Skip to content Skip to footer

Researchers from HuggingFace present Docmatix, a dataset designed for visual question answering in documents, featuring 2.4 million images and 9.5 million question/answer pairs.

Leave a comment

0.0/5