Whisper

• Requirement
  • python >=3.7, < 3.11
  • ffmpeg

install (GPU ver)

pip install git+https://github.com/openai/whisper.git 
pip install --upgrade --no-deps --force-reinstall git+https://github.com/openai/whisper.git
pip uninstall torch
pip cache purge
pip install torch -f https://download.pytorch.org/whl/torch_stable.html

use whisper (command ver)

need to source activate script

<<WhisperInstallPath>>\Scripts\Activate.ps1
whisper intput.m4a --language zh --model medium > test1.txt

use whisper (py ver)

python convert.py

other reference:

https://www.youtube.com/watch?v=YhLyeOJIoGE