ARBML

ARBML is a community of +700 researchers working on Arabic NLP research and development. Although we come from different walks of life, we all strive to achieve one goal: empowering our beloved language with open-source Arabic tools and applications.

Mission

Our mission is to democratize Arabic NLP research and development through open research and collaboration.

Highlighted Projects

...
CIDAR Paper Code Demo

Culturally Relevant Instruction Dataset For Arabic. CIDAR contains 10,000 instructions and their output.

...
Ashaar Paper Code Demo

An extension of qawafi that contains poetry analysis and generation. Using this platform we provide four datasets and five models all available for free.

...
Taqyim Paper Code

A library for evaluting Arabic NLP tasks on chatgpt models. We provide examples for doing evaluation on classification, part of speech tagging, translation, summarization and diacritization>

...
Dar Code

dar or دار which means house in Araibc is a simple semi-supervised approach for creating huggingface data script loaders and upload to the hub.

...
Masader Paper Code Web

The largest public catalogue for Arabic NLP and speech datasets. There are +500 datasets annotated with more than 25 attributes.

...
Klaam Code

Arabic speech recognition, classification and text-to-speech library. The models are built on top of Wave2Vec and FastSpeech.

...
Calliar Paper Code Web

A dataset for online Arabic calligraphy. A collection of 2500 strokes manully annotated calligraphic styles.

...
tkseem Paper Code

Arabic Tokenization Library. It provides many algorithms to tokenize and segment Arabic text.

...
ARBML Paper Code Web

Implementation of many Arabic NLP and CV projects. Providing real time experience using many interfaces like web, command line and notebooks.

News

  • We are Sponsored by ML Collective and Maqsam, May 2023.
  • Adawat for Arabic tools is launched here .
  • Masader Plus website is up and running here .
  • We launched Masader hackathon on the 10 June to 20 June on our discord in collaboration with HuggingFace.
  • Masader highlighted in The Washington Post .
  • MIT Techonology Review wrote an article about Calliar .

Our Contributors

Collaborators

Sponsors