https://github.com/pyannote/pyannote-audio/archive/develop.zip
speechbrain == 0.5.10
gradio
Jinja2
transformers