OpenCV C++ and Yolo v5

I need your help with a project that I need to do object detection. I need to detect an object in the project. As hardware, I will use the i.MX8M Plus module with an internal npu unit. That’s why I searched for the fastest and most efficient working algorithms. As can be seen in the link below, the yolov5s model seems to be more performant than the Tensorflow SSD MobileNet v2 model.
So I started researching how I can run the yolov5s model with OpenCV c++. But I couldn’t find any clear source or sample code. How can I do this?


sadly, no working support for yolov5 in the dnn module, so far,
but folks are working on it

Thank you for the answer. So, how would you recommend me to go about the project? Is there an object detection method that I can run with opencv for small hardware?

I work it with C++ and OpenCV with libtorch. Those who need it can use the links below.