whisper-playground

A playground to use whisper python package for transcription. A dev container is used to set up all that is needed included whisper, pyannote, ffmpeg and pydub.

Requirements:

• Docker installed locally
• If you want to use diarization, pyannote package reuires Hugging Face access token

1. Open folder in VS Code
2. copy .devcontainer/env.sample into .devcontainer/.env file and update environment variables as needed.
3. Run 'Reopen in Container' from the command pallete

Transcription & Diarization

To transcribe a video or audio file, run:

python transcription.py --file "path/to/audio/or/video/file"

If you also want to add diarization and align the transcription with each speaker segment run"

python transcription.py --file "path/to/audio/or/video/file" --diarization True