CFR Poker AI Source Code

MasterAI is a research codebase for heads-up no-limit Texas Hold'em (HUNL), focused on counterfactual regret minimization, self-play, counterfactual value estimation and strategy evaluation.

View source on GitHubResponsible use

Repository-backed research components

CFR and game trees

csrc/ contains performance-oriented computation and search components.

Counterfactual value

cfv/ holds counterfactual value related implementations.

Training strategies

rela/ and supervised_strategy/ provide learning components.

Evaluation

benchmark/ and tests/ support experiments and validation.

Research interface screenshots

MasterAI CFR poker AI research interface
Poker strategy research interface
Texas Holdem AI strategy analysis
Strategy analysis and state information
HUNL poker AI experiment dashboard
Research and experiment view

Scope and responsible evaluation

This repository targets game theory, imperfect-information games and offline strategy experiments. Verify dependencies, configurations, seeds and benchmark protocols before drawing conclusions. Do not use it for real-time assistance, automated play on third-party services or rule circumvention.