All Tools
M
AgentsFreeOpen Source
MARLLIB
Multi-agent RL library on Ray RLlib
MIT
ABOUT
MARL research splits across algorithm repos and env-specific forks, so comparing methods is painful. MARLlib wraps popular multi-agent tasks for RLlib and ships many algorithms in one library so experiments share the same training interface.
INSTALL
pip install marllibINTEGRATION GUIDE
1. Train MAPPO, QMIX, or other MARL algorithms through a shared RLlib interface
2. Wrap SMAC, MPE, or PettingZoo tasks without maintaining per-env forks
3. Compare cooperative and competitive agents with MARLlib configs
4. Reproduce multi-agent baselines on Ray without assembling algorithm code
TAGS
pythonpytorchmarlmulti-agentrllibray