Datasets:

henrygagnier
/

kazakh-ocr

language:
  - kk
tags:
  - kazakh
  - ocr
  - benchmark
license: mit
task_categories:
  - image-text-to-text
  - image-to-text
pretty_name: KazakhOCR
size_categories:
  - 1K<n<10K

KazakhOCR: A Synthetic Benchmark for Evaluating Multimodal Models in Low-Resource Kazakh Script OCR

Synthetic OCR benchmark for Arabic, Cyrillic, and Latin script Kazakh OCR. Labels are in metadata.csv.

Curated by: Henry Gagnier, Sophie Gagnier, Ashwin Kirubakaran
Language(s) (NLP): Kazakh (kk)
License: MIT

Citation

@inproceedings{gagnier-etal-2026-kazakhocr,
    title = "{K}azakh{OCR}: A Synthetic Benchmark for Evaluating Multimodal Models in Low-Resource {K}azakh Script {OCR}",
    author = "Gagnier, Henry  and
      Gagnier, Sophie  and
      Kirubakaran, Ashwin",
    booktitle = "Proceedings of the 2nd Workshop on {NLP} for Languages Using {A}rabic Script",
    month = mar,
    year = "2026",
    address = "Rabat, Morocco",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2026.abjadnlp-1.8/",
    pages = "60--65"