Dataset · ieasybooks-org
Digitized Islamic library in Arabic
ieasybooks-org/waqfeya-library
Over 10 thousand Arabic books in PDF, TXT, and DOCX formats, recognized via Google Document AI.
What's inside
An Arabic-language corpus gathered from the Waqfeya online library, covering over 80 thematic categories. It contains 22,443 PDF files (nearly 9 million pages) representing 10,150 books, along with extracted text in TXT and DOCX formats obtained using the Google Document AI API.
What you can do with it
- Training and testing OCR systems and complex Arabic layout recognition
- Pretraining LLMs and RAG systems on large-scale Arabic literature
- Training text extraction models from multi-page PDF documents
- Tasks
- Image → text
- Languages
- Arabic
- Size
- 10K–100K rows
- Downloads this month
- 210.6K
- Likes
- 23
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("ieasybooks-org/waqfeya-library")