Update llama.cpp submodule to latest release b3078 #14
Workflow file for this run
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: CI Quality Gate Submodule | |
| on: | |
| pull_request: | |
| types: [opened, synchronize, reopened] | |
| workflow_dispatch: | |
| env: | |
| LLM_MODEL_URL: https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GGUF/resolve/main/mistral-7b-instruct-v0.2.Q4_K_M.gguf | |
| EMBEDDING_MODEL_URL: https://catalog.jan.ai/dist/models/embeds/nomic-embed-text-v1.5.f16.gguf | |
| jobs: | |
| build-and-test: | |
| if: startsWith(github.head_ref, 'update-submodule') | |
| runs-on: ${{ matrix.runs-on }} | |
| timeout-minutes: 60 | |
| strategy: | |
| fail-fast: false | |
| matrix: | |
| include: | |
| - os: "linux" | |
| name: "amd64-avx2" | |
| runs-on: "ubuntu-18-04" | |
| cmake-flags: "-DLLAMA_NATIVE=OFF" | |
| run-e2e: true | |
| vulkan: false | |
| - os: "linux" | |
| name: "amd64-avx" | |
| runs-on: "ubuntu-18-04" | |
| cmake-flags: "-DLLAMA_AVX2=OFF -DLLAMA_NATIVE=OFF" | |
| run-e2e: false | |
| vulkan: false | |
| - os: "linux" | |
| name: "amd64-avx512" | |
| runs-on: "ubuntu-18-04" | |
| cmake-flags: "-DLLAMA_AVX512=ON -DLLAMA_NATIVE=OFF" | |
| run-e2e: false | |
| vulkan: false | |
| - os: "linux" | |
| name: "amd64-vulkan" | |
| runs-on: "ubuntu-18-04-cuda-11-7" | |
| cmake-flags: "-DLLAMA_VULKAN=ON -DLLAMA_NATIVE=OFF" | |
| run-e2e: false | |
| vulkan: true | |
| - os: "linux" | |
| name: "amd64-cuda-11-7" | |
| runs-on: "ubuntu-18-04-cuda-11-7" | |
| cmake-flags: "-DLLAMA_NATIVE=OFF -DLLAMA_CUDA=ON" | |
| run-e2e: false | |
| vulkan: false | |
| - os: "linux" | |
| name: "amd64-cuda-12-0" | |
| runs-on: "ubuntu-18-04-cuda-12-0" | |
| cmake-flags: "-DLLAMA_NATIVE=OFF -DLLAMA_CUDA=ON" | |
| run-e2e: false | |
| vulkan: false | |
| - os: "mac" | |
| name: "amd64" | |
| runs-on: "macos-13" | |
| cmake-flags: "-DLLAMA_METAL=OFF" | |
| run-e2e: true | |
| vulkan: false | |
| - os: "mac" | |
| name: "arm64" | |
| runs-on: "mac-silicon" | |
| cmake-flags: "-DLLAMA_METAL_EMBED_LIBRARY=ON" | |
| run-e2e: true | |
| vulkan: false | |
| - os: "windows" | |
| name: "amd64-avx2" | |
| runs-on: "windows-latest" | |
| cmake-flags: "-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_BLAS=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE=RELEASE" | |
| run-e2e: true | |
| vulkan: false | |
| - os: "windows" | |
| name: "amd64-avx" | |
| runs-on: "windows-latest" | |
| cmake-flags: "-DLLAMA_AVX2=OFF -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_BLAS=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE=RELEASE" | |
| run-e2e: true | |
| vulkan: false | |
| - os: "windows" | |
| name: "amd64-avx512" | |
| runs-on: "windows-cuda-11-7" | |
| cmake-flags: "-DLLAMA_AVX512=ON -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_BLAS=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE=RELEASE" | |
| run-e2e: true | |
| vulkan: false | |
| - os: "windows" | |
| name: "amd64-vulkan" | |
| runs-on: "windows-latest" | |
| cmake-flags: "-DLLAMA_VULKAN=ON -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE=RELEASE" | |
| run-e2e: false | |
| vulkan: true | |
| - os: "windows" | |
| name: "amd64-avx2-cuda-12-0" | |
| runs-on: "windows-cuda-12-0" | |
| cmake-flags: "-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE=RELEASE" | |
| run-e2e: false | |
| vulkan: false | |
| - os: "windows" | |
| name: "amd64-avx-cuda-12-0" | |
| runs-on: "windows-cuda-12-0" | |
| cmake-flags: "-DLLAMA_AVX2=OFF -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE=RELEASE" | |
| run-e2e: false | |
| vulkan: false | |
| - os: "windows" | |
| name: "amd64-avx512-cuda-12-0" | |
| runs-on: "windows-cuda-12-0" | |
| cmake-flags: "-DLLAMA_AVX512=ON -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE=RELEASE" | |
| run-e2e: false | |
| vulkan: false | |
| - os: "windows" | |
| name: "amd64-avx2-cuda-11-7" | |
| runs-on: "windows-cuda-11-7" | |
| cmake-flags: "-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE=RELEASE" | |
| run-e2e: false | |
| vulkan: false | |
| - os: "windows" | |
| name: "amd64-avx-cuda-11-7" | |
| runs-on: "windows-cuda-11-7" | |
| cmake-flags: "-DLLAMA_AVX2=OFF -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE=RELEASE" | |
| run-e2e: false | |
| vulkan: false | |
| - os: "windows" | |
| name: "amd64-avx512-cuda-11-7" | |
| runs-on: "windows-cuda-11-7" | |
| cmake-flags: "-DLLAMA_AVX512=ON -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE=RELEASE" | |
| run-e2e: false | |
| vulkan: false | |
| steps: | |
| - name: Clone | |
| id: checkout | |
| uses: actions/checkout@v3 | |
| with: | |
| submodules: recursive | |
| - name: Prepare Vulkan SDK | |
| if: ${{ matrix.vulkan }} | |
| uses: humbletim/[email protected] | |
| with: | |
| vulkan-query-version: 1.3.275.0 | |
| vulkan-components: Vulkan-Headers, Vulkan-Loader | |
| vulkan-use-cache: true | |
| - name: Install choco on Windows | |
| if: runner.os == 'Windows' | |
| run: | | |
| choco install make -y | |
| - name: Build | |
| run: | | |
| make build-example-server CMAKE_EXTRA_FLAGS="${{ matrix.cmake-flags }}" | |
| - name: Pre Package | |
| run: | | |
| make pre-package | |
| - name: Package | |
| run: | | |
| make package | |
| - name: Install Python | |
| if: ${{matrix.run-e2e}} | |
| uses: actions/setup-python@v4 | |
| with: | |
| python-version: '3.10' | |
| - name: Run e2e submodule testing | |
| if: ${{ matrix.run-e2e }} | |
| run: | | |
| make run-e2e-submodule-test LLM_MODEL_URL=${{ env.LLM_MODEL_URL }} EMBEDDING_MODEL_URL=${{ env.EMBEDDING_MODEL_URL }} | |
| - name: Upload Artifact | |
| uses: actions/upload-artifact@v2 | |
| with: | |
| name: cortex.llamacpp-${{ matrix.os }}-${{ matrix.name }} | |
| path: ./cortex.llamacpp |