han20192019/skill-composition-benchmark

Benchmarks for Compose by Focus: Scene Graph-based Atomic Skills

24
/ 100
Experimental

This project provides standardized testing environments for robot manipulation skills. It takes simulated robot actions and scene information, and outputs performance metrics for both simple, short-term actions (like picking up an object) and complex, multi-step tasks (like sorting blocks). Robotics researchers and engineers developing control policies for robots in varied environments would use this.

Use this if you are developing and evaluating robot control algorithms for complex, sequential manipulation tasks in a simulated environment.

Not ideal if you are working with physical robots or simple, non-sequential tasks, or if you need a real-world dataset rather than a simulation benchmark.

robotics robot-manipulation reinforcement-learning task-planning robot-simulation
No Package No Dependents
Maintenance 6 / 25
Adoption 5 / 25
Maturity 13 / 25
Community 0 / 25

How are scores calculated?

Stars

10

Forks

Language

Python

License

Apache-2.0

Last pushed

Jan 02, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/han20192019/skill-composition-benchmark"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.