Local Whisper Model

[eFroD]

In the current implementation, an OpenAI key is still needed for video transcription to power the whisper model, even if the rest runs locally.
In order to allow full self-hosting, we should extend the deployment to the local whisper model.

Objectives

• Extend the Codebase and configs to handle local whisper models
• Add it to the docker compose to directly ship the model