ML//multimodal//YOLO

YOLO (You Only Look Once) is a family of real-time object detection models that find every object in an image in a single forward pass, with no region proposals. It is at heart a CNN, detecting in one shot from convolutional feature maps, though newer versions incorporate transformer components. It comes from Redmon et al. (2016) and is now up to YOLOv11, with the lineage fragmented across maintainers after v5.

YOLO (You Only Look Once) is a family of real-time object detection models that find every object in an image in a single forward pass, with no region proposals. It is at heart a CNN, detecting in one shot from convolutional feature maps, though newer versions incorporate transformer components. It comes from Redmon et al. (2016) and is now up to YOLOv11, with the lineage fragmented across maintainers after v5.

It is still the default for real-time production (robotics, surveillance, autonomous driving): DETR (transformer-based) catches up academically, but YOLO dominates deployment.