Update Dockerfile

Dockerfile

@@ -1,15 +1,15 @@
 ARG UBUNTU_VERSION=22.04
-ARG CUDA_VERSION=12.2.0
+ARG CUDA_VERSION=12.3.1
 ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION}
 ARG BASE_CUDA_RUN_CONTAINER=nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION}
 
-# Unless otherwise specified, we make a fat build.
-ARG CUDA_DOCKER_ARCH=all
-
 FROM ${BASE_CUDA_DEV_CONTAINER} as build
 
-RUN apt-get update && \
-    apt-get install -y build-essential git cmake wget
+ARG CUDA_DOCKER_ARCH=all
+
+RUN apt-get update --fix-missing && \
+    apt-get install -y --no-install-recommends git build-essential gcc cmake && \
+    rm -rf /var/lib/apt/lists/*
 
 WORKDIR /build
 
@@ -17,30 +17,30 @@ RUN git clone https://github.com/ggerganov/llama.cpp.git
 
 WORKDIR /build/llama.cpp
 
-# Set nvcc architecture
 ENV CUDA_DOCKER_ARCH=${CUDA_DOCKER_ARCH}
-# Enable cuBLAS
 ENV LLAMA_CUBLAS=1
 
-RUN make LLAMA_CUBLAS=1
-
-WORKDIR /data
-RUN wget https://huggingface.co/IlyaGusev/saiga2_70b_gguf/resolve/main/ggml-model-q2_K.gguf -nv -O model.gguf
+RUN mkdir build && \
+    cd build && \
+    cmake .. -DLLAMA_CUBLAS=ON && \
+    cmake --build . --config Release
 
 FROM ${BASE_CUDA_RUN_CONTAINER} as runtime
-
+RUN apt-get update --fix-missing && \
+    apt-get install -y --no-install-recommends curl && \
+    rm -rf /var/lib/apt/lists/*
 
 WORKDIR /app
 
 # Copy the executable from the build stage
-COPY --from=build /build/llama.cpp/server /app
-COPY --from=build /data/model.gguf /data/model.gguf
+COPY --from=build /build/llama.cpp/build/bin/server /app
+COPY --from=build /build/llama.cpp/examples/server/public /app/public
 COPY ./run.sh /app/run.sh
-
-EXPOSE 7860
+WORKDIR /app
+EXPOSE 7867
 
 # Make the script executable
 RUN chmod +x run.sh
-ENV LC_ALL=C.utf8
+
 # CMD to run your script
-CMD /app/run.sh
+CMD ./run.sh