Datasets:

prothmane
/

amawal-dataset

language:
  - ber
  - fr
  - ar
multilinguality:
  - multilingual
task_categories:
  - translation
  - text-generation
pretty_name: amawal
tags:
  - amazigh
  - tamazight
  - berber
  - tifinagh
  - low-resource-language
  - multilingual
license: cc-by-4.0
size_categories:
  - 10K<n<100K

ⴰⵎⴰⵡⴰⵍ Amawal

A trilingual Amazigh-French-Arabic lexicon to support AI research for Berber languages.

Description

This dataset brings together Amazigh vocabulary in Tifinagh script (ⵜⵉⴼⵉⵏⴰⵖ) with corresponding French and/or Arabic translations. The goal is to make it easier to develop language technologies for Amazigh, a language spoken by over 30 million people in North Africa but lacking digital resources.

Languages included

Amazigh (Tamazight) [ber]: In Tifinagh script (ⵜⵉⴼⵉⵏⴰⵖ)
French [fr]: French translations
Arabic [ar]: Modern Standard Arabic translations
Phonetic: Latin alphabet transcription

Data format

Each entry follows this structure:

Example with French translation:

{
  "amazigh": "ⴰⴱⵓⵔⴽⴰⵏ",
  "french": "volcan",
  "pronunciation": "aburkan"
}

Example with Arabic translation:

{
  "amazigh": "ⴽⵍⴽⵍ",
  "french": "",
  "arabic": "بدن",
  "pronunciation": "klkl"
}

Usage

Loading the dataset

from datasets import load_dataset

dataset = load_dataset("prothmane/amawal-dataset")
print(dataset['train'][0])

Use cases

This dataset can be used for:

Machine translation systems
Language models for Amazigh
Speech recognition tools
Multilingual embeddings

License

Released under Creative Commons Attribution 4.0 International (CC-BY-4.0).

You can freely share, adapt, and use commercially, as long as you credit the author.

Citation

If you use this dataset:

@dataset{tamazight_trilingual_lexicon_2025,
  author       = {Othmane Daif},
  title        = {amawal},
  year         = {2025},
  publisher    = {Hugging Face},
  url          = {https://huggingface.co/datasets/prothmane/amawal-dataset}
}

ⵜⴰⵏⵎⵎⵉⵔⵜ (Tanmmirt) - Thank you! 🙏