Darija YouTube Dataset
Domain:
natural language processingRecord type:
dataset
Creator:
DrI
Host:
A dataset of Moroccan Arabic (Darija) speech scraped from YouTube, with transcriptions in Arabic script, Latin script (Arabizi), and English translations.
This dataset contains sentence-level audio segments of Darija speech, paired with:
Arabic transcription (modern Moroccan Arabic script)
Latin transliteration (Arabizi format: 3=ع, 7=ح, 9=ق, etc.)
English translation
Column
Type
Description