We’re excited to announce the release ofAMD Schola v2, a major update to AMD’s open-source reinforcement learning plugin for Unreal® Engine 5. This release represents a significant leap forward in capabilities, performance, and ease of use for training and deploying AI agents in Unreal Engine environments.

What is AMD Schola?

AMD Schola is a toolkit that bridges the gap between Unreal Engine and popular Python-based RL frameworks like Stable Baselines 3 and Ray RLlib. Whether you’re training NPCs for games, developing robotic simulations, or exploring sim-to-real transfer, Schola provides the tools you need to create intelligent agents that learn from experience.

What’s new in AMD Schola v2?

Flexible inference architecture with agent/policy/stepper system

AMD Schola v2 introduces a powerful and flexible architecture that decouples the inference process into components for maximum flexibility and reusability. This modular design allows you to mix and match different policies, stepping strategies, and agent implementations to suit your specific needs.

Key components:

• Agent Interface - Define an agent that takes actions and makes observations.

  • UInferenceComponent - Add inference to any actor.
  • AInferencePawn - Standalone pawn-based agents.
  • AInferenceController - AI controller pattern for complex behaviors.

• Policy Interface - Plug in different inference backends, to turn observations into actions.

  • UNNEPolicy - Native ONNX inference with Unreal Engine’s Neural Network Engine.
  • UBlueprintPolicy - Custom Blueprint-based decision making.
  • Extensible interface for custom policy implementations, and new inference providers.

• Stepper Objects - Control inference execution patterns by coordinating agents and policies.

  • SimpleStepper - Synchronous, straightforward inference.
  • PipelinedStepper - Overlap inference with simulation for better throughput.
  • Build custom steppers for specialized performance requirements.

This architecture means you can easilyswitch between inference backends,optimize performance characteristics, andcompose behaviors without rewriting your agent logic. Whether you’re prototyping with Blueprints or deploying production-ready neural networks, the same agent interface works seamlessly with your chosen policy and execution strategy.

Minari dataset support

AMD Schola v2 introduces native support for theMinari dataset format, the standard for offline RL and imitation learning datasets. Minari provides a unified interface for storing and loading trajectory data, making it easier to share demonstrations and datasets across different projects and research communities.

Dynamic agent management

One of the most powerful improvements in AMD Schola v2 is robust support foragents being spawned and deleted mid-episode. Previous versions required a static set of agents throughout an episode, or a predefined spawning function to spawn agents but v2 can now handle dynamic populations seamlessly.

This enables realistic scenarios like:

• Battle Royale/Survival Games - Agents can be eliminated and removed from training without breaking the episode.
• Population Simulations - Spawn new agents based on game events or environmental triggers.
• Dynamic Team Composition - Add or remove team members on the fly.
• Procedural Scenarios - Dynamically create agents as players progress through procedurally generated content.

The system lets you manage lifecycles the way you want, simply mark the agents as terminated when they die, or start reporting observations when they spawn. This makes it much easier to build realistic, dynamic environments that mirror actual game scenarios.

Enhanced command-line interface

Training from the command line is now more intuitive than ever:

# Stable Baselines 3
scholasb3trainppo...
# Ray RLlib
scholarllibtrainppo...
# Utilities
scholacompile-proto
scholabuild-docs

The new CLI built withcyclopts provides better error messages, auto-completion support, and a more consistent interface across different RL frameworks.

Unreal Blueprints improvements

Working in Unreal Engine Blueprints is smoother than ever:

• Instanced Struct Based Objects Full Blueprint Support for all points and spaces.
• Enhanced Blueprint Utilities for spaces and points.

Updated framework support

AMD Schola v2 has been updated to support the latest versions of all major RL frameworks and libraries:

• Gymnasium - Full support for the latest Gymnasium API (1.1+).
• Ray RLlib New API Stack - Compatible with the latest Ray RLlib features and algorithms.
• Stable-Baselines3 2.x - Updated to work with the newest SB3 release.

These updates ensure you can leverage the latest features, bug fixes, and performance improvements from the RL ecosystem while training your agents in Unreal Engine.

Getting started with AMD Schola v2

Prerequisites

• Unreal® Engine 5.5+ (tested with 5.5 - 5.6)
• Python® 3.10 - 3.12
• Visual Studio® 2022 with MSVC v143 build tools (Windows®)

Installation

1. Clone or download AMD Schola v2 from the repository.
2. Copy to your project’s/Plugins folder.
3. Install the Python package:

pipinstall-e<pathtoSchola>/Resources/python[all]

4. Enable the plugin in your Unreal Engine project.

Compatibility

Additional resources, community, and support

AMD Schola is open source and we welcome contributions!

• Learn more about AMD Schola:Product page.
• GitHub:GPUOpen-LibrariesAndSDKs/Schola.
• Documentation:gpuopen.com/manuals/schola.
• Training an X-ARM 5 robotic arm with AMD Schola and Unreal Engine:Tutorial article.
• Issues & Discussions: Submit on our GitHub repository or discuss on theAMD Developer Community Discord.

Acknowledgments

AMD Schola v2 is the result of contributions from the AMD Software Technologies team, especially Tian Yue Liu, Mehdi Saeedi, and Nathan Hung, as well as the broader open-source community. Special thanks to all contributors who have helped make this release possible.

Try it today

AMD Schola v2 is available now under the MIT license. Whether you’re building the next generation of game AI, training robots, or exploring cutting-edge research, Schola provides the tools you need to succeed.

Download AMD Schola v2 and start building intelligent agents in Unreal Engine today!

Related software

AMD Schola

AMD Schola is a library for developing reinforcement learning (RL) agents in Unreal Engine and training with your favorite python-based RL Frameworks.

AMD FSR Frame Generation

AMD FSR™ Frame Generation uses state-of-the-art machine learning to generate high-quality intermediate frames, delivering ultra-smooth gameplay on AMD Radeon RX 9000 Series GPUs.

AMD FSR Upscaling

AMD FSR™ Upscaling is our cutting-edge ML-based upscaler. It delivers significant image quality improvements with reduced ghosting, better particle preservation, and superior detail.

AMD FSR Radiance Caching

Related news and technical articles

HIP Threads: GPU power for teams without GPU experts

HIP Threads lets C++ teams eliminate CPU hotspots by running familiar multithreaded code on AMD GPUs, no kernel rewrites or GPU expertise required, delivering real speedups in days.

Dyson Sphere Program - new multithreading dev log & full AMD Ryzen Threadripper PRO breakdown

Youthcat Games overhauled the Dyson Sphere Program game's multithreading system, boosting performance by up to 88% through custom core binding, dynamic task allocation, and enhanced thread synchronization.

AMD FSR™ Redstone expands with the latest AMD FSR™ SDK 2.1

AMD FSR 'Redstone' SDK 2.1 enables developers to integrate FSR features into games with easy-to-use APIs and access to advanced neural rendering technologies.

Training an X-ARM 5 robotic arm with AMD Schola and Unreal Engine

Train a robot arm with reinforcement learning in AMD Schola using Unreal® Engine, progressively increasing task complexity to adapt to changing conditions.

Introducing AMD GI-1.2 with multibounce indirect real-time rendering

AMD GI-1.2, our real-time Global Illumination solution, is available now as part of the AMD Capsaicin Framework v1.3.

AMD FidelityFX Super Resolution 4 plugin updated for Unreal Engine 5.7

Our AMD FSR 4 plugin has been updated to support Unreal® Engine 5.7, empowering you to build expansive, lifelike, and high-performance worlds.

AMD Radeon Developer Tool Suite updated with new GPU support and significant RGP enhancements

The latest version of the RDTS is designed to empower developers with enhanced capabilities for profiling, analyzing, and optimizing their GPU applications.

ONNX and DirectML execution provider guide - part 1

Learn how to optimize neural network inference on AMD hardware using the ONNX Runtime with the DirectML execution provider and DirectX 12 in the first part of our guide.

Related videos

Advancing AI in video games with AMD Schola | HTEC Days 2025 - YouTube link

Join Alexander Cann, Lead Developer at Schola, and Mehdi Saeedi, AI Lead at Schola, as they take you through the fascinating world of reinforcement learning (RL) and its transformative impact on gaming. They'll be joined by Gabor Sines, Sr. Fellow Engineer at AMD, as moderator.

Gears 5 – High-Gear Visuals On Multiple Platforms - YouTube link

This talk will discuss Direct3D® 12 in general, as well as some of the features that were leveraged to accomplish this goal, such as Async Compute, Tiled Resources, Debugging, Copy Queues, and HDR.