Comments (6)
Perhaps some analysis could be done on the wav audio to see if its silence or not
from whisper_mic.
@BakingBrains That's a known issue with a lot of AI models, called hallucination. Not much can be done about it unfortunately at this stage afaik
from whisper_mic.
Yep @SunnyOd , I knew, I thought there might be some method so that we can suppress the junk output. But thanks though👍
from whisper_mic.
Hi. Has anyone managed to suppress the junk output? I keep getting "Thank you" and a couple of other phrases pop up during inactivity/silence. I've played with energy levels and that has helped some but I think there might be more mileage with the fix below - possibly
I've found some threads where people have managed to fix it, this one looks to give the solution as including --suppress_tokens
in the command when running whisper
. Not sure how to add this flag to whisper_mic since the move to pip installation of whisper_mic. Is it worth looking into?
Thanks!
S
from whisper_mic.
from whisper_mic.
So this is due to hallucinations. There doesn't seem to be a way to fully fix it due to the way that the model was trained. It hallucinates when there is little to no audio so it makes something up.
I added a new flag that helps with this.
I'm going to leave this issue open. If anyone finds a real solution, please ping me here. However, I think with the current flags that exist with the tool that you can find a configuration that limits hallucinations.
from whisper_mic.
Related Issues (20)
- How to terminate program HOT 1
- [Feature request][Bug] Improvement needed on the listen() method. Issues related to setup_mic() and listen_loop() method. HOT 3
- pip install whisper_mic not working HOT 7
- Feature request: Provide code sample for Web UI Mic recording HOT 1
- no transcript output HOT 2
- cannot import the function to another project HOT 2
- Takes considerable time to actually setup the mic and start transcribing HOT 1
- [Fix] Keyboard interrupt for listen_loop HOT 1
- mic.py does not exist in file or directory HOT 1
- Error while using the save_file to save the transcribed data HOT 1
- add large-v3 HOT 1
- Issues with Python Setup HOT 1
- ALSA lib error, invalid card HOT 4
- Proccess hanging in infinite loop when input audio is not loud enough HOT 2
- Feature Request: Use isolated-env to make the app bind to the GPU automatically on windows
- ModuleNotFoundError: No module named 'distutils' HOT 1
- Many incomplete segments, what is it even returning? predicted_text referenced before assignment. HOT 5
- Can't use mic in linux - ALSA errors HOT 1
- Sonoma 14.4.1 - Python 3.12 - Running whisper_mic returns errors HOT 1
- Feature requests HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from whisper_mic.