- https://huggingface.co/pyannote/speaker-diarization-3.1 - https://github.com/narcotic-sh/senko
I personally love senko since it can run in seconds, whereas py-annote took hours, but there is a 10% WER (word error rate) that is tough to get around.