leggedrobotics/legged_gym

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

3102 stars · 584 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

A Python library providing Isaac Gym environments for training legged robots (e.g., ANYmal) to walk on rough terrain with reinforcement learning. It includes sim-to-real transfer components such as actuator networks, domain randomization, noisy observations, and random pushes during training.

Use cases

• train a quadruped robot to walk on rough terrain with reinforcement learning
• run PPO training for legged locomotion in Isaac Gym
• set up sim-to-real transfer with friction and mass randomization
• create custom robot environments by extending the legged robot config classes
• reproduce the ANYmal rough-terrain locomotion results from the paper

When to choose

• you want to train legged robot locomotion policies with NVIDIA Isaac Gym
• you need a proven sim-to-real pipeline for quadruped robots
• you want a configurable, inheritance-based RL environment framework for robotics research

When to avoid

• you are starting a new project - the maintainers recommend migrating to Isaac Lab
• you need a simulator other than Isaac Gym
• you require long-term support or frequent updates, as the repo receives limited maintenance

Facets

library · maturity maintenance

reinforcement-learning simulation machine-learning robotics robotics reinforcement-learning simulation machine-learning python isaac-gym legged-robots sim-to-real locomotion ppo anymal robot-learning linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="leggedrobotics/legged_gym")