torch
transformers
gradio==5.49.1
datasets
librosa
ffmpeg-python
python-dotenv
spaces