Update Dockerfile

Dockerfile

@@ -20,23 +20,28 @@ WORKDIR /build/llama.cpp
 RUN git checkout b3465 
 
 ENV CUDA_DOCKER_ARCH=${CUDA_DOCKER_ARCH}
-ENV LLAMA_CUBLAS=1
+# ENV LLAMA_CUBLAS=1
+ENV GGML_CUDA=1
 
+# Use the default CUDA archs if not specified
 RUN mkdir build && \
-    cd build && \
-    cmake .. -DLLAMA_CUBLAS=ON && \
-    cmake --build . --config Release
+    if [ "${CUDA_DOCKER_ARCH}" != "default" ]; then \
+    export CMAKE_ARGS="-DCMAKE_CUDA_ARCHITECTURES=${CUDA_DOCKER_ARCH}"; \
+    fi && \
+    cmake -B build -DGGML_CUDA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \
+    cmake --build build --config Release --target llama-server -j$(nproc)
 
 FROM ${BASE_CUDA_RUN_CONTAINER} as runtime
 RUN apt-get update --fix-missing && \
-    apt-get install -y --no-install-recommends wget && \
+    apt-get install -y --no-install-recommends wget libgomp1 libcurl4-openssl-dev && \
     rm -rf /var/lib/apt/lists/*
 
 WORKDIR /app
 
-# Copy the executable from the build stage
-COPY --from=build /build/llama.cpp/build/bin/server /app
-COPY --from=build /build/llama.cpp/examples/server/public /app/public
+COPY --from=build /build/llama.cpp/build/ggml/src/libggml.so /app/libggml.so
+COPY --from=build /build/llama.cpp/build/src/libllama.so /app/libllama.so
+COPY --from=build /build/llama.cpp/build/bin/llama-server /app/server
+
 COPY ./run.sh /app/run.sh
 WORKDIR /app
 EXPOSE 7867