what is the recommended input size and number of pyramid levels for general object detection?
what is the recommended input size and number of pyramid levels for general object detection?