streamlit/speach-to-text silently failing

[dimadatasiv]

I am trying diarization of wav file (on my Macos machine)
My steps:
docker build . -t streamlit_app:latest
docker run --rm -it -p 8501:8501 --user=42420:42420 streamlit_app:latest
set token in UI
Provide wav file
select "differentiate speakers" option
Result:
silently fail after downloading models

SampleOrderTakingCustomerSupportPhilippines.wav.zip

[dimadatasiv]

Note i use streamlit/speech-to-text code as is

[Mathieu Busquet]

Hi @dimadatasiv

Thank you for trying our application.
Here is the code which loads the diarization model:

# Load Diarization model (Differentiate speakers)
    with st.spinner("Loading Diarization Model"):
        try:
            dia_pipeline = pickle.load(open("models/dia_pipeline.sav", 'rb'))
        except FileNotFoundError:
            dia_pipeline = Pipeline.from_pretrained("pyannote/speaker-diarization@2.1",
                                                    use_auth_token=st.session_state["my_HF_token"])
            # If the token hasn't been modified, dia_pipeline will automatically be set to None. The functionality will then be disabled.

If you can't load pyannote.audio's diarization model, there are 2 possible reasons:

• Either you have a "models/dia_pipeline.sav" named file, which is therefore imported by the code above, and which is the result of an error, as pyannote's diarization can normally no longer be imported from a file, since version 2>= of the pyannote library.

• Your access token is not functional. Have you tried loading the diarization model in another environment, such as a notebook or local script?

[Mathieu Busquet]

This issue has been closed due to inactivity. If you believe it is still relevant, please leave a comment below.