NeuralSPOT is currently available as an open source technical preview.
An open source AI developer-focused SDK called neuralSPOT is now available from Ambiq. It was created for the company’s newest Apollo4 Plus system-on-chip (SoC) family.
For the electronic devices of end customers, neuralSPOT unlocks a game-changing, multi-fold boost in energy efficiency by utilising Ambiq’s own SPOT platform. For developers to expand AI features for their applications, the SDK includes necessary software and tools. NeuralSPOT connects everything by providing libraries for interacting with sensors, managing SoC peripherals, managing power and memory configurations, tools for quickly debugging AI models from a laptop or PC, and examples.
The multibillion-dollar intelligent Internet of Things (IoT) sector now heavily relies on AI for everything from industrial anomaly detection to speech-based user interfaces. According to Ericsson, the installed base of IoT endpoints will increase from 14.6 billion devices in 2021 to 30.0 billion devices by 2027. This growth will be fueled by the adoption of IoT applications like industrial IoT (IIoT), intelligent manufacturing, healthcare, smart manufacturing, smart homes, smart buildings, and automotive.
By providing Ambiq-specific optimised libraries for audio, I2C, power and memory management, accelerated inter-process communication, and other peripherals, neuralSPOT uses a turnkey approach to help unlock the Apollo4 Plus’s highly efficient AI capabilities. Additionally, by enabling PC-based AI models to connect effortlessly with Ambiq’s embedded platforms, it offers tailored tools to simplify typical cross-platform AI development issues, such as validating a deployed model against open datasets. In order to jumpstart the development of AI features on Ambiq products, it also includes ready-to-use AI examples.
NeuralSPOT’s open source and modular design is tailored for embedded AI application developers, and it flexibly supports a variety of workflows to promote collaboration and development.