Fetching the paper…

Florence-2: Advancing a Unified Representation for a Variety of Vision Tasks · Around