search

MahmoudAshraf97 / whisper-diarization

Automatic Speech Recognition with Speaker Diarization based on OpenAI Whisper
BSD 2-Clause "Simplified" License
3.53k stars 301 forks source link

Progress on diarisation #58

Closed mirix closed 1 year ago

mirix commented 1 year ago

I have made some (modest) progress on this, if anyone wishes to have a look:

https://github.com/mirix/approaches-to-diarisation/tree/main

TheGermanEngie commented 1 year ago

Hi, I'm interested in using this but can't figure out the exact procedure. You say "with the latest pip versions of all the requirements", and then the default device and pytorch version is cpu only. Should I edit the requirements.txt to fit accordingly, or does it not matter?

I'm not a developer either, just a hobbyist. I'd assume fast_whisper, pyannote and SpeechBrain can be installed with pip also.

mirix commented 1 year ago

Hi, there is no requirements.txt. At this stage it is not a library you can install and use. Just code snippets.

The new procedure works very well in the testing samples:

https://github.com/mirix/approaches-to-diarisation

TheGermanEngie commented 1 year ago

I see you have updated the repo. Are these meant to be used on their own or with the original whisper-diarization repo? Stupid question, I'm very new to github.

mirix commented 1 year ago

Hi, when you install the requirements, namely stable-ts, it will automatically pull Whisper as a dependency.