Comment by sebastianvoelkl

The only thing Whisper misses is speaker diarization. I'm currently working on a model that uses Whisper + pyannote to transcribe Interviews and also detects who is speaking. It's working but damn it takes so long