bytedance/GR-MG

Official implementation of GR-MG

35
/ 100
Emerging

This project helps robotics researchers and engineers train robots to perform complex manipulation tasks more effectively, especially when full instructional data is scarce. It takes partially annotated video demonstrations and specific task goals (like an image of the desired outcome) as input, then generates robust robot control policies that can achieve those goals. It is designed for researchers in robotics, automation, and machine learning focused on real-world robot task learning.

No commits in the last 6 months.

Use this if you are a robotics researcher working on teaching robots new manipulation skills with limited complete demonstration data and want to generate goal-conditioned policies.

Not ideal if you are a hobbyist or someone looking for a plug-and-play solution for off-the-shelf robots, as it requires significant technical setup and expertise in machine learning and robotics.

robot-manipulation robot-learning partially-annotated-data goal-conditioned-robotics robot-control
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 10 / 25

How are scores calculated?

Stars

93

Forks

7

Language

Python

License

Apache-2.0

Last pushed

Jan 12, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/bytedance/GR-MG"

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