Usage tips
OWL-ViT is a zero-shot text-conditioned object detection model.