MicroAI

Principle

The MicroAI Foundation Library defines a low-level Machine Learning framework for embedded devices. It allows to run inferences on trained Machine Learning models.

Functional Description

A typical Application using MicroAI will load a model binary file, read its input/output characteristics, and finally perform an inference.

The MicroAI integration in a VEE port relies on a native AI framework (Tensorflow-lite, ONNX Runtime, etc…) to implement all of the above functionalities.

Dependencies

LLML_impl.h implementation (see LLML: MicroAI).
A port of MicroAI for TensorFlow Lite can be found in MicroAI Abstraction Layer for TensorFlow Lite.

Installation

MicroAI is an additional module. To enable it, the MicroAI Pack must be installed in your VEE Port:

microejPack("ej.api:microai:2.1.0")

Use

See MicroAI API chapter in Application Developer Guide.