Optimize hyperparameters in pyannote 3.1.1

[ChristianNSchmitz]

Tested versions

Reproducible in 3.1.1

System information

Ubuntu 23.04, Intel i7; MacOS 13.6, M3

Issue description

Dear pyannote-developers,

we are currently working with pyannote to transcribe dialogues of two to three people. However, I noticed that small "single-word-sentences" like "yes" or "I see" are neglected by pyannote. Thus, I would like to ask you whether there is the option to optimise the hyper parameters of pyannote in order to get a higher sensitivity. We currently use the command:

pipeline = Pipeline.from_pretrained('pyannote/speaker-diarization-3.1', use_auth_token=authorization_token)
diarization = pipeline(audio_file_path)

Thank you very much!
Best
Christian

Minimal reproduction example (MRE)

See above