---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: answer
    dtype: string
  - name: source
    dtype: string
  splits:
  - name: train
    num_bytes: 20267981.0
    num_examples: 200
  download_size: 16302920
  dataset_size: 20267981.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
language:
- ar
task_categories:
- image-to-text
- document-understanding
- visual-question-answering
license: apache-2.0
tags:
- ocr
- arabic
---

**Please see paper & code for more information:**
- [KITAB-Bench: A Comprehensive Multi-Domain Benchmark for Arabic OCR and Document Understanding](https://arxiv.org/abs/2502.14949)
- [GitHub Repository](https://github.com/mbzuai-oryx/KITAB-Bench)