langchain
gradio
transformers
accelerate
tensorflow>=2.0
torch
flax
flask