Dataset

NayanaOCRBench_Synthetic

Name: NayanaOCRBench_Synthetic
Creator: Cognitive-Lab
License: CC-BY-NC-4.0

by Cognitive-Lab

NayanaOCRBench · Synthetic Held-out evaluation split of the NayanaOCR 2026 pipeline, 22 languages, ~770 pages per language. Every page is the same source document re-typeset in each language with the layout preserved, so the set is parallel across languages…

CC-BY-NC-4.0image-to-text 0 010K < n < 100K rows

published 17 Sept 2026

Open on Hugging Face

Preview

The dataset does not exist, or is not accessible without authentication (private or gated). Please check the spelling of the dataset name or retry with authentication.

Cite this

CC-BY-NC-4.0imagetextCite

Tags

task_categories:image-to-texttask_categories:image-text-to-textlanguage:enlanguage:hilanguage:bnlanguage:mrlanguage:gulanguage:palanguage:orlanguage:knlanguage:talanguage:telanguage:mllanguage:salanguage:zhlanguage:jalanguage:kolanguage:delanguage:frlanguage:eslanguage:itlanguage:rulanguage:arlanguage:thlicense:cc-by-nc-4.0size_categories:10K<n<100Kformat:parquetmodality:imagemodality:textlibrary:datasetslibrary:pandaslibrary:polarslibrary:mlcroissantregion:usocrdocument-aimultilingualindicbenchmarknayana