Datasets
Dataset
indic-ocr-bench
by sarvamai
Sarvam Indic OCR Bench Global benchmarks focus heavily on English document parsing, and to the best of our knowledge there is no Indic OCR benchmark of comparable breadth and rigor. Sarvam Indic OCR Bench fills this gap with 6,909 curated text-block samples…
Apache-2.0image-to-text 20 11K < n < 10K rows
published 25 May 2026
Preview
First 5 rows of default / test, from the Hugging Face dataset viewer.
| image | image_name | gt | language |
|---|---|---|---|
| {"src":"https://datasets-server.huggingface.co/assets/sarvamai/indic-ocr-bench/--/84ce7ce447456a92bcbf25f3c0a55d6a5a44a24b/--/default/test/0/image/image.jpg?Expires=1790023260&Signature=e-wF4pq-HLGs-7c0DHpADKYehxTWjcvhvnmlff4dQSbyvrfA66Y57wccga-aZu7jBi7VjfUZJc2fSR4873Gf--ZpC895i~bKUHYomAsOXtB3svG8m… | indic_ocr_bench_test_ben_119 | উৎৎড়ৎ! টহশহড়হি ংরিঃপয ধৎমঁসবহঃ. | Bengali |
| {"src":"https://datasets-server.huggingface.co/assets/sarvamai/indic-ocr-bench/--/84ce7ce447456a92bcbf25f3c0a55d6a5a44a24b/--/default/test/1/image/image.jpg?Expires=1790023260&Signature=A4-aOlHktAzMuK5BHl38MygxkZ7~17rVMMZ-I4ZQFFfpG~-rhIo22jvGeWHrs69W4CmhlS8dG92TtT8lSmx3P15Bc9q6pu1yRXMBFr89ZfCySnmfy… | indic_ocr_bench_test_ben_171 | স্মারক নং:- ৩৭.০২.০০০০.১০৭.৩১.১৫০.২০২৩-৩৩১৩ | Bengali |
| {"src":"https://datasets-server.huggingface.co/assets/sarvamai/indic-ocr-bench/--/84ce7ce447456a92bcbf25f3c0a55d6a5a44a24b/--/default/test/2/image/image.jpg?Expires=1790023260&Signature=LDLQCz-VoC0ZUOUYHKsjPkWgFdDeZWTFH4moUEEybQcJx~2O-Vdh39bB6Ykdl6ROa4gybqttIxQUiqch3tFSrqV2Djt2pEMdqZaLHia25~DWHkZil… | indic_ocr_bench_test_ben_194 | “অ্যাঁ —বলেন কি!” | Bengali |
| {"src":"https://datasets-server.huggingface.co/assets/sarvamai/indic-ocr-bench/--/84ce7ce447456a92bcbf25f3c0a55d6a5a44a24b/--/default/test/3/image/image.jpg?Expires=1790023260&Signature=FWbbHpX6BnQ7kxR9s21jWS3hBIh4bZIDPvKRUZQqCOKwaM8UMGrgBEM~l8rRtzaXUAjbKQmpMPypdyfdgImaSQILscR5llyBbUPWqV7GqwIsNKuuf… | indic_ocr_bench_test_ben_26 | একইভাবে, 6 এবং 4 পরিমেয় সংখ্যা দুটি যোগ করে দেখা গেল যে 6 + 4 = 10 একটি পরিমেয় সংখ্যা। আবার (-9) ও (-5) পরিমেয় সংখ্যা দুটির ক্ষেত্রে (-9)+(-5)=-14 একটি পরিমেয় সংখ্যা পাওয়া গেল। | Bengali |
| {"src":"https://datasets-server.huggingface.co/assets/sarvamai/indic-ocr-bench/--/84ce7ce447456a92bcbf25f3c0a55d6a5a44a24b/--/default/test/4/image/image.jpg?Expires=1790023260&Signature=O2LY9AAcWbVwMwJ0Ap3VVdDm658iXkyraIu-mfNlNWiYdFXYA8sDpuHLnR6v1qR-etDETQS2bfbX77lYYhFW8Aer9HwUe~akLVulBL2pUSMWLBBge… | indic_ocr_bench_test_ben_163 | ভলেন্টিয়ার—দেখুন, দশ বছর ছেড়ে একশো বছর কল্কাতায় থাক্লেও কেউ | Bengali |
Cite this
Apache-2.0imagetextCite
Tags
task_categories:image-to-texttask_categories:text-generationlanguage:aslanguage:bnlanguage:brxlanguage:doilanguage:enlanguage:gulanguage:hilanguage:knlanguage:kslanguage:koklanguage:mailanguage:mllanguage:mnilanguage:mrlanguage:nelanguage:orlanguage:palanguage:salanguage:satlanguage:sdlanguage:talanguage:telanguage:urlicense:apache-2.0size_categories:1K<n<10Kformat:parquetmodality:imagemodality:textlibrary:datasetslibrary:dasklibrary:polarslibrary:mlcroissantregion:usocrindicdocument-understandingmultilingual