unitreerobotics/unifolm-world-model-action

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

1161 stars · 146 forks observed · 2026-10-02

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

UnifoLM-WMA-0 is Unitree's open-source world-model-action framework for general-purpose robot learning across multiple robotic embodiments. Its world model serves as an interactive simulation engine for synthetic data generation and enhances action policies by predicting future robot-environment interactions.

Use cases

• train a world model for robot manipulation
• generate synthetic training data for robot policies
• improve robot action decision-making with future prediction
• deploy learned policies on Unitree robot arms
• fine-tune a foundation model on Open-X robotics datasets

When to choose

• you work with Unitree robots or similar embodiments
• you need a world model to simulate robot-environment interactions
• you want to combine video prediction with action policy learning

When to avoid

• you need a lightweight general robotics simulator like MuJoCo or Isaac
• you have no GPU resources for training large models
• your project does not involve robot learning

Facets

library · maturity active

machine-learning deep-learning simulation llm-training robotics machine-learning artificial-intelligence autonomous-vehicles python world-model robot-learning robot-policy unitree video-prediction imitation-learning linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="unitreerobotics/unifolm-world-model-action")