Getting Started With Schola Examples

This guide will walk you through setting up the Schola Experiments Project to try out some premade reinforcement learning examples with Schola.

Install Prerequisites

LinuxWindows

• Ubuntu 22.04. (22.04.4 Desktop x86 64-bit is recommended for reproducibility)

• Unreal Engine 5.5. (5.5.2 is recommended for reproducibility)

• Python 3.9 to 3.12

While other Integrated Development Environments (IDEs) can be used, Visual Studio Code is recommended along with the C/C++ Extension Pack. Please see the Unreal Engine documentation for more information.

• Windows 11 or Windows 10

• Unreal Engine 5.5. (5.5.2 is recommended for reproducibility)

• Python 3.9 to 3.12

While other Integrated Development Environments (IDEs) can be used, Visual Studio is recommended. (Visual Studio Professional 2022 (64-bit) - LTSC 17.8 is recommended for reproducibility)

Follow the Visual Studio setup guide from Epic Games to setup your Visual Studio environment for working with Unreal.

Install Schola Examples

1. Get the ScholaExamples project, either by downloading the source, or via git.

2. Make sure pip is updated before installing the schola python package from Plugins/Schola/Resources/python.

python -m pip install --upgrade pip
pip install ./Plugins/Schola/Resources/python[all]

Schola installs the cpu version of pytorch by default, to install other versions of pytorch follow the instructions at Pytorch Get Started.

To install Pytorch with ROCm on Linux, we recommend following the guide at Install Pytorch for Radeon GPUs.

3. Launch the Project.

For setting up your project files to run ScholaExample, please refer to the guide on Project Files for IDEs from Epic Games.