API hosted RL environments for multi-turn/multi-agent interaction with LLMs.
pip install rl-environments
These environments can be plugged into:
- verl (original intent), see
examples/smallest_box_verl.py. - trl via verifiers, see
examples/smallest_box_trl.py.
For inquiries, please reach out at corefranciscopark@g.harvard.edu.