LaboroTVSpeech

홈페이지 · 논문 2편

LaboroTVSpeech is a large-scale Japanese speech corpus built from broadcast TV recordings and their subtitles. It contains over 2,000 hours of speech.