m-bain / whisperX

WhisperX: Automatic Speech Recognition with Word-level Timestamps (& Diarization)
BSD 2-Clause "Simplified" License
12.58k stars 1.33k forks source link

Use whisperx diarization offline #873

Open itaipee opened 2 months ago

itaipee commented 2 months ago

WhisperX diarization is done with Pyannote .
I'm using whisper-X for transcription in closed environment, no internet access. It works well with whisper transcription , since we can download the model.

However, Pyannote requires HF token . Can Pyannote be used offline ? downloaded somehow and than used without syncing with HF ?

jrknox1977 commented 1 month ago

I am trying to achieve this as well.

https://github.com/pyannote/pyannote-audio/blob/develop/tutorials/applying_a_model.ipynb

https://github.com/pyannote/pyannote-audio/blob/develop/tutorials/applying_a_pipeline.ipynb

jrknox1977 commented 1 month ago

This one appears to be the most promising: https://github.com/pyannote/pyannote-audio/blob/develop/tutorials/community/offline_usage_speaker_diarization.ipynb