This repo is based on the work done here by OpenAI. This repo allows you use use a mic as demo. This repo copies some of the README from the original project.
The latest video tutorial for this repo can be seen here
An older video tutorial for this repo can be seen here
If are in need of paid professional help, that is available through this email
Now a pip package!
- Create a venv of your choice.
- Run
pip install whisper-mic
There are five model sizes, four with English-only versions, offering speed and accuracy tradeoffs. Below are the names of the available models and their approximate memory requirements and relative speed.
| Size | Parameters | English-only model | Multilingual model | Required VRAM | Relative speed |
|---|---|---|---|---|---|
| tiny | 39 M | tiny.en |
tiny |
~1 GB | ~32x |
| base | 74 M | base.en |
base |
~1 GB | ~16x |
| small | 244 M | small.en |
small |
~2 GB | ~6x |
| medium | 769 M | medium.en |
medium |
~5 GB | ~2x |
| large | 1550 M | N/A | large |
~10 GB | 1x |
For English-only applications, the .en models tend to perform better, especially for the tiny.en and base.en models. We observed that the difference becomes less significant for the small.en and medium.en models.
You can use the model with a microphone using the whisper_mic program. Use -h to see flag options.
Some of the more important flags are the --model and --english flags.
Using the command: whisper_mic --loop --dictate will type the words you say on your active cursor.
You can use this code in other projects rather than just use it for a demo. You can do this with the listen method.
from whisper_mic.whisper_mic import WhisperMic
mic = WhisperMic()
result = mic.listen()
print(result)Check out what the possible arguments are by looking at the cli.py file
If you are having issues, try the following:
sudo apt install portaudio19-dev python3-pyaudio
Some ideas that you can add are:
- Supporting different implementations of Whisper
- Adding additional optional functionality.
- Use Pyaudio to get the audio for the listen method to speed things up
The model weights of Whisper are released under the MIT License. See their repo for more information.
This code under this repo is under the MIT license. See LICENSE for further details.
Until recently, access to high performing speech to text models was only available through paid serviecs. With this release, I am excited for the many applications that will come.