I trained a YOLOv8 nano detection model with imgsz=320 and 36 classes. When I conver it to ONNX, I get the following message:

PyTorch: starting from 'best.pt' with input shape (1, 3, 320, 320) BCHW and output shape(s) (1, 40, 2100) (5.9 MB)

What is the rational behind output of (1, 40, 2100)? I have 36 classes, so why we see the number 40 here? What is that 2100? And how are these related to the input size?

I see this output when I try to visualize with Netron:

enter image description here



You must log in to answer this question.

Browse other questions tagged .