---
language:
- ar
- ary
license: cc-by-4.0
task_categories:
- automatic-speech-recognition
pretty_name: Darija ASR Dataset
size_categories:
- n<1K
tags:
- darija
- moroccan-arabic
- asr
- speech
- audio
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

# Darija ASR Dataset

Dataset de reconnaissance automatique de la parole en Darija marocaine.

## Description

- Langue: Darija marocaine (`ary`)
- Tache: automatic speech recognition
- Audio: WAV mono 16 kHz stocke en Parquet
- Colonnes: `audio`, `sentence`

## Structure

| Colonne | Type | Description |
|---|---|---|
| `audio` | Audio | Segment audio WAV mono 16 kHz |
| `sentence` | string | Transcription en darija |

## License

[CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)