clalanliu/youtubeshadowingpractice

Shadowing Practice Using Python

Make your own VoiceTube. Demo Video
This is an applicaion using whisper.

GUI support

Installation

pip install -r requirements.txt

For cpu user:

1. Install torch first: pip3 install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cpu

Usage

GUI

python gui.py

For Youtube link:

python main.py [--url YOUTUBE_URL] [--pause] [--force] [--model=base] [--words=5] [--speed=1.0]

For mp4 or mp3 files:

python mp4_to_mp3.py mp4filename.mp4  
python video2script.py mp4filename.mp4
python main.py [--pause] [--force] [--model=base] [--words=5] [--speed=1.0]

--url: Specify Youtube link. It is required for the first time use. If not set, play cached video.
--pause: Manually set pause duration. If not set, It will equal the original sentence duration.
--force: Re-processing video
--model: Choose Whisper model size: tiny | base | small | medium | large. Default base
--words: At least # words in a segment. Default 5.
--speed: Adjust speed to #-X. Default 1-X. If negative, represents for #-words per min.