Self-play reinforcement learning AI for Agent Avenue — 17M+ games trained, C++/CUDA, evaluation suite, and playable Windows opponent.
reinforcement-learning cpp cuda game-ai actor-critic self-play multi-agent-reinforcement-learning ai-agent hidden-information-game board-game-ai card-game-ai agent-avenue
-
Updated
Sep 18, 2026 - C++