README.md

metadata

configs:
  - config_name: Awal
    data_files: parallel/awal/raw-dumps/Contribution-24.03.14.json
    default: true
  - config_name: IRCAM-clean-tifinagh
    data_files: mono/IRCAM-clean-tifinagh.txt
  - config_name: tc_wajdm_v1
    data_files: mono/tc_wajdm_v1.txt
  - config_name: CV-localization
    data_files: parallel/CV-localization/pontoon-CV-zgh-en.tsv
license: cc-by-2.0
language:
  - zgh
  - fr
  - ca
  - en
  - es
  - ary
  - ar
task_categories:
  - translation
  - text-generation
pretty_name: Tamazight language data
size_categories:
  - 100K<n<1M

This repository contains various Tamazight language datasets created by Col·lectivaT in collaboration with CIEMEN and with funding from Municipality of Barcelona and Government of Catalonia.

Under mono you can find monolingual sentences.

• tc_wajdm_v1.txt - Texts from language learning material “tc wawjdm”
• IRCAM-clean-tifinagh.txt - Tifinagh scripted sentences extracted from IRCAM's text corpus

Under parallel you can find sentences with translations.

• AWAL contains data extracted from contributions made to AWAL
• tatoeba-translit contains parallel sentences from Tatoeba.org transliterated into Tifinagh using a Python script.
• proverbs contains Tamazight proverbs with translations in Catalan.
• CV-localization contains parallel English and Tamazight segments from localization of Common Voice, extracted from Pontoon