vocos
torch>=2
av
soundfile
huggingface_hub
fastprogress
fastcore
numpy

[speaker]
speechbrain>=1.0
