torch
omegaconf
torchaudio
einops
numpy
transformers==4.48.1
sentencepiece
tqdm
tensorboard
descript-audiotools>=0.7.2
descript-audio-codec
mmgp==3.1.4-post15
scipy
gradio