gradio
pinecone-client
transformers
torch
sentence-transformers==2.2.2
pinecone-text
accelerate
optimum
auto-gptq
llama-cpp-python