Add subtitles to your video using machine learning!
-
Download repository:
gh repo clone mathemusician/AudioTranscription
-
Install python requirements (Tested mainly on Python 3.9.0):
pip install -r requirements.txt
-
Install ffmpeg
Using Homebrew on Mac:
brew install ffmpeg
Windows: Install from https://ffmpeg.org/download.html
The app can be split into three parts: 1) model 2) user interface 3) video-maker. For the model, I used a pytorch-lightning implementation of wav2vec2 to convert audio into text. For the user interface, I used streamlit. And for rendering the final video, I use ffmpeg.