Machine learning software module MPLAB®

for productionAI-assistedplug-in

Overview
MPLAB® Machine Learning Development Suite Model Builder enables engineers to design, train and convert compact machine learning models for tiny-edge embedded IoT sensor recognition. Delivered as a Visual Studio Code extension (via MPLAB for VS Code) and as a plug-in for MPLAB® X IDE, the Model Builder outputs optimized inference code and firmware artifacts targeted to Microchip MCUs, dsPIC® DSCs and MPUs.

Key capabilities

• Build compact supervised and anomaly-detection algorithms suited for resource-constrained edge devices (MCUs, dsPIC DSCs, MPUs)
• Automate the ML workflow: data preparation, feature extraction, training, validation and model conversion to firmware
• Integrated data labeling and filtering tools for dataset preparation
• AutoML-assisted prototype code generation to accelerate development timelines
• Parameter tuning options for advanced ML developers and embedded engineers
• Pipeline creation for production-ready ML deployments
• Generate production-grade firmware compatible with 8-, 16- and 32-bit devices and Microchip XC toolchains

Documentation and delivery
Comprehensive Getting Started materials and supporting documentation (including a Getting Started PDF) are provided. The Model Builder is distributed as editor extensions and IDE plug-ins that integrate with common Microchip development toolchains and generate firmware-ready code and deployment artifacts.

Usage scenarios

• Edge sensor recognition (voice/activity/motion and anomaly detection) on constrained hardware
• Rapid prototyping of ML-enabled embedded products using AutoML-generated code
• Production firmware generation for MCU/MPU-based products with support for Microchip XC compilers

Technical specifications

• Delivery modes: VS Code extension (MPLAB for VS Code) and MPLAB X IDE plug-in
• Target devices: 8-, 16-, 32-bit MCUs, dsPIC® DSCs and MPUs
• Supported ML types: supervised learning and anomaly detection optimized for tiny-edge
• Automated workflow: data prep, feature extraction, training, validation and model conversion
• AutoML features to accelerate prototype code generation
• Outputs: optimized inference code, firmware-ready artifacts and compatibility with Microchip XC compilers
• Includes documentation and a Getting Started PDF guide