What development tools are available for STM32 Edge AI?
STM32 has a complete ecosystem of hardware and software for Edge AI, aimed at helping developers efficiently implement Edge AI applications based on STM32 MCUs and MPUs.
The STM32 Edge AI ecosystem consists of STM32 hardware platforms (MCUs and MPUs) and software development tools. A typical representative of the STM32 hardware platform is the ultra-high-performance MCU STM32N6; on the software side, it includes six free AI development tools, over 50 case studies, and more than 20 resource documents, covering the entire process from data collection, model optimization, to deployment verification, and is compatible with mainstream AI frameworks such as TensorFlow Lite, PyTorch, and ONNX, forming a closed-loop support from algorithm development to hardware implementation.
What development tools are available for STM32 Edge AI?
In Edge AI development based on STM32, what tools can assist developers in efficiently completing tasks? We introducesix distinctive core tools, each playing an important role in different development stages.
NanoEdge AI StudioThis is a free AutoML low-code tool aimed at time-series data (such as accelerometer and gyroscope data), guiding users to easily find suitable AI models that meet their needs, automatically generating optimized machine learning algorithms without deep programming, suitable for rapid development of AI applications at the node level, such as device status monitoring and motion recognition.
STM32Cube.AI (X-CUBE-AI)
X-CUBE-AI is a free STM32Cube expansion package that helps developers automatically convert pre-trained Edge AI algorithms (such as neural networks and machine learning models) into optimized STM32 C code.
AI for OpenSTLinux
X-LINUX-AI is an embedded Linux system development kit for STM32 MPUs, supporting time-series, audio, and visual data processing, integrating Linux AI frameworks and application examples, assisting developers in deploying AI models on STM32 MPUs to meet high scalability application needs.
ST Edge AI Model Zoo
The Edge AI Model Zoo is a collection of reference Edge AI models optimized for STMicroelectronics devices, providing AI models optimized for STM32 MCUs, MPUs, and other devices, covering various application scenarios such as vision and sensor data; it comes with deployment scripts that support direct execution on target devices, allowing models to be retrained based on datasets or flexibly expanded through “Bring Your Own Model (BYOM)” and “Bring Your Own Data (BYOD)” options, greatly enhancing convenience for developers; it includes STM32 model libraries, MLC model libraries, and ISPU model libraries, facilitating quick invocation by hardware type, aiding efficient development and deployment of embedded Edge AI functionalities.
ST Edge AI Core
A command-line interface (CLI) tool for optimizing and compiling Edge AI models, suitable for various ST devices including microcontrollers and microprocessors. It supports cross-framework model import of AI models; can perform algorithm analysis, model validation, and optimization; and also supports code generation and deployment.
ST Edge AI Developer Cloud
A free online platform that allows easy optimization and benchmarking of Edge AI models, suitable for various ST devices. This platform relies on Edge AI Core to execute model optimization and validation, conducting benchmarking through cloud development boards; users can directly access the platform, simplifying the Edge AI development process and enhancing model deployment efficiency.
Through the collaborative use of the above tools, developers can achieve full-link development from data collection, model training to hardware deployment within the STM32 Edge AI ecosystem, significantly lowering the technical threshold for Edge AI applications and accelerating the innovation and implementation of smart devices.
·
© THE END
What tools are used for developing Renesas MCUs?
What are Reality AI tools applied to MCUs/MPUs?
Adding the Keil AC6 compiler in the MCU integrated development environment e2 studio