Florence 2

Microsoft open unified vision foundation model for captioning, detection, segmentation and OCR.