tazz4843 / whisper-rs

Rust bindings to https://github.com/ggerganov/whisper.cpp
The Unlicense
652 stars 107 forks source link

Real-time audio input example #23

Open bruskajp opened 1 year ago

bruskajp commented 1 year ago

The whisper.cpp repo has a real time audio input example. I think we should have a version of that in the rust code here.

I'm on my second iteration of it, following similar techniques to those used in the the cpp example code.

How do you feel about this being a part of the whisper-rs repo?

The other packages that I use are: cpal: microphone recording webrtc-vad: voice recognition detection ringbuf: a ring buffer

tazz4843 commented 1 year ago

This would be certainly a very good idea I think. If and when you do make a PR, request a review from me that way I see it quicker.

ShayBox commented 1 year ago

This is exactly what I need, I want to add live captioning to VRChat.
I spent all day getting vosk to work but results are poor, Whisper looks promising.

I can't wait for this 😄

EDIT: Does this fork/pull request currently work? it gives me various actions like breathing, but no words.