Transcribe (locally on a GPU) long WhatsApp recordings, or any other audio file, using OpenAI's Whisper model.
This code was written in a virtual environment created by PyCharm.
Run the package-installer-py. It will install what's necessary for the UI and local transcription (pytorch etc.)
Run GUI.py, it should pop up a window that's currently a work in progress.
Click the folder icon to load an audio file. Wait for the GPU to cook a little bit, and the transcription will pop up in the text field.
- Semi-Realtime transcription using computer audio
- Transcription of youtube videos by pasting the URL