2025-02-18 06:16 AM
I have seen a lot of videos on YouTube where STM32N6 demonstrates incredible AI features such as multiple pose estimation, multiple object detection and multiple object classification at the same time.
But I haven't seen anything similar for STM32MP257. So my question is: If I want to start with AI on embedded systems, which one of them two should I use?
2025-02-18 06:32 AM - edited 2025-02-18 06:32 AM
Well, N6 is built with AI in mind and has dedicated resources for that !
2025-02-18 06:47 AM
Can you elaborate that?
2025-02-18 07:20 AM
From STM web site STM32N6 series - STMicroelectronics :
The STM32N6 is the first STM32 MCU to embed the ST Neural-ART accelerator™, an in-house developed neural processing unit (NPU) engineered for power-efficient edge AI applications. Clocked at 1 GHz and providing up to 600 GOPS, it enables real-time neural network inference for computer vision and audio applications.
A dedicated computer vision pipeline with a MIPI CSI-2 interface and image signal processing (ISP) ensures compatibility with a wide range of cameras. The STM32N6 also features an H264 hardware encoder and the NeoChrom™ Accelerator for graphics, making it suitable for feature-rich products.
It offers 4.2 Mbytes of contiguous embedded RAM, ideal for neural networks or graphics applications, complemented by high-speed external memory interfaces (hexa-SPI, OCTOSPI, FMC).
2025-02-19 11:43 PM
Can anyone from ST show me some demonstrations for the STM32MP257 when AI is used?
2025-02-20 01:03 AM
Hello @DMårt,
Please find this small presentation video describing the AI capabilities on stm32mp257 platforms :
There, you have a description of the embedded NPU on stm32mp257: STM32MP25 NPU wiki (up to 1.35TOPS)
Application examples: https://wiki.st.com/stm32mpu/wiki/Category:AI_-_Application_examples
X-Linux-AI expansion package: https://wiki.st.com/stm32mpu/wiki/Category:X-LINUX-AI_expansion_package
Best regards,
Gatien