khursanirevo/multiturn_ks
khursanirevo/multiturn_ks Dataset Description Multiturn dialogue dataset with speaker-separated stereo audio and multi-language transcripts from 139 YouTube videos. Features Audio: Stereo audio with speaker separation (speaker 0 = left channel, speaker 1 = right channel) Segments: Speaker turn-level annotations with timestamps for English and Malay Multi-language: Transcripts in 9 languages (en, ms, zh-Hans, zh-Hant, ru, id, ar, ja, ko) Video ID:… See the full description on the dataset page: https://huggingface.co/datasets/khursanirevo/multiturn_ks.
Add quality confidence columns (26169 clean, 214 flagged, avg confidence 0.948)
Add quality confidence columns (26169 clean, 214 flagged, avg confidence 0.948) (part 00002-of-00003)
Add quality confidence columns (26169 clean, 214 flagged, avg confidence 0.948) (part 00001-of-00003)
Add quality confidence columns (26169 clean, 214 flagged, avg confidence 0.948) (part 00000-of-00003)
Add README
Upload multiturn dataset (26383 chunks from 139 videos)
initial commit