Microsoft Releases Florence-2: A Novel Vision Foundation Model with a Unified Cue-Based Representation for a Variety of Computer Vision and Vision Language Tasks 06/21/2024