search

Vaibhavs10 / insanely-fast-whisper

Apache License 2.0
6.94k stars 505 forks source link

Change pyannote settings using Insanely-Fast-Whisper? #152

Open Maldoror1900 opened 6 months ago

Maldoror1900 commented 6 months ago

I use this line of code to transcribe and diarize at the same time : 

!pipx run insanely-fast-whisper --file-name "/content/drive/MyDrive/aurore.wav" --hf_token <HF TOKEN>

but I get more speakers than there are on the audio. Knowing that on pyannote, the parameter that handles this is num_speakers=2:

diarization = pipeline("audio.wav", num_speakers=2) , or diarization = pipeline("audio.wav", min_speakers=2, max_speakers=5)

Is there a way to implement this parameter in the line, using insanely-fast-whisper?

Vaibhavs10 commented 6 months ago

Good idea. I'll add support for this in the next release. Sorry for the delay in responding to you.

koromkorom commented 5 months ago

Yes! +1 on this! Just pass through. Any News on this. Could offer a PR as well.

koromkorom commented 4 months ago

@Vaibhavs10 Any news on this?

felixcarmona commented 4 months ago

I have opened a PR to allow that: https://github.com/Vaibhavs10/insanely-fast-whisper/pull/180