Microsoft has released a set of vision models called Florence 2 is a prompt-based vision model designed for computer vision and image processing tasks such as image description, object recognition, localization, and segmentation.