added docker-multi-stage builds

.devops/full.Dockerfile → .devops/cpu.Dockerfile

@@ -14,7 +14,7 @@ RUN cmake -S . -B build -DGGML_BACKEND_DL=ON -DGGML_NATIVE=OFF -DGGML_CPU_ALL_VA
     mkdir -p /app/lib && \
     find build -name "*.so" -exec cp {} /app/lib/ \;
 
-FROM ubuntu:$UBUNTU_VERSION as runtime
+FROM ubuntu:$UBUNTU_VERSION as full
 
 WORKDIR /app
 
@@ -36,3 +36,37 @@ COPY --from=build /app/gguf-py /app/gguf-py
 ENV LC_ALL=C.utf8
 
 ENTRYPOINT ["/app/tools.sh"]
+
+
+FROM ubuntu:$UBUNTU_VERSION AS light
+
+WORKDIR /app
+
+RUN apt-get update && \
+    apt-get install -y libgomp1
+
+COPY --from=build /app/build/bin/llama-cli /app/
+COPY --from=build /app/lib/ /app/
+
+ENV LC_ALL=C.utf8
+
+ENTRYPOINT [ "/app/llama-cli" ]
+
+
+FROM ubuntu:$UBUNTU_VERSION AS server
+
+WORKDIR /app
+
+RUN apt-get update && \
+    apt-get install -y libcurl4-openssl-dev libgomp1 curl
+
+COPY --from=build /app/build/bin/llama-server /app/
+COPY --from=build /app/lib/ /app/
+
+ENV LC_ALL=C.utf8
+# Must be set to 0.0.0.0 so it can listen to requests from host machine
+ENV LLAMA_ARG_HOST=0.0.0.0
+
+HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
+
+ENTRYPOINT [ "/app/llama-server" ]

.devops/cuda.Dockerfile

@@ -0,0 +1,91 @@
+ARG UBUNTU_VERSION=22.04
+# This needs to generally match the container host's environment.
+ARG CUDA_VERSION=12.6.0
+# Target the CUDA build image
+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}
+
+FROM ${BASE_CUDA_DEV_CONTAINER} AS build
+
+# CUDA architecture to build for (defaults to all supported archs)
+ARG CUDA_DOCKER_ARCH=default
+
+RUN apt-get update && \
+    apt-get install -y build-essential cmake python3 python3-pip git libcurl4-openssl-dev libgomp1
+
+WORKDIR /app
+
+COPY . .
+
+RUN if [ "${CUDA_DOCKER_ARCH}" != "default" ]; then \
+    export CMAKE_ARGS="-DCMAKE_CUDA_ARCHITECTURES=${CUDA_DOCKER_ARCH}"; \
+    fi && \
+    cmake -B build -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \
+    cmake --build build --config Release -j$(nproc) && \
+    cp build/bin/* .
+
+RUN mkdir -p /app/lib && \
+    find build -name "*.so" -exec cp {} /app/lib \;
+
+
+FROM ${BASE_CUDA_RUN_CONTAINER} AS full
+
+COPY --from=build /app /app
+
+WORKDIR /app
+
+RUN apt-get update \
+    && apt-get install -y \
+    python3 \
+    python3-pip \
+    git \
+    libgomp1 \
+    && pip install --upgrade pip setuptools wheel \
+    && pip install -r requirements.txt \
+    && apt autoremove -y \
+    && apt clean -y \
+    && rm -rf /tmp/* /var/tmp/* \
+    && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
+    && find /var/cache -type f -delete
+
+ENTRYPOINT ["/app/.devops/tools.sh"]
+
+FROM ${BASE_CUDA_RUN_CONTAINER} AS light
+
+RUN apt-get update \
+    && apt-get install -y libgomp1 \
+    && apt autoremove -y \
+    && apt clean -y \
+    && rm -rf /tmp/* /var/tmp/* \
+    && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
+    && find /var/cache -type f -delete
+
+COPY --from=build /app/lib/ /app/
+COPY --from=build /app/build/bin/llama-cli /app/
+
+WORKDIR /app
+
+ENTRYPOINT [ "/app/llama-cli" ]
+
+FROM ${BASE_CUDA_RUN_CONTAINER} AS server
+
+RUN apt-get update \
+    && apt-get install -y libgomp1 curl \
+    && apt autoremove -y \
+    && apt clean -y \
+    && rm -rf /tmp/* /var/tmp/* \
+    && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
+    && find /var/cache -type f -delete
+
+COPY --from=build /app/lib/ /app/
+COPY --from=build /app/build/bin/llama-server /app/
+
+WORKDIR /app
+
+# Must be set to 0.0.0.0 so it can listen to requests from host machine
+ENV LLAMA_ARG_HOST=0.0.0.0
+
+HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
+
+ENTRYPOINT [ "/app/llama-server" ]

.devops/intel.Dockerfile

@@ -0,0 +1,87 @@
+ARG ONEAPI_VERSION=2025.0.0-0-devel-ubuntu22.04
+
+## Build Image
+
+FROM intel/oneapi-basekit:$ONEAPI_VERSION AS build
+
+ARG GGML_SYCL_F16=OFF
+RUN apt-get update && \
+    apt-get install -y git libcurl4-openssl-dev
+
+WORKDIR /app
+
+COPY . .
+
+RUN if [ "${GGML_SYCL_F16}" = "ON" ]; then \
+        echo "GGML_SYCL_F16 is set" && \
+        export OPT_SYCL_F16="-DGGML_SYCL_F16=ON"; \
+    fi && \
+    echo "Building with dynamic libs" && \
+    cmake -B build -DGGML_NATIVE=OFF -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DLLAMA_CURL=ON ${OPT_SYCL_F16} && \
+    cmake --build build --config Release -j$(nproc) && \
+    cp build/bin/* .
+
+RUN mkdir -p /app/lib && \
+    find build -name "*.so" -exec cp {} /app/lib \;
+
+FROM intel/oneapi-basekit:$ONEAPI_VERSION as full
+
+COPY --from=build /app /app
+
+WORKDIR /app
+
+RUN apt-get update \
+    && apt-get install -y \
+    python3 \
+    python3-pip \
+    git \
+    libgomp1 \
+    && pip install --upgrade pip setuptools wheel \
+    && pip install -r requirements.txt \
+    && apt autoremove -y \
+    && apt clean -y \
+    && rm -rf /tmp/* /var/tmp/* \
+    && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
+    && find /var/cache -type f -delete
+
+ENTRYPOINT ["/app/.devops/tools.sh"]
+
+FROM intel/oneapi-basekit:$ONEAPI_VERSION AS light
+
+RUN apt-get update \
+    && apt-get install -y libgomp1 \
+    && apt autoremove -y \
+    && apt clean -y \
+    && rm -rf /tmp/* /var/tmp/* \
+    && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
+    && find /var/cache -type f -delete
+
+COPY --from=build /app/lib/ /app/
+COPY --from=build /app/build/bin/llama-cli /app/
+
+WORKDIR /app
+
+ENTRYPOINT [ "/app/llama-cli" ]
+
+
+FROM intel/oneapi-basekit:$ONEAPI_VERSION AS server
+
+RUN apt-get update \
+    && apt-get install -y libgomp1 curl \
+    && apt autoremove -y \
+    && apt clean -y \
+    && rm -rf /tmp/* /var/tmp/* \
+    && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
+    && find /var/cache -type f -delete
+
+COPY --from=build /app/lib/ /app/
+COPY --from=build /app/build/bin/llama-server /app/
+
+WORKDIR /app
+
+# Must be set to 0.0.0.0 so it can listen to requests from host machine
+ENV LLAMA_ARG_HOST=0.0.0.0
+
+HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
+
+ENTRYPOINT [ "/app/llama-server" ]