Victor Ma
Posts AI Python Go K8s Aws Azure Tags About
Victor Ma

Use Whisper for Dictation

Install python

Refer Install python with Miniconda to install python 3.10.10

Install PyTorch

Get the following conda command from https://pytorch.org/get-started/locally/

1
2
3
4

# Start "Anaconda prompt", or 
%windir%\System32\cmd.exe "/K" C:\Users\vma\miniconda3\Scripts\activate.bat C:\Users\vma\miniconda3

conda install pytorch torchvision torchaudio cpuonly -c pytorch

Install Chocolatey

Refer https://chocolatey.org/Finstall to get a one liner (from Administrator prompt) like this:

1

Set-ExecutionPolicy Bypass -Scope Process -Force; [System.Net.ServicePointManager]::SecurityProtocol = [System.Net.ServicePointManager]::SecurityProtocol -bor 3072; iex ((New-Object System.Net.WebClient).DownloadString('https://community.chocolatey.org/install.ps1'))

Install FFMPEG

To read different formats of audio files. run the following command from Administrator prompt to install.

1

choco install ffmpeg

Install Whisper

Refer https://github.com/openai/whisper to get instructions.

1

pip install -U openai-whisper

Use Whisper

  1. set language can save time to detect languages.
  2. fp16 False is for CPU machine.
  3. Refer to Available models and languages for –model choices.
1
2
3
4

# Start "Anaconda prompt", or 
%windir%\System32\cmd.exe "/K" C:\Users\vma\miniconda3\Scripts\activate.bat C:\Users\vma\miniconda3

(base) C:\Users\vma\Documents\Sound recordings>whisper Recording.m4a --language Chinese --fp16 False --model medium

medium is the sweet spot, much better than default and faster than large.

You can also translate (only to English) with –task translate option.