Skip to content

Dataset · jhu-clsp

Parallel corpus for translation into Creole languages

jhu-clsp/kreyol-mt

A text dataset of 1 million to 10 million rows for machine translation of low-resource Creole languages.

What's inside

A repository from researchers at Johns Hopkins University (JHU CLSP) containing parallel text pairs for machine translation involving Creole languages (including acf, aoa, bah, bzj, bzk, cri, and others). The total size of the presented data ranges from 1 million to 10 million records. The authors note that the publication is currently incomplete, as the remaining part release from the LDC consortium is pending.

What you can do with it

  • Training machine translation models for low-resource languages
  • Fine-tuning multilingual LLMs to understand Creole dialects
  • Studying cross-lingual knowledge transfer in NLP
Tasks
Translation
Languages
acf, aoa, bah, bzj, bzk, cri, Seselwa Creole French, dcr and 4 more
Size
1M–10M rows
Downloads this month
1.3K
Likes
14

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("jhu-clsp/kreyol-mt")