes a motion-tracking workflow for Unitree G1 that converts retargeted motion CSV files into NPZ format and replays them.
  • Offers experimental AMP dance and handstand tasks, plus symmetry augmentation and teacher-student distillation for Anymal D.
  • Can run training across multiple GPUs and nodes, and provides a Docker setup built on a locally built Isaac Lab base image.

Written by Robotheca from the project's README (opens in a new tab) at commit 500399e.

Getting started

Difficulty
Advanced

Maintenance

Status
Maintained: a commit in the last 12 months
Newest commit
27 May 2026 on main
Latest release
v2.3.2 (3 Feb 2026)

Checked on GitHub on 16 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.