fix: mv vllm to cuda-enable requirement file and update Dockerfile

fix: mv vllm to cuda-enable requirement file and update Dockerfile #17