PARME Parallel Corpora

Year: 2025
Size: Southern Kurdish: 9,797; Laki: 3,418; Hawrami: 7,794; Zazaki: 4,401 aligned records as documented by the project
License: Corpus/code repository: MIT; model weights have separate licensing
Domain: Machine Translation
Languages:
Southern Kurdish Laki Hawrami Zazaki English Persian
Formats:
TSV JSONL

Resource Description

Community-created parallel resources for low-resource Middle Eastern language varieties. Kurdish-related resources identified by the publication include Southern Kurdish, Laki, Hawrami and Zazaki data.

Data Structure

Parallel TSV resources with language-variety and orthographic metadata, train-validation-test splits and training files.

Primary Publication

PARME: Parallel Corpora for Low-Resourced Middle Eastern Languages

Sina Ahmadi (2025)

Presents community-created parallel corpora for low-resource Middle Eastern language varieties and machine translation experiments using NLLB-600M. Kurdish-related resources include Southern Kurdish, …

Related Publications

How to Cite

Ahmadi, S., Sennrich, R., Karami, E., et al. (2025). PARME: Parallel Corpora for Low-Resourced Middle Eastern Languages.

Dataset Information

Total Size Southern Kurdish: 9,797; Laki: 3,418; Hawrami: 7,794; Zazaki: 4,401 aligned records as documented by the project
Languages 6 languages
Formats 2 formats
Related Papers 1 publication

Data Access

Download Dataset

Licensed under Corpus/code repository: MIT; model weights have separate licensing

View Main Paper Back to Datasets