vocos
torch>=2
torchaudio
soundfile
huggingface_hub
fastprogress
fastcore
numpy

[speaker]
speechbrain<1.0
