dots.ocr: Multilingual Document Layout Parsing in a Single Vision-Language Model - Explained Simply | ArXiv Explained