Skip to content

Dataset · M-A-D

Large Arabic Text Corpus

M-A-D/Mixed-Arabic-Datasets-Repo

A large-scale collection of diverse Arabic texts from various sources for a wide range of NLP tasks.

What's inside

The collection combines between 100 million and 1 billion text and tabular records in Arabic. The data was gathered by the community from various online platforms and existing sets to solve the problem of fragmented Arabic corpora on the web.

What you can do with it

  • Pre-training and fine-tuning Arabic language models
  • Machine translation, summarization, classification, and question-answering tasks
Tasks
Text classification, Question answering, Translation, Summarization, Text generation, Fill mask
Languages
Arabic
Size
100M–1B rows
Downloads this month
4.5K
Likes
39

How to load it

from datasets import load_dataset

ds = load_dataset("M-A-D/Mixed-Arabic-Datasets-Repo")