Skip to content

Dataset · tahrirchi

Uzbek books v2 corpus (OCR)

tahrirchi/uz-books-v2

Bilingual corpus of about 40,000 books in Uzbek across Latin and Cyrillic versions.

What's inside

The collection includes 38,339 books in Latin transliteration and the same number in Cyrillic, extracted using an improved Google Cloud Vision OCR engine. The text is gathered as text data and divided into two parts according to the alphabet.

What you can do with it

  • Pre-training language models for the Uzbek language
  • Fill-mask and text generation tasks
  • Improving optical character recognition (OCR) systems
Tasks
Text generation, Fill mask
Languages
Uzbek
Size
10K–100K rows
Downloads this month
944
Likes
7

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("tahrirchi/uz-books-v2")