AMD Schola v2 is a major update to the open-source reinforcement learning plugin for Unreal® Engine 5, offering significant improvements in capabilities, performance, and ease of use for training and deploying AI agents.

AMD GPUOpen

AMD Schola v2 is an open-source reinforcement learning plugin for Unreal Engine 5 that enables training and deploying AI agents. The major update introduces a flexible inference architecture with modular agent/policy/stepper components, native Minari dataset support for offline RL, dynamic agent spawning and deletion during episodes, an enhanced command-line interface, improved Blueprint integration, and updated support for Gymnasium 1.1+, Ray RLlib, and Stable-Baselines3 2.x. The plugin bridges Unreal Engine with Python-based RL frameworks, supporting use cases from game NPC training to robotic simulations.

Announcing AMD Schola v2: Next-generation reinforcement learning for Unreal Engine

Additional resources, community, and support