fastapi
uvicorn[standard]
torch
torchaudio
transformers
speechbrain
soundfile
numpy
pydantic