alpc91/SGRL

[ICML 2023 Oral] Official environments and implementations for "Subequivariant Graph Reinforcement Learning in 3D Environments"

27
/ 100
Experimental

This project provides an environment and models to train AI agents to navigate and control physically simulated robots like hoppers, walkers, humanoids, and cheetahs in complex 3D spaces. It takes descriptions of robot morphologies and desired behaviors, then outputs trained control policies that can adapt to different robot designs and starting conditions. Robotics researchers and control engineers would use this to develop advanced locomotion and manipulation algorithms.

No commits in the last 6 months.

Use this if you need to train reinforcement learning policies for 3D robotic agents that can generalize across various physical configurations and initial orientations.

Not ideal if your primary goal is 2D simulation or if you require an off-the-shelf, pre-trained policy for a specific robot without further research.

robotics reinforcement-learning 3D-simulation robot-locomotion control-engineering
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 6 / 25
Maturity 16 / 25
Community 5 / 25

How are scores calculated?

Stars

19

Forks

1

Language

Python

License

MIT

Last pushed

Jul 24, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/alpc91/SGRL"

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