跳到正文
HuggingFace Daily Papers(社區熱門論文)·· 10 天前AI 評分51

Kezek 將雙人單聲道對話拆分為每位説話者各一條音軌,供全雙工對話數據使用

Kezek: Splitting Mono Two-Person Conversations into Per-Speaker Tracks for Full-Duplex Dialogue Data

AI 導讀

Kezek 將雙人單聲道錄音拆分為兩條對應説話者的音軌,供全雙工語音模型使用;音軌保持原錄音的長度和時間位置,重疊發言及短回覆留在相應説話者音軌,兩條音軌相加後與輸入完全一致。

正文

Kezek splits a mono recording of two people into one track per speaker, same length and timing. Overlaps and short replies ("yeah", "mm-hm") stay on the right person's track, and the two tracks sum exactly to the input.

We built it to get two-channel conversation data for full-duplex voice models, since almost all real conversations online exist only as a mono mix.

On Open Yap test (121 real conversations), compared to pyannote → DialogueSidon (DuplexChat pipeline, run unchanged): 0.83 vs 0.69 of short replies kept, 0.93 vs 0.79 recordings without swapped channels, 231 vs 51 hours of audio per GPU-hour.

41M params, weights Apache-2.0, code MIT.
Demo: https://huggingface.co/spaces/manifestro/kezek
Weights: https://huggingface.co/manifestro/kezek-v1

來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co