Datasets:

Omar-youssef
/

arabic-document-ocr-markdown

dataset_info:
  features:
    - name: image
      dtype: image
    - name: markdown
      dtype: string
  splits:
    - name: train
      num_bytes: 634447055
      num_examples: 1256
  download_size: 554295266
  dataset_size: 634447055
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Arabic Document OCR Markdown Dataset

Dataset Description

This dataset contains 1,256 pairs of document images and their corresponding Markdown representations, specifically designed for Arabic document OCR tasks. The dataset is intended for training and evaluating models that convert document images into structured Markdown text (image-to-markdown OCR).

Features

The dataset consists of two main features:

image: Document images in various formats
markdown: Corresponding Markdown text extracted from the images

Use Cases

This dataset can be used for:

Training image-to-text OCR models for Arabic documents
Evaluating document understanding models

Loading the Dataset

from datasets import load_dataset

dataset = load_dataset("Omar-youssef/arabic-document-ocr-markdown")

Citation

If you use this dataset in your research, please cite:

@dataset{arabic_document_ocr_markdown,
  author = {Omar Youssef},
  title = {Arabic Document OCR Markdown Dataset},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/Omar-youssef/arabic-document-ocr-markdown}
}