OpenPipe Agent Reinforcement Training (ART) is a library for fine-tuning and training LLM agents using reinforcement learning. Version 0.5.18 supports Python >=3.11. It provides tools for supervised fine-tuning, RLHF, and evaluation of agentic workflows. The library is relatively new with active development.
pip install openpipe-artNo compatibility data collected yet for this library.
Verified import paths — ran on the pinned version, not inferred.
Initialize ART client, create trainer, add data, and train. Requires OPENPIPE_API_KEY environment variable.
Use import openpipe_art instead of openpipe.art.
Upgrade Python to 3.11 or higher.
Use 'art run --config config.yaml' instead of 'art train ...'.
No resource links recorded.