Doesn't accept num_speakers as argument

[jhdeov]

I'm running whisperx on an interview audio file. But I can't pass num_speakers as an argument

diarize_model = whisperx.DiarizationPipeline( use_auth_token=YOUR_HF_TOKEN, device=device)
diarize_segments = diarize_model( audio_path, num_speakers = 2
)

I get the error:

TypeError: DiarizationPipeline.call() got an unexpected keyword argument 'num_speakers'

[SeeknnDestroy]

Hey @jhdeov, diarize model does not accept num_speakers param but min_speakers and max_speakers instead.

diarize_model(input_audio_path, min_speakers=min_speakers, max_speakers=max_speakers)

I believe you can try to set them both to 2 and experiment what happens. Please let me know if it works!

[jhdeov]

It works.

[NebularNerd]

Hey @jhdeov, diarize model does not accept num_speakers param but min_speakers and max_speakers instead.

diarize_model(input_audio_path, min_speakers=min_speakers, max_speakers=max_speakers)

I believe you can try to set them both to 2 and experiment what happens. Please let me know if it works!

That did not work for me, I got the following error:
ERROR: DiarizationPipeline.__init__() got an unexpected keyword argument 'min_speakers'

Ah, I see I'm using the example python from the homepage which calls diarize differently:
diarize_model = whisperx.DiarizationPipeline(use_auth_token=YOUR_HF_TOKEN, device=device, min_speakers=args.minspeakers, max_speakers=args.maxspeakers)

Let's see how we fix that....

Well I fixed it with a pretty horrific bodge , no way I could submit it as a PR in its current state, need to improve it.