torch
torchaudio
torchvision
huggingface-hub==0.16.4
pytube
youtube-transcript-api
librosa
soundfile
transformers
langchain==0.0.277
gradio
gradio_client
accelerate
tiktoken