Update llama-server-cuda.Dockerfile

Author: Williamyzd

.devops/llama-server-cuda.Dockerfile

@@ -1,11 +1,11 @@
 ARG UBUNTU_VERSION=20.04
 
 # This needs to generally match the container host's environment.
-ARG CUDA_VERSION=11.3.1
+ARG CUDA_VERSION=11.4.0
 # Target the CUDA build image
-ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION}
+ARG BASE_CUDA_DEV_CONTAINER=registry.cn-hangzhou.aliyuncs.com/reg_pub/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION}
 # Target the CUDA runtime image
-ARG BASE_CUDA_RUN_CONTAINER=nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION}
+ARG BASE_CUDA_RUN_CONTAINER=registry.cn-hangzhou.aliyuncs.com/reg_pub/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION}
 
 FROM ${BASE_CUDA_DEV_CONTAINER} AS build
 ARG DEBIAN_FRONTEND=noninteractive