StanfordVL/GibsonEnv

Gibson Environments: Real-World Perception for Embodied Agents

49
/ 100
Emerging

This project offers a realistic virtual environment for training AI models in embodied perception and sensorimotor control. It takes real-world 3D spaces as input and simulates an agent's interaction within them, outputting improved AI models capable of complex visual understanding and physical navigation. This tool is designed for AI researchers and robotics engineers developing active perception systems.

936 stars. No commits in the last 6 months.

Use this if you need to efficiently train AI agents to perceive and move realistically within complex, real-world environments without the fragility and cost of physical robots.

Not ideal if your primary goal is developing AI for tasks that don't involve physical embodiment or require learning solely from abstract data representations.

robotics simulation AI training embodied AI sensorimotor learning perceptual robotics
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 23 / 25

How are scores calculated?

Stars

936

Forks

149

Language

C

License

MIT

Last pushed

Apr 15, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/computer-vision/StanfordVL/GibsonEnv"

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