Performance of STM32N6

[NMO]

Hello, I am working on a visual inertial odometry system. It basically contains a camera, a vision processing unit and an IMU. For vision processing part I was thinking of the STM32N6. Can you elaborate on the performance of the ST Neural-ART processsor? Can you elaborate on the inference time? How many images per second can it process?

[Julian E.]

Hi ​@NMO,

 

In ST Model Zoo, you will find a lot of model architectures for different use cases.

For each one, you should find performance metrics.

 

For example, for the Yolov8n, look at the differents arrays on the page:

stm32ai-modelzoo/object_detection/yolov8n at main · STMicroelectronics/stm32ai-modelzoo

 

We are working on improving the browsing of models.

 

Also you have the ST Developper cloud where you can benchmark your models on a real N6DK board (and others): ST Edge AI Developer Cloud - STMicroelectronics - STM32 AI

 

Have a good day,

Julian