Installation of all the packages I'm using into the collab environment worked fine However whenever I try to run the following code cell

from pyannote.audio.pipelines.speaker_verification import PretrainedSpeakerEmbedding

it automatically crashes the session.

The logs indicate this is due to frozen modules which as far as I understand it means some version of some package is not playing nicely with some other package's version.

Now I've been experimenting with stuff all day ; in the middle it seemed like uninstallating torch, torchaudio, & torchmetrics (and letting them get reinstalled as dependencies needed by pyannote/other packages) was working, but it isn't any longer.

If anyone has versions of the following packages & their dependencies (especially torch related) that they know play nicely with each other in collab (or otherwise knows something that'd be useful to solve this), it would be much appreciated

• faster-whisper
• pyannote.audio
• whisper
• ctranslate2==4.4.0
• librosa
• soundfile
• scikit-learn
• speechbrain==0.5.16