inspirai/TimeChamber
A Massively Parallel Large Scale Self-Play Framework
This project helps robotics researchers and AI developers efficiently train physically simulated agents to learn complex competitive behaviors. By leveraging parallel simulation on a single GPU, it can quickly train agents, such as humanoids with swords or multi-legged 'ants,' to compete in virtual arenas. The input is a 3D simulated environment and the desired competitive task, and the output is a highly skilled agent policy that can outperform others in combat or strategy.
361 stars. No commits in the last 6 months.
Use this if you are developing AI for multi-agent competitive robotics tasks and need to rapidly train and evaluate policies with limited GPU resources.
Not ideal if your agents are not physically simulated in a 3D environment or if your primary focus is on non-competitive, single-agent tasks.
Stars
361
Forks
38
Language
Python
License
MIT
Category
Last pushed
Jan 09, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/agents/inspirai/TimeChamber"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
google-deepmind/concordia
A library for generative social simulation
Mai-xiyu/Minecraft_AI
AI Play Minecraft
mikelma/craftium
A framework for creating rich, 3D, Minecraft-like single and multi-agent environments for AI...
cocacola-lab/MineLand
Simulating Large-Scale Multi-Agent Interactions with Limited Multimodal Senses and Physical Needs
Dominien/brunnfeld-agentic-world
Medieval village economy simulation — up to 1000 LLM agents across multiple villages, driven by...