How to use pyannote-audio for diarization ?

[mkr00]

Hello ,
Could anyone tell me how to use pyannote-audio for diarization, please ?. If I want to use my wave file and I want to get the result of diarization in the rttm file.

BR
mkr00

[Yagna24]

To load the diarization pipeline use this :

pipeline = torch.hub.load('pyannote/pyannote-audio', 'dia_ami')
Your .wav file should be passed like this :

AUDIO = {'uri': 'ES2004a.Mix-Headset', 'audio': '/content/ES2004a.Mix-Headset.wav'}

To run the pipeline :

diarized = pipeline(AUDIO)

Results :

segments = diarized.for_json()["content"] --> to get the timestamps and labels

diarized --> To view the plot

You could then convert the output to rttm format