'speechbrain' must be installed to use 'speechbrain/spkrec-ecapa-voxceleb' embeddings.

[ijean]

Tested versions

pyannote.audio version 3.1.1
speechbrain 1.0.0

System information

Windows 11 x64

Issue description

Hello,
I've successfully installed pyannote-audio (accepted the licenses) and was able to do speaker diarization on an audio file with two speakers.

I try to follow one of your tutorials https://github.com/pyannote/pyannote-audio/blob/develop/tutorials/speaker_verification.ipynb

The code crashes at:

model = PretrainedSpeakerEmbedding(
    "speechbrain/spkrec-ecapa-voxceleb",
    device=torch.device("cuda"))

with 'speechbrain' must be installed to use 'speechbrain/spkrec-ecapa-voxceleb' embeddings. Visit https://speechbrain.github.io for installation instructions.

I've installed speechbrain with "pip install speechbrain" but I still get the same error.

Any suggestion ?

Thanks for all your great work,
Jean

Minimal reproduction example (MRE)

https://github.com/pyannote/pyannote-audio/blob/develop/tutorials/speaker_verification.ipynb

[ijean]

I found a temporary workaround:

pip uninstall speechbrain
pip install speechbrain==0.5.16

[davidlee1102]

pip install speechbrain==0.5.16

Amazing, I faced with the problem also, and thank you for your solution

[ahmetkipkip]

I found a temporary workaround:

pip uninstall speechbrain
pip install speechbrain==0.5.16

That works, thanks