GitHub

Scaling Agentic Reinforcement Learning with a Multi-Turn, Multi-Task Framework

For a minimal example of how to use the environment framework, refer to examples/simple-calculator. For the environment and training data used in our paper, see ...