chutes

rednote-hilab/dots.ocr

Dots.ocr bf16

dots.ocr is a 1.7B parameter vision-language model that achieves state-of-the-art multilingual document parsing including layout detection, text recognition, table extraction, and reading order preservation.