faster-whisper==1.1.0
websockets
onnxruntime==1.17.0
numba
kaldialign
soundfile
scipy
av
jiwer
evaluate
numpy<2
openai-whisper==20240930
tokenizers==0.20.3