Minimal extension of OpenAI's Whisper adding speaker diarization with special tokens - View it on GitHub