# roboterax/humanoid-gym

Humanoid-Gym: Reinforcement Learning for Humanoid Robot with Zero-Shot Sim2Real Transfer https://arxiv.org/abs/2404.05695

Repository: https://github.com/roboterax/humanoid-gym
Canonical: https://ross.abutalabs.com/products/humanoid-gym
Homepage: https://sites.google.com/view/humanoid-gym/
Language: Python
License Family: other
Topics: ai, artificial-intelligence, control-systems, humanoid-robot, locomotion, machine-learning, robotics
Last push: 2025-01-26T08:11:29+00:00

## Health v2 (maintenance only)
Score: 27/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 3, release rhythm 35, longevity 65
- inputs: {"age_days": 912, "days_push": 584, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2075, forks 254 (observed 2026-08-28T04:06:10.777586+00:00)

## What it is
Humanoid-Gym is a reinforcement learning framework built on Nvidia Isaac Gym for training locomotion skills in humanoid robots, with emphasis on zero-shot sim-to-real transfer. It includes a sim-to-sim pipeline to Mujoco for validating trained policies across different physics simulators.

## Use cases
- train humanoid robot walking policies with reinforcement learning
- transfer RL locomotion policies from simulation to real robots
- validate trained policies in Mujoco before real-world deployment
- train locomotion for stairs, slopes, and uneven terrain
- adapt the framework to a custom humanoid robot

## When to choose
- you are training locomotion skills for a humanoid robot
- you need zero-shot sim-to-real transfer of RL policies
- you want a sim-to-sim verification pipeline between Isaac Gym and Mujoco

## When to avoid
- you need manipulation or dexterous hand training, which is still under development
- you lack a GPU capable of running Isaac Gym
- you work with non-humanoid robots requiring substantial rework

## Facets
- artifact type: framework
- maturity: active
- function: reinforcement-learning, simulation, robotics
- domain: robotics, reinforcement-learning, machine-learning, simulation
- platform: python
- tags: humanoid-robots, sim2real, sim2sim, isaac-gym, mujoco, locomotion, linux, gpu

## Member repositories
- roboterax/humanoid-gym (main) score 27

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:10.777586+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:56:17.855399+00:00, confidence not recorded.
  - readme: https://github.com/roboterax/humanoid-gym (fetched 2026-08-28T04:06:10.777586+00:00, sha e3972867f238)
  - homepage: https://sites.google.com/view/humanoid-gym/ (fetched 2026-08-29T10:36:27.756601+00:00, sha 0b8ffb7f1434)
- Data as of 2026-08-30T08:39:29.467469+00:00.