How to do for yolov8n in STM32n6

[yiulsup]

Dear ST,

i convert my tflite of yolov8n without nms to neural-art related files in STM32N6 using stedgeai cli command as below

after that, i have to apply it to cubeISE which import NUCLEO-N657X0-Q object detection example.

i have a below question to be answered from you.

1. what is ll_aton? which looks like a runtime such as cuda runtime for nvidia gpu. any link for detailed?

2. please describe the overall on how to apply my weight with fine tunning from yolov8n to STM32N6 neural-art.

thanks 

[Julian E.]

Hello @yiulsup,

To work with a Yolov8n and the N6, I would suggest you look at this document:

stm32ai-modelzoo-services/object_detection/deployment/doc/tuto/How_to_deploy_yolov8_yolov5_object_detection.md at main · STMicroelectronics/stm32ai-modelzoo-services · GitHub

Concerning your questions:

1. The ll_aton library is related to the use of the NPU on the N6. So yes, I think your comparison is right.

You can find documentation about it here: https://stedgeai-dc.st.com/assets/embedded-docs/stneuralart_programming_model.html

2. I think the first document I linked should answer your question. You first retrain your model on your usecase, then quantize it, use the st edge ai core to convert it to a C model and build your application around it (camera pipeline, preprocessing, postprocessing and display if you need).

Have a good day,

Julian