yinruiqing / pyannote-whisper

457 stars 72 forks source link

pyannote can't tag every sentence with a speaker label #20

Open JiangN6 opened 6 months ago

JiangN6 commented 6 months ago

As you can see in the figure, pyannote does not label each sentence as speaker. There are even more severe cases in which a 5-minute audio clip has no labels for the previous minute of conversation Dingtalk_20231221132801

yinruiqing commented 4 months ago

This repo relies on pyannote-audio for speaker diarization. Please open an issue and provide an example in pyannote-audio.