mimicking the output of a YOLO object detection model prompts

very few results