☝ Click on the top “STMicroelectronics China” to follow us
-
The mainstream MCU model library in the market has been expanded to accelerate embedded AI development
-
Over 140 models to speed up the deployment of visual, audio, and perceptual AI applications on edge devices applications
Recently, STMicroelectronics (ST) released new AI models and enhanced support for STM32 AI model library in development projects to accelerate the prototyping and product development of embedded AI applications.This marks another increase in the STM32 AI model library, which is already among the best in the industry for models that can be embedded in wearable devices, smart cameras, smart sensors, security monitoring devices, safety protection devices, and robots for visual, audio, and perceptual applications.
☞ Stephane Henry, Vice President of Edge AI Solutions at STMicroelectronics, stated
Transforming data science into practical applications suitable for embedded platforms is a complex engineering challenge, and throughout the process, developers need technical support. We are increasing the variety of available models in the library to help the STM32 developer community quickly kickstart projects, while STM32 AI model library 4.0 accelerates the entire development process from model to application deployment, which is one of our commitments to realizing physical AI.
Embedding AI into everyday accessories, home appliances, and other electronic devices not only unlocks the transformative advantages of this technology but also improves efficiency and saves energy. The core components of these devices are small microcontrollers, which have limited processing power and storage capacity, making it challenging for product developers to optimize AI models while balancing performance and energy efficiency. The newly launched AI model library from STMicroelectronics helps designers maximize the use of existing resources to create efficient and low-power models. This model library is a component of the ST Edge AI suite, which provides developers with a complete set of development tools, software libraries, and utilities to further simplify and accelerate the development and deployment of AI algorithms on STMicroelectronics hardware, ensuring seamless integration of AI from prototype to production.
For over a decade, STMicroelectronics has been at the forefront of edge AI research and innovation, dedicated to providing hardware-accelerated models that help developers tackle the complexities of deploying AI at the network edge. Today, STMicroelectronics’ AI tools continue to support over 160,000 projects annually.
STMicroelectronics’ STM32 microcontrollers are widely used in various fields, including consumer electronics, wearable devices, communication infrastructure, smart grids, smart cities, industrial automation, and even low Earth orbit satellites. By strategically deploying AI on these microcontrollers, STMicroelectronics can quickly provide end-users with cost-effective cutting-edge technology while enhancing product sustainability.
To further consolidate its leading position in the embedded AI field, STMicroelectronics has expanded its product lineup, launching microcontrollers like the STM32N6 series with NPU hardware accelerators to seize opportunities in the rapidly growing embedded AI or edge AI market.
Technical Information
-
The number of models has doubled from 30 to 60, boasting the industry’s largest ready-to-use MCU model family, with over 140 models.
-
Not limited to a single model directory, it also provides a full-process development solution, where scripts can assist in training models and integrating them with application software libraries, helping customers deploy models on MCUs more efficiently.
-
New native support for PyTorch models has been added, enhancing existing support for TensorFlow Lite, Keras AI frameworks, LiteRT, and ONNX formats.
-
Improved execution efficiency of embedded AI through model compression and sub-byte quantization.
-
For more product details, please visit:
https://github.com/STMicroelectronics/stm32ai-modelzoo/
ENDRelated Articles
▷ STMicroelectronics x Autotrak: Edge AI for Distracted Driving Prevention, Ensuring Road Safety
▷ STMicroelectronics STM32N6: 800MHz Clock Speed, 600GOPS Computing Power Creates a Strong Engine for Edge AI
▷ STM32N6 “Breaks the Game” in AIoT: How One Chip Can Ignite a Trillion-Dollar Market for Smart Glasses, Robots, and Home Appliances?
Click to read the original article for more information