gradio
pyannote.audio==3.1.1
pympi-ling
transformers
torch