datasets[audio]
transformers==4.40.1
torchaudio
accelerate
evaluate
jiwer
tensorboard
gradio
spaces