STM32 AI Model Zoo:
- Latest extensions to industry’s largest library of models for market-leading MCU family accelerate embedded AI development
- More than 140 models for vision, audio, and sensing AI applications at the edge
STMicroelectronics, a global semiconductor leader serving customers across the spectrum of electronics applications, has unveiled new models and enhanced project support for its STM32 AI Model Zoo to accelerate the prototyping and development of embedded AI applications. This marks a significant expansion for what is already the industry’s largest library of models for vision, audio, and sensing to be embedded in equipment such as wearables, smart cameras and sensors, security and safety devices, and robotics.
“Turning data science into a working application tuned for an embedded platform is a complex engineering challenge, and developers need support throughout the journey,” said Stephane Henry, Edge AI Solution Group VP at STMicroelectronics. “While expanding the selection of models available, to help the STM32 developer community jump-start their projects, we are also strengthening the infrastructure all the way to deployment with STM32 AI Model Zoo 4.0. This is part of our commitment to make Physical AI a reality.”
Embedding AI into everyday accessories, appliances, and other electronic equipment unlocks the technology’s transformative benefits while increasing efficiency and saving energy. At the core of these devices lie compact microcontrollers, with limited processing and memory by design, challenging product developers to optimize AI models for both performance with efficiency. ST’s latest AI Model Zoo empowers designers to maximize available resources, enabling the creation of highly efficient models that operate with minimal power consumption.
This Model Zoo is part of ST Edge AI Suite, which offers a comprehensive collection of tools, libraries, and utilities that further simplify and accelerate the development and deployment of AI algorithms on ST hardware, ensuring seamless integration from prototype to production.
For over a decade, ST has been at the forefront of research, innovation, and development in edge AI, with the goal of helping developers overcome the complexities of deploying AI at the edge with both software and hardware accelerated models. Today, ST’s AI tools continue to support over 160,000 projects annually.
ST’s STM32 family features the world’s most widely adopted microcontrollers, used in a diverse range of applications, including consumer appliances, wearables, communication infrastructure, smart grids, smart cities, industrial automation, and even low-earth-orbit satellites. By strategically enabling AI deployment on general-purpose MCUs across these sectors, ST delivers cutting-edge technology to end users rapidly and cost-effectively, while enhancing sustainability.
Further advancing its leadership in embedded AI, ST expands its portfolio with AI-accelerated MCUs, such as the STM32N6 series. This positions ST at the forefront of the fast-growing embedded AI, or edge AI market.
Technical information
- Doubles the number of model families from 30 to 60, representing the industry’s largest collection of ready-made models for MCUs with more than 140 models
- Goes beyond a mere model catalog to a full workflow solution with scripts to assist training and integrating models with application libraries for optimal performance and efficiency
- Adds native support for PyTorch models, complementing the existing support for TensorFlow Lite, Keras AI frameworks, LiteRT, and ONNX formats
- Increases embedded AI efficiency with compressed and sub-byte quantized models
- Available as a standalone solution at: https://github.com/STMicroelectronics/stm32ai-modelzoo/
