transformers
torch
accelerate
auto-gptq
sentence-transformers
faiss-cpu
gradio
optimum