Hierarchical Deep CFR

GitHub Link

Codebase for my paper: Hierarchical Deep Counterfactual Regret Minimization

Language: Python

The following parts are included:

• Hold’em Poker benchmarks built with PokerRL, including Leduc, FHP, HULH.
• An implementation of the hierarchical strategy learning algorithm proposed in our paper and its ablations.
• Implementations of the SOTA model-free algorithms for zero-sum imperfect information games, including DREAM, NFSP, DCFR.

The codebase is huge and so understanding it may take time.