RealDoc-Bench and RealDoc-Bench-Layout — documents under their owners' terms

Source: https://huggingface.co/datasets/Extend-AI/RealDoc-Bench
        https://huggingface.co/datasets/Extend-AI/RealDoc-Bench-Layout
Collected by: Extend AI.
Suites: realdoc_pdfs (RealDoc-Bench), realdoc_pages (RealDoc-Bench-Layout), the rotated copies of
realdoc_pdfs pages in rotation_skew, and the synth_copies pages rebuilt from either.

Upstream licenses its annotations under CC BY 4.0; those annotations are not used here, and this
benchmark's tests were made for it. The documents and page images are not covered by that license:
they keep their owners' rights. They are publicly available documents, credited here to the
collection they come from.
