File size: 1,932 Bytes
801347d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ac04ad1
 
 
801347d
ac04ad1
801347d
 
ac04ad1
801347d
ac04ad1
801347d
bf82f50
801347d
ac04ad1
801347d
 
ac04ad1
801347d
ac04ad1
801347d
ac04ad1
9de3933
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
---
language:
- kmr
pretty_name: Northern Kurdish
tags:
- Kurdish
- Badini
license: cc0-1.0
task_categories:
- automatic-speech-recognition
---

# Northern Kurdish (Arabic Script) ASR Dataset

## Dataset Description

Northern Kurdish is the most widely spoken variant of the Kurdish language and is used across all parts of Kurdistan. Although it is mainly written today in the **Latin script**, it was historically written in the **Arabic script**. The Arabic script is still used for this dialect in **Southern Kurdistan**, particularly in the **Duhok province of the Kurdistan Regional Government (KRG)**.  
Similarly, the primary writing system for all Kurdish dialects in **Eastern Kurdistan (Iran)** is the Arabic script.

To provide **automatic speech recognition (ASR) resources** for Northern Kurdish in the Arabic script, we transliterated the **Northern Kurdish Common Voice 23** transcriptions from Latin to Arabic script and manually revised the transliterations to ensure consistency with **standard Kurdish orthography in the Arabic alphabet**.

This dataset can be used for the development of **Northern Kurdish ASR systems in the Arabic script**, addressing a critical gap in Kurdish speech technology resources.

## Audio Source

The audio files are **not redistributed** in this repository.  
They are available in **Northern Kurdish Common Voice 23** and can be retrieved from:

https://datacollective.mozillafoundation.org/datasets?q=common+voice

## Languages

- **Code:** `kmr`  
- **Name:** Northern Kurdish

## Tasks

- Automatic Speech Recognition (ASR)

## License

- **cc0-1.0** for the provided transliterated transcriptions  
- Audio files follow the **Common Voice license** (see Mozilla Common Voice)

## Reference

If you use this dataset, please cite:

## Reference

> **Mohammad Mohammadamini (2025).**  
> *Exploring the reusability of Northern Kurdish resources for Badini speech recognition*.