Docmatix - a huge dataset for Document Visual Question Answering

In three linesHugging Face releases Docmatix, a large-scale dataset for Document Visual Question Answering (DocVQA). The dataset contains thousands of documents with visual question-answer pairs to train models for understanding and extracting information from documents.

Summary generated by Claude — human-verified