RoboticsData/score_lerobot_episodes

A lightweight toolkit for quantitatively scoring LeRobot episodes.

42
/ 100
Emerging

This toolkit helps robotics engineers and researchers evaluate the quality of their robot demonstration videos, known as LeRobot episodes. It takes your raw robot interaction videos and, using computer vision and optional AI analysis, produces a quality score (from 0 to 1) for each episode. This score reflects visual clarity, motion smoothness, collision detection, and task success, allowing you to filter out poor-quality data before training your robot models.

Use this if you need to automatically assess and improve the quality of your robot demonstration datasets to enhance the performance of your robot learning models.

Not ideal if your datasets are not in the LeRobot episode format or if you primarily need real-time scoring during data capture rather than post-collection analysis.

robot-learning robotics-data-curation demonstration-learning robot-performance dataset-quality
No Package No Dependents
Maintenance 10 / 25
Adoption 8 / 25
Maturity 15 / 25
Community 9 / 25

How are scores calculated?

Stars

49

Forks

4

Language

Python

License

Apache-2.0

Last pushed

Mar 13, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/RoboticsData/score_lerobot_episodes"

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