Edge ML Support Integrated Development Kit
Achieving Security and Power Efficiency with ML

Microchip supports embedded system designers with SW tools that facilitate the development of machine learning models.

Microchip Technology Inc. today announced the availability of the MPLAB® Machine Learning Development Suite, an integrated workflow that enables simplified machine learning model development.

This software tool kit can be utilized across Microchip's MCU and MPU product families. It is expected that embedded engineers will be able to implement machine learning inference tasks more quickly and efficiently.

“Machine learning is becoming the new standard for embedded controllers, and leveraging machine learning at the edge enables the creation of products that are more secure and power-efficient than those that rely on cloud communications,” said Rodger Richey, director of Microchip’s development systems business unit. “This integrated solution supports 32-bit MCUs and MPUs, as well as, for the first time, 8-bit and 16-bit devices, enabling efficient product development.”

Microchip said it will help "more accurately predict potential problems in equipment used in industrial, manufacturing, consumer and automotive applications."

The MPLAB® Machine Learning Development Suite enables engineers to build highly efficient and space-saving machine learning models. The solution’s toolkit is also based on Automatic Machine Learning (AutoML), which reduces extraction, training, validation, and testing times, thereby reducing costs. Models can be optimized to take into account the memory constraints of MCUs and MPUs.

This newly developed development suite can be used with the MPLAB X Integrated Development Environment (IDE). This allows even developers without machine learning programming expertise to easily build solutions.

Microchip also offers the option to import models from TensorFlow Lite and use them in MPLAB Harmony v3 projects.

MPLAB Harmony v3 is an integrated embedded software development framework that provides flexible, interoperable software modules.

The VectorBlox™ Accelerator Software Development Kit (SDK) enables the most power-efficient execution of convolutional neural network (CNN)-based artificial intelligence/machine learning (AI/ML) inference workloads when used with PolarFire® FPGAs.