ghananlpcommunity/asante-twi-yarngpt-aligned
Domain:
natural language processingRecord type:
dataset
Creator:
gha
Host:
This is an intermediate dataset produced during the preparation of a YarnGPT-style TTS model for Asante Twi. It contains the original audio from ghananlpcommunity/asante-twi-bible-speech-text enriched with word-level forced alignment timestamps — i.e. the start time, end time, and alignment confidence score for every word in every sentence.