csc301-2026-f/20-amd-schola

Schola is an Unreal Engine toolkit for defining agent environments, exchanging observations and actions with Python reinforcement-learning frameworks, and running trained policies in games. The principal loop runs from Unreal environment and agent definitions through Gym/gRPC communication to Python training or evaluation, then returns actions to advance the environment. Separate branches support in-game inference and imitation-data collection. README and tree establish these product capabilities; sampled implementation detail is limited, so the graph omits unverified internal call links.