Update Dockerfile

Dockerfile

@@ -1,11 +1,3 @@
-ARG UBUNTU_VERSION=22.04
-# This needs to generally match the container host's environment.
-ARG CUDA_VERSION=11.7.1
-# Target the CUDA build image
-ARG BASE_CUDA_DEV_CONTAINER=nvidia/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}
-
 FROM ${BASE_CUDA_DEV_CONTAINER} as build
 
 # Unless otherwise specified, we make a fat build.
@@ -14,6 +6,9 @@ ARG CUDA_DOCKER_ARCH=all
 RUN apt-get update && \
     apt-get install -y build-essential git
 
+# Install Python3 and pip
+RUN apt-get install -y python3 python3-pip
+
 WORKDIR /app
 
 COPY . .
@@ -34,8 +29,11 @@ RUN apt-get update && \
     pkg-config \
     curl
 
-RUN pip install -U pip setuptools wheel && \
-  pip install --verbose llama-cpp-python[server]
+# Install Python3 and pip for the runtime container as well
+RUN apt-get install -y python3 python3-pip
+
+RUN pip3 install -U pip setuptools wheel && \
+  pip3 install --verbose llama-cpp-python[server]
 
 # Download model
 RUN mkdir model && \
@@ -56,4 +54,4 @@ ENV PORT=7860
 EXPOSE ${PORT}
 
 # Run the server start script
-CMD ["/bin/sh", "./start_server.sh"]
+CMD ["/bin/sh", "./start_server.sh"]