voidful/tkt-song

TKT Song Public Taiwanese Hokkien song dataset in Hugging Face AudioFolder layout. This dataset contains rows where all three resources are available and line-aligned: song audio converted to 16 kHz mono FLAC Han-character lyrics in lyrics.han Hokkien romanization in lyrics.hokkien Current build row count: 29384. Schema audio: Hugging Face Audio feature generated from file_name source_id: source lyrics id title: source song title duration_seconds: YouTube… See the full description on the dataset page: https://huggingface.co/datasets/voidful/tkt-song.