The ML-Agents toolkit allows you to use pre-trained neural network models inside your Unity games. This support is possible thanks to the Unity Inference Engine. The Unity Inference Engine is using compute shaders to run the neural network within Unity.
Scripting Backends : The Unity Inference Engine is generally faster with IL2CPP than with Mono for Standalone builds. In the Editor, It is not possible to use the Unity Inference Engine with GPU device selected when Editor Graphics Emulation is set to OpenGL(ES) 3.0 or 2.0 emulation. Also there might be non-fatal build time errors when target platform includes Graphics API that does not support Unity Compute Shaders. The Unity Inference Engine supposedly works on any Unity supported platform but we only tested for the following platforms :
- Linux 64 bits
- Mac OS X 64 bits (
OpenGLCore
Graphics API is not supported) - Windows 64 bits
- iOS
- Android
When using a Learning Brain, drag the .nn
file into the Model field
in the Inspector.
Uncheck the Control
checkbox for the corresponding Brain in the
BroadcastHub of the Academy.
Select the Inference Device : CPU or GPU you want to use for Inference.
Note: For most of the models generated with the ML-Agents toolkit, CPU will be faster than GPU. Only use GPU if you have a large number of agents using visual observations.