A standardized bidirectional script transliteration and schwa (e) vowel restoration benchmark for Kabyle (Taqbaylit, kab, Latin & Tifinagh scripts), created by the AƔBALU project.
KabTifinagh normalises, repairs, deduplicates, and structures 497,944 parallel sentence entries matching Neo-Tifinagh (kab_Tfng) to canonical Kabyle Latin (kab_Latn), alongside 123,852 English and 205,637 French trilingual sentence alignments.
from datasets import load_dataset
script =… See the full description on the dataset page:
https://huggingface.co/datasets/agbalu/KabTifinagh.
