bekan/english_karakalpak_parallel_corpus_v3-4

English-Karakalpak Parallel Corpus

This dataset contains parallel sentences in English and Karakalpak language. It is created to support AI development for the Karakalpak language.

Dataset Description

English-Karakalpak Parallel Corpus v3-4 is a high-quality dataset containing 2,722 carefully aligned sentence pairs in English (en) and Karakalpak (kaa).

This dataset is designed to advance the representation and capability of the Karakalpak language in large-scale AI models (LLMs) and Neural Machine Translation (NMT) systems, enabling them to better understand and generate Karakalpak text. This resource specifically utilizes the official Latin script for Karakalpak.

• —Language(s): English (en), Karakalpak (kaa)
• —Format: CSV (Comma-Separated Values)
• —License: MIT
• —Script: Latin (Karakalpak)

Dataset Structure

Data Fields

The dataset is configured to follow the standard translation schema:

• —en (string): The source sentence in English.
• —kaa (string): The corresponding translation in Karakalpak (Latin script).

Example

The data is structured for direct use in translation tasks:

{
  "en": "Heading is a common technique and is used by players in practically every match.",
  "kaa": "Bas penen urıw keń tarqalǵan texnika bolıp, oyınshılar tárepinen is júzinde hár bir matchta qollanıladı."
}