Sound

GitHub Repo stars FastWhisper This is an optimized implementation of OpenAI's Whisper

Uses a greedy decode for multilingual transcription. It supports all sizes of the Whisper model (from tiny to large).