AI & Machine Learning
★ 9.0k
Python
OpenPipe/ART
9.0k
Stars
768
Forks
101
Issues
Python
Language
Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen3.5, GPT-OSS, Llama, and more!
View on GitHub
git clone https://github.com/OpenPipe/ART.git
Quick Start Example
bash
# OpenPipe/ART
git clone https://github.com/OpenPipe/ART
cd ART
# See README for setup instructions