ctranslate2>=4.0,<5
huggingface_hub>=0.13
tokenizers>=0.13,<1
onnxruntime>=1.14,<2 
pyannote-audio
torch
torchaudio
tqdm