Schola V1.3.0

What's New in Schola v1.3

This release of Schola contains support for behavior cloning with the imitation library, easily compiling your environment into a standalone executable from Python and Unreal® Engine 5.6.

New Features

• Helpers for building standalone executables from Python scripts.

Bug Fixes

• Fixed a bug where the Unreal Process doesn't close when running standalone on Linux®.
• Fixed a bug with merging discrete spaces into a new MultiDiscreteSpace.
• Fixed an import bug in StatLoggerComponent.h.

Known Incompatibilities and Issues

• Models trained with Schola v1.0.0 must be re-exported to ONNX using Schola v1.1+.
• Schola v1.1+ is not compatible with Unreal® Engine 5.4.
• RLlib does not support camera observers without an additional wrapper to convert observations to channels last format.

Prerequisites

• Unreal Engine 5.5 to 5.6.
• Python 3.9 to 3.12.

Schola V1.2.0

What's New in Schola v1.2

This release of Schola contains bug fixes, support for single agent gym environments, some helpful environment wrappers and additional documentation.

New Features

• Support for Non-Vectorized Gym Environments
• Transpose Wrapper for using Image Observations with RLlib
• Pop Action Wrapper for using Non-Vectorized gym environments with Stable-Baselines3

Documentation

• Improved labeling of UPROPERTIES
• Minor improvements

Bug Fixes

• Frame Stacker no longer flattens shaped Box Spaces
• Raycast Sensor now properly rotates with characters

Known Incompatibilities and Issues

• Models trained with Schola v1.0.0 must be re-exported to ONNX using Schola v1.1+
• Schola v1.1+ is not compatible with Unreal 5.4
• RLlib does not support camera observers without an additional wrapper to convert observations to channels last format.

Schola v1.1.1

What's New in Schola v1.1.1

This minor release of Schola contains bug fixes, and additional documentation.

Bug Fixes

• Updated Unreal version in documentation
• Added Python<=3.12 to the requirements list for installing the schola python package.
• Fixed Bug causing Launch.py to not include tensorboard logging on SB3 even when enabled

Documentation

• Added Tips and Tricks Page
• Added detailed documentation on setting up inference
• Added detailed documentation on running from the CLI (i.e. headless mode)
• Updated install instructions to more clearly cover using ROCm, and avoid a few possible edge cases

Known Incompatibilities and Issues

• Models trained with Schola v1.0.0 must be re-exported to ONNX using Schola v1.1+
• Schola v1.1+ is not compatible with Unreal 5.4
• RLlib does not support camera observers without an additional wrapper to convert observations to channels last format.

Prerequisites

• Unreal Engine 5.5
• Python 3.9 to 3.12

Schola v1.1.0

What's New in Schola v1.1.0

This release of Schola contains support for Unreal 5.5, several new features, and stability improvements.

New Features

• Launch Script Extensions: You can now inject callbacks into launch.py using Python entrypoint plugins.
• Camera Sensors: Schola now includes a camera sensor, which can integrate with an existing SceneCaptureComponent2D to provide vision or depth input to an agent.
• (Beta) Behavioral Tree + RL: Use the new AgentStep BTTask to embed RL models into your behavioral trees.
• Support for Adding and Removing Agents: Agents can now be dynamically added or removed during training episodes. Inference Agents can now be dynamically added or removed from a level at any time.
• Observer/Actuator Wrappers: Observers/Actuators can now be wrapped to easily build reusable utilities. This update also includes some premade wrappers such as frame stacking, clipping, and normalization.

Tip

Check out the latest ScholaExamples for sample environments using some of these new features.

Improvements

• Improved ONNX Format: Models exported to ONNX now use a more human legible format to simplify debugging and enable name-based mapping of observers/actuators to model inputs/outputs.
• Shaped Box Spaces: Schola now supports BoxSpaces with a non-flat shape (e.g. Images of shape (3,64,64), instead of (1, 12288))
• Documentation: Python commands and scripts are now comprehensively documented with Sphinx.
• Support for Unreal 5.5

Bug Fixes

• Fixed several script arguments not working correctly.

Known Incompatibilities and Issues

• Models trained with Schola v1.0.0 must be re-exported to ONNX using Schola v1.1.
• Schola v1.1 is not compatible with Unreal 5.4
• RLlib does not support camera observers without an additional wrapper to convert observations to channels last format.

Prerequisites

• Unreal Engine 5.5
• Python 3.9 or later

Schola v1.0.0

Introducing Schola v1.0.0

Welcome to the Schola SDK. The initial release of Schola features tools for building reinforcement learning environments in Unreal Engine and training with your favorite python-based RL Frameworks (Gym, RLlib and Stable Baselines 3).

Features

Inference

Schola provides tools for connecting and controlling agents with ONNX models inside Unreal Engine. Allowing for inference with or without Python.

Simple Interfaces in Unreal

Schola exposes simple interfaces in Unreal Engine for the user to implement. Allowing you to quickly build and develop reinforcement learning environments.

Headless Training

Run training without rendering to significantly improve training throughput.

Multi-Agent Training

Train multiple agents to compete against each other at the same time using RLlib and multi-agent environments built with Schola.

Vectorized Training

Run multiple copies of your environment within the same Unreal Engine process to accelerate training.

Reusable Actions and Sensors

Schola supports building reusable sensors and actuators, to quickly design new agents from existing components.

Documentation

Comprehensive interactive html documentation including guides, API Documentation, and installation instructions can be found at GPUOpen.

Prerequisites

• Unreal Engine 5.4

• Python 3.9+