# ZiwenZhuang/parkour

[CoRL 2023] Robot Parkour Learning

Repository: https://github.com/ZiwenZhuang/parkour
Canonical: https://ross.abutalabs.com/products/parkour
Homepage: https://robot-parkour.github.io
Language: Python
License: MIT
License Family: permissive
Last push: 2025-10-26T00:01:50+00:00

## Health v2 (maintenance only)
Score: 50/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 48, release rhythm 35, longevity 79
- inputs: {"age_days": 1106, "days_push": 312, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1114, forks 151 (observed 2026-08-28T04:03:38.199720+00:00)

## What it is
Official code for 'Robot Parkour Learning' (CoRL 2023), a reinforcement learning system that trains quadrupedal robots to perform vision-based parkour skills like climbing, leaping, and crawling. It provides IsaacGym simulation environments, RL training and distillation code, and deployment instructions for Unitree Go1 and Go2 robots.

## Use cases
- train quadruped robots to climb and jump obstacles in simulation
- learn vision-based locomotion policies with reinforcement learning
- deploy RL-trained policies on a Unitree Go1 or Go2 robot
- generate parkour skills without reference motion data
- distill privileged RL policies into depth-camera-based policies
- research sim-to-real transfer for legged robots

## When to choose
- you want to reproduce or extend the CoRL 2023 robot parkour results
- you need IsaacGym-based training environments for legged locomotion
- you own a Unitree Go1/Go2 and want to run vision-based parkour policies
- you are researching RL skill generation and policy distillation for quadrupeds

## When to avoid
- you need a production-ready robotics framework rather than research code
- you work with non-quadruped robots or humanoids
- you lack a GPU or NVIDIA IsaacGym-compatible setup
- you need plug-and-play deployment without simulation training

## Facets
- artifact type: library
- maturity: active
- function: reinforcement-learning, simulation, robotics, machine-learning
- domain: robotics, reinforcement-learning, machine-learning, simulation
- platform: python
- tags: legged-locomotion, quadruped-robots, isaacgym, sim-to-real, parkour, unitree-go1, unitree-go2, depth-camera, policy-distillation, research-code, linux, gpu

## Member repositories
- ZiwenZhuang/parkour (main) score 50

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:38.199720+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 2026-08-30T06:42:13.338296+00:00, confidence not recorded.
  - readme: https://github.com/ZiwenZhuang/parkour (fetched 2026-08-28T04:03:38.199720+00:00, sha e84d664e304c)
  - homepage: https://robot-parkour.github.io (fetched 2026-08-29T12:46:24.699249+00:00, sha 2fc17e9370f1)
- Data as of 2026-08-30T08:39:29.467469+00:00.