transformers
torch
gradio
huggingface_hub
accelerate
scikit-learn