hijkzzz/pymarl2
PyMARL2 provides optimized multi-agent RL algorithm implementations achieving near-perfect win rates on SMAC scenarios.
Not currently ranked — collecting fresh signals.
star history
This repository offers fine-tuned implementations of cooperative multi-agent reinforcement learning algorithms including QMIX, VDN, and MAPPO variants. It focuses on StarCraft Multi-agent Challenge (SMAC) as the primary benchmark, incorporating implementation tricks such as value function clipping, orthogonal initialization, n-step returns, and large batch training. The project also supports Google Football environments as an alternative multi-agent task.
Frequently asked
- What is hijkzzz/pymarl2?
- PyMARL2 provides optimized multi-agent RL algorithm implementations achieving near-perfect win rates on SMAC scenarios.
- Is pymarl2 open source?
- Yes — hijkzzz/pymarl2 is open source, released under the Apache-2.0 license.
- What language is pymarl2 written in?
- hijkzzz/pymarl2 is primarily written in Python.
- How popular is pymarl2?
- hijkzzz/pymarl2 has 709 stars on GitHub.
- Where can I find pymarl2?
- hijkzzz/pymarl2 is on GitHub at https://github.com/hijkzzz/pymarl2.