A Framework for LLM-based Multi-Agent Reinforced Training and Inference

Train smarter: adaptive task selection that targets your model's learning frontier. Plugs directly into TRL, OpenRLHF, Axolotl, and lm-evaluation-harness.

🌐 Streamline LLM development with ready-to-use environment templates for efficient setup and deployment.

A list of uv environments templates for LLM development.