Merge pull request #430 from containers/llama-cpp-python-server

Install llama-cpp-python[server]

container-images/asahi/Containerfile

@@ -1,25 +1,11 @@
 FROM fedora:41
 
-# renovate: datasource=github-releases depName=containers/omlmd extractVersion=^v(?<version>.*)
-ARG OMLMD_VERSION=0.1.6
 ARG LLAMA_CPP_SHA=1329c0a75e6a7defc5c380eaf80d8e0f66d7da78
 # renovate: datasource=git-refs depName=ggerganov/whisper.cpp packageName=https://github.com/ggerganov/whisper.cpp gitRef=master versioning=loose type=digest
 ARG WHISPER_CPP_SHA=31aea563a83803c710691fed3e8d700e06ae6788
 
-RUN dnf install -y dnf-plugins-core && \
-    dnf copr enable -y @asahi/fedora-remix-branding && \
-    dnf install -y asahi-repos && \
-    dnf install -y mesa-vulkan-drivers vulkan-headers vulkan-loader-devel \
-      vulkan-tools spirv-tools glslc glslang git procps-ng vim cmake gcc-c++ \
-      python3-pip python3-argcomplete && \
-    dnf clean all && \
-    rm -rf /var/cache/*dnf*
-
-RUN /usr/bin/python3 --version
-RUN pip install "omlmd==${OMLMD_VERSION}"
-
 COPY ../scripts /scripts
 RUN chmod +x /scripts/*.sh && \
-    /scripts/build_llama_and_whisper.sh "$LLAMA_CPP_SHA" "$WHISPER_CPP_SHA" \
-      "/usr" "-DGGML_VULKAN=1"
+    /scripts/build_llama_and_whisper.sh "asahi" "$LLAMA_CPP_SHA" \
+      "$WHISPER_CPP_SHA" "/usr" "-DGGML_VULKAN=1"
 

container-images/cuda/Containerfile

@@ -5,32 +5,17 @@ ARG LLAMA_CPP_SHA=1329c0a75e6a7defc5c380eaf80d8e0f66d7da78
 # renovate: datasource=git-refs depName=ggerganov/whisper.cpp packageName=https://github.com/ggerganov/whisper.cpp gitRef=master versioning=loose type=digest
 ARG WHISPER_CPP_SHA=31aea563a83803c710691fed3e8d700e06ae6788
 
-# Install dependencies only needed for building
-RUN dnf install -y git cmake gcc-c++ && \
-    dnf clean all && rm -rf /var/cache/*dnf*
-
-# Set the temporary installation directory
-ARG INSTALL_PREFIX=/tmp/install
-
 COPY ../scripts /scripts
 RUN chmod +x /scripts/*.sh && \
-    /scripts/build_llama_and_whisper.sh "$LLAMA_CPP_SHA" "$WHISPER_CPP_SHA" \
-      "$INSTALL_PREFIX" \
+    /scripts/build_llama_and_whisper.sh "cuda" "$LLAMA_CPP_SHA" \
+      "$WHISPER_CPP_SHA" "/tmp/install" \
       "-DGGML_CUDA=1" "-DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined"
 
 # Final runtime image
 FROM docker.io/nvidia/cuda:12.6.2-runtime-ubi9
 
-# renovate: datasource=github-releases depName=containers/omlmd extractVersion=^v(?<version>.*)
-ARG OMLMD_VERSION=0.1.6
-
-# Install minimal runtime dependencies
-RUN dnf install -y python3 python3-pip && \
-    dnf clean all && \
-    rm -rf /var/cache/*dnf*
-
-# Install Python packages in the runtime image
-RUN pip install "omlmd==${OMLMD_VERSION}"
+RUN dnf install -y python3 && \
+    dnf clean all && rm -rf /var/cache/*dnf*
 
 # Copy the entire installation directory from the builder
 COPY --from=builder /tmp/install /usr

container-images/ramalama/Containerfile

@@ -1,42 +1,13 @@
 FROM registry.access.redhat.com/ubi9/ubi:9.4-1214.1729773476
 
-# renovate: datasource=github-releases depName=containers/omlmd extractVersion=^v(?<version>.*)
-ARG OMLMD_VERSION=0.1.6
 ARG LLAMA_CPP_SHA=1329c0a75e6a7defc5c380eaf80d8e0f66d7da78
 # renovate: datasource=git-refs depName=ggerganov/whisper.cpp packageName=https://github.com/ggerganov/whisper.cpp gitRef=master versioning=loose type=digest
 ARG WHISPER_CPP_SHA=31aea563a83803c710691fed3e8d700e06ae6788
 
-# vulkan-headers vulkan-loader-devel vulkan-tools glslc glslang python3-pip mesa-libOpenCL-$MESA_VER.aarch64
-RUN dnf install -y https://dl.fedoraproject.org/pub/epel/epel-release-latest-9.noarch.rpm && \
-    crb enable && \
-    dnf install -y epel-release && \
-    dnf --enablerepo=ubi-9-appstream-rpms install -y git procps-ng vim \
-      dnf-plugins-core python3-dnf-plugin-versionlock cmake gcc-c++ \
-      python3-pip python3-argcomplete && \
-    dnf copr enable -y slp/mesa-krunkit epel-9-$(uname -m) && \
-    dnf install -y mesa-vulkan-drivers-23.3.3-102.el9 \
-      vulkan-headers vulkan-loader-devel vulkan-tools spirv-tools glslc && \
-    dnf clean all && \
-    rm -rf /var/cache/*dnf*
-
-RUN /usr/bin/python3 --version
-RUN pip install "omlmd==${OMLMD_VERSION}"
-
-RUN dnf config-manager --add-repo \
-      https://mirror.stream.centos.org/9-stream/AppStream/$(uname -m)/os/
-RUN curl --retry 8 --retry-all-errors -o \
-      /etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-Official \
-      http://mirror.centos.org/centos/RPM-GPG-KEY-CentOS-Official && \
-      cat /etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-Official
-RUN rpm --import /etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-Official
-RUN dnf install -y glslang && \
-    dnf clean all && \
-    rm -rf /var/cache/*dnf*
-
 COPY ../scripts /scripts
 RUN chmod +x /scripts/*.sh && \
-    /scripts/build_llama_and_whisper.sh "$LLAMA_CPP_SHA" "$WHISPER_CPP_SHA" \
-      "/usr" "-DGGML_KOMPUTE=1"
+    /scripts/build_llama_and_whisper.sh "ramalama" "$LLAMA_CPP_SHA" \
+      "$WHISPER_CPP_SHA" "/usr" "-DGGML_KOMPUTE=1"
 
 ENV WHISPER_CPP_SHA=${WHISPER_CPP_SHA}
 ENV LLAMA_CPP_SHA=${LLAMA_CPP_SHA}

container-images/rocm/Containerfile

@@ -1,27 +1,12 @@
 FROM quay.io/ramalama/ramalama:latest
 
-RUN /usr/bin/python3 --version
-
 ARG ROCM_VERSION=6.2.2
 ARG AMDGPU_VERSION=6.2.2
 
 COPY rocm/amdgpu.repo /etc/yum.repos.d/
 COPY rocm/rocm.repo /etc/yum.repos.d/
-
-RUN dnf config-manager --add-repo \
-      https://mirror.stream.centos.org/9-stream/AppStream/$(uname -m)/os/
-RUN curl --retry 8 --retry-all-errors -o \
-      /etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-Official \
-      http://mirror.centos.org/centos/RPM-GPG-KEY-CentOS-Official && \
-      cat /etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-Official
-RUN rpm --import /etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-Official
-
 COPY scripts /scripts
-RUN dnf install -y rocm-dev hipblas-devel rocblas-devel && \
-    dnf clean all && \
-    chmod +x /scripts/*.sh && \
-    /scripts/build_llama_and_whisper.sh "$LLAMA_CPP_SHA" "$WHISPER_CPP_SHA" \
-      "/usr" "-DGGML_HIPBLAS=1" && \
-    rm -rf /var/cache/*dnf* /opt/rocm-*/lib/llvm \
-      /opt/rocm-*/lib/rocblas/library/*gfx9*
+RUN chmod +x /scripts/*.sh && \
+    /scripts/build_llama_and_whisper.sh "rocm" "$LLAMA_CPP_SHA" \
+      "$WHISPER_CPP_SHA" "/usr" "-DGGML_HIPBLAS=1"
 

container-images/scripts/build_llama_and_whisper.sh

@@ -1,34 +1,85 @@
 #!/bin/bash
 
+dnf_install() {
+  local rpm_list=("python3" "python3-pip" "gcc-c++" "cmake" "vim" "procps-ng" \
+                  "git")
+  if [ "$containerfile" = "ramalama" ]; then
+    dnf install -y https://dl.fedoraproject.org/pub/epel/epel-release-latest-9.noarch.rpm
+    crb enable
+    dnf install -y epel-release
+    rpm_list+=("dnf-plugins-core" "python3-dnf-plugin-versionlock" \
+               "python3-argcomplete")
+    dnf --enablerepo=ubi-9-appstream-rpms install -y "${rpm_list[@]}"
+    local uname_m
+    uname_m="$(uname -m)"
+    dnf copr enable -y slp/mesa-krunkit "epel-9-$uname_m"
+    dnf install -y mesa-vulkan-drivers-23.3.3-102.el9 vulkan-headers \
+      vulkan-loader-devel vulkan-tools spirv-tools glslc
+    local url="https://mirror.stream.centos.org/9-stream/AppStream/$uname_m/os/"
+    dnf config-manager --add-repo "$url"
+    url="http://mirror.centos.org/centos/RPM-GPG-KEY-CentOS-Official"
+    curl --retry 8 --retry-all-errors -o \
+      /etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-Official "$url"
+    rpm --import /etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-Official
+    dnf install -y glslang
+  elif [ "$containerfile" = "asahi" ]; then
+    dnf install -y dnf-plugins-core
+    dnf copr enable -y @asahi/fedora-remix-branding
+    dnf install -y asahi-repos
+    dnf install -y mesa-vulkan-drivers vulkan-headers vulkan-loader-devel \
+      vulkan-tools spirv-tools glslc glslang python3-argcomplete \
+      "${rpm_list[@]}"
+  elif [ "$containerfile" = "rocm" ]; then
+    dnf install -y rocm-dev hipblas-devel rocblas-devel
+  else
+    dnf install -y "${rpm_list[@]}"
+  fi
+}
+
+cmake_steps() {
+  local flag="$1"
+  cmake -B build "${common_flags[@]}" "$flag"
+  cmake --build build --config Release -j"$(nproc)"
+  cmake --install build
+}
+
 main() {
   set -e
 
-  local llama_cpp_sha="$1"
-  local whisper_cpp_sha="$2"
-  local install_prefix="$3"
-  local build_flag_1="$4"
-  local build_flag_2="$5"
+  local containerfile="$1"
+  local llama_cpp_sha="$2"
+  local whisper_cpp_sha="$3"
+  local install_prefix="$4"
+  local build_flag_1="$5"
+  local build_flag_2="$6"
+  local common_flags=("-DGGML_CCACHE=0" \
+                      "-DCMAKE_INSTALL_PREFIX=$install_prefix" "$build_flag_1")
+  if [ -n "$build_flag_2" ]; then
+    common_flags+=("$build_flag_2")
+  fi
+
+  dnf_install
+
   git clone https://github.com/ggerganov/llama.cpp
   cd llama.cpp
   git reset --hard "$llama_cpp_sha"
-  cmake -B build -DGGML_CCACHE=0 "$build_flag_1" "$build_flag_2" \
-    -DCMAKE_INSTALL_PREFIX="$install_prefix"
-  cmake --build build --config Release -j"$(nproc)"
-  cmake --install build
+  cmake_steps
   cd ..
-  rm -rf llama.cpp
 
   git clone https://github.com/ggerganov/whisper.cpp
   cd whisper.cpp
   git reset --hard "$whisper_cpp_sha"
-  cmake -B build -DGGML_CCACHE=0 "$build_flag_1" "$build_flag_2" \
-    -DBUILD_SHARED_LIBS=NO -DCMAKE_INSTALL_PREFIX="$install_prefix"
-  cmake --build build --config Release -j"$(nproc)"
-  cmake --install build
+  cmake_steps "-DBUILD_SHARED_LIBS=NO"
   mv build/bin/main "$install_prefix/bin/whisper-main"
   mv build/bin/server "$install_prefix/bin/whisper-server"
   cd ..
-  rm -rf whisper.cpp
+
+  CMAKE_ARGS="${common_flags[*]}" FORCE_CMAKE=1 \
+    pip install --prefix="$install_prefix" 'llama-cpp-python[server]'
+
+  dnf clean all
+  rm -rf /var/cache/*dnf* /opt/rocm-*/lib/llvm \
+    /opt/rocm-*/lib/rocblas/library/*gfx9* llama.cpp whisper.cpp
 }
 
 main "$@"

container-images/vulkan/Containerfile

@@ -2,6 +2,6 @@ FROM quay.io/ramalama/ramalama:latest
 
 COPY ../scripts /scripts
 RUN chmod +x /scripts/*.sh && \
-    /scripts/build_llama_and_whisper.sh "$LLAMA_CPP_SHA" "$WHISPER_CPP_SHA" \
-      "/usr" "-DGGML_VULKAN=1"
+    /scripts/build_llama_and_whisper.sh "vulkan" "$LLAMA_CPP_SHA" \
+      "$WHISPER_CPP_SHA" "/usr" "-DGGML_VULKAN=1"