Datasets

Dataset

NayanaIR-CrossBench

by Cognitive-Lab

NayanaIR-CrossBench NayanaIR-CrossBench is a comprehensive multilingual multimodal document retrieval benchmark designed to evaluate cross-lingual visual search capabilities. It is part of the M3DR (Multilingual Multimodal Document Retrieval) framework.

No licenseother 210 110K < n < 100K rows

published 1 Dec 2025

Preview

First 5 rows of corpus / test, from the Hugging Face dataset viewer.

corpus-idimagedoc-id
0{"src":"https://datasets-server.huggingface.co/assets/Cognitive-Lab/NayanaIR-CrossBench/--/4b395c757c95f6481a0eb5419355b16437e54fe7/--/corpus/test/0/image/image.jpg?Expires=1792268622&Signature=JCKWTaragr0iyT3lwxClCnPEtMVNc7QUN9oZJg6V3FPJFX~cElVHGcIeXQJ7gWYftqBHbvN~He--oSrbqVMAivkN3L~GVx0-ndsExaKD9…document_15579_page_98_en
1{"src":"https://datasets-server.huggingface.co/assets/Cognitive-Lab/NayanaIR-CrossBench/--/4b395c757c95f6481a0eb5419355b16437e54fe7/--/corpus/test/1/image/image.jpg?Expires=1792268622&Signature=hBoMHmUv4b33QbP332bhEiG7U8gcuRipaCfopIqIc7JR88FZh8hqW2rjp7laO8HNzvvztX3Gn-n65LOfTkLE-u~jg9kBQMk67sqM41z6x…document_15579_page_98_it
2{"src":"https://datasets-server.huggingface.co/assets/Cognitive-Lab/NayanaIR-CrossBench/--/4b395c757c95f6481a0eb5419355b16437e54fe7/--/corpus/test/2/image/image.jpg?Expires=1792268622&Signature=rfO3RrhQ0AsjE65uEVvVubpkZw4eIBEB9jkM12uapEwsl02~27iReRCmtHGKc9F3XqZBOD-t84A-C2DX~aefbhuBNmMa-3-DcbHxHkwem…document_14755_page_99_th
3{"src":"https://datasets-server.huggingface.co/assets/Cognitive-Lab/NayanaIR-CrossBench/--/4b395c757c95f6481a0eb5419355b16437e54fe7/--/corpus/test/3/image/image.jpg?Expires=1792268622&Signature=kj2Zh6t4~L18psDrd8cXB3Y9Xl6NIdDk6UhKThRanUVVH~gKTKnmX2zdBBpzdkkjfn432zH-EHuAjydaGrACZHdXAnVF8tYewjVvPptHG…document_18066_page_98_or
4{"src":"https://datasets-server.huggingface.co/assets/Cognitive-Lab/NayanaIR-CrossBench/--/4b395c757c95f6481a0eb5419355b16437e54fe7/--/corpus/test/4/image/image.jpg?Expires=1792268622&Signature=W3aT3jGzV8e0rOfkXE8RbHW6-GTUTyuOVVF5yolPW4zUG2BPpy9VFOmpvsmPS1UV31gpBUXsGHptDEzJwIly9LVKf0zmERD7nYlldH60W…document_18066_page_98_fr

Cite this

No licenseimagetabulartextCite

Tags

language:enlanguage:bnlanguage:delanguage:eslanguage:frlanguage:hilanguage:itlanguage:knlanguage:mrlanguage:mllanguage:orlanguage:palanguage:telanguage:talanguage:salanguage:gulanguage:thlanguage:zhlanguage:jalanguage:kosize_categories:10K<n<100Kformat:parquetformat:optimized-parquetmodality:imagemodality:tabularmodality:textlibrary:datasetslibrary:dasklibrary:polarslibrary:mlcroissantarxiv:2512.03514region:us