Skip to content

Update submodule to latest release b3266 (#127) #11

Update submodule to latest release b3266 (#127)

Update submodule to latest release b3266 (#127) #11

Workflow file for this run

name: CI Nightly Build
on:
push:
tags: ["v[0-9]+.[0-9]+.[0-9]+-[0-9]+.[0-9]+.[0-9]+"]
paths:
[
".github/scripts/**",
".github/workflows/build.yml",
"**/CMakeLists.txt",
"**/Makefile",
"**/*.h",
"**/*.hpp",
"**/*.c",
"**/*.cpp",
"**/*.cu",
"**/*.cc",
"**/*.cxx",
"llama.cpp",
"!docs/**",
"!.gitignore",
"!README.md",
]
workflow_dispatch:
env:
LLM_MODEL_URL: https://delta.jan.ai/tinyllama-1.1b-chat-v0.3.Q2_K.gguf
EMBEDDING_MODEL_URL: https://catalog.jan.ai/dist/models/embeds/nomic-embed-text-v1.5.f16.gguf
jobs:
create-draft-release:
runs-on: ubuntu-latest
if: github.event_name == 'push' && startsWith(github.ref, 'refs/tags/')
outputs:
upload_url: ${{ steps.create_release.outputs.upload_url }}
version: ${{ steps.get_version.outputs.version }}
permissions:
contents: write
steps:
- name: Extract tag name without v prefix
id: get_version
run: echo "VERSION=${GITHUB_REF#refs/tags/v}" >> $GITHUB_ENV && echo "::set-output name=version::${GITHUB_REF#refs/tags/v}"
env:
GITHUB_REF: ${{ github.ref }}
- name: Create Draft Release
id: create_release
uses: softprops/action-gh-release@v2
with:
tag_name: ${{ github.ref_name }}
token: ${{ secrets.GITHUB_TOKEN }}
name: "${{ env.VERSION }}"
draft: false
generate_release_notes: true
prerelease: true
build-and-test:
runs-on: ${{ matrix.runs-on }}
needs: [create-draft-release]
timeout-minutes: 60
strategy:
matrix:
include:
- os: "linux"
name: "amd64-avx2"
runs-on: "ubuntu-20-04"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_NATIVE=OFF -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: true
vulkan: false
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "linux"
name: "amd64-noavx"
runs-on: "ubuntu-20-04"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_AVX=OFF -DGGML_AVX2=OFF -DGGML_FMA=OFF -DGGML_NATIVE=OFF -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "linux"
name: "amd64-avx"
runs-on: "ubuntu-20-04"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_AVX2=OFF -DGGML_NATIVE=OFF -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "linux"
name: "amd64-avx512"
runs-on: "ubuntu-20-04"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_AVX512=ON -DGGML_NATIVE=OFF -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "linux"
name: "amd64-vulkan"
runs-on: "ubuntu-20-04-cuda-11-7"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_VULKAN=ON -DGGML_NATIVE=OFF -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: false
vulkan: true
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "linux"
name: "amd64-avx2-cuda-11-7"
runs-on: "ubuntu-20-04-cuda-11-7"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "linux"
name: "amd64-avx-cuda-11-7"
runs-on: "ubuntu-20-04-cuda-11-7"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_AVX2=OFF -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "linux"
name: "amd64-avx512-cuda-11-7"
runs-on: "ubuntu-20-04-cuda-11-7"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_AVX512=ON -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "linux"
name: "amd64-avx2-cuda-12-0"
runs-on: "ubuntu-20-04-cuda-12-0"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "linux"
name: "amd64-avx-cuda-12-0"
runs-on: "ubuntu-20-04-cuda-12-0"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_AVX2=OFF -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "linux"
name: "amd64-avx512-cuda-12-0"
runs-on: "ubuntu-20-04-cuda-12-0"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_AVX512=ON -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_BUILD_TYPE='Release' -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: "/tmp/sccache.conf"
- os: "mac"
name: "amd64"
runs-on: "macos-13"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_METAL=OFF"
run-e2e: true
vulkan: false
sccache: false
sccache-conf-path: ""
- os: "mac"
name: "arm64"
runs-on: "macos-latest"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DBUILD_SHARED_LIBS=OFF -DGGML_METAL_EMBED_LIBRARY=ON"
run-e2e: false
vulkan: false
sccache: false
sccache-conf-path: ""
- os: "windows"
name: "amd64-avx2"
runs-on: "windows-2019"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_BLAS=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE='Release' -DCMAKE_CXX_COMPILER=cl -DCMAKE_C_COMPILER=cl -GNinja"
run-e2e: true
vulkan: false
sccache: false
sccache-conf-path: ""
- os: "windows"
name: "amd64-avx"
runs-on: "windows-2019"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DGGML_AVX2=OFF -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_BLAS=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE='Release' -DCMAKE_CXX_COMPILER=cl -DCMAKE_C_COMPILER=cl -GNinja"
run-e2e: true
vulkan: false
sccache: false
sccache-conf-path: ""
- os: "windows"
name: "amd64-avx512"
runs-on: "windows-2019"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DGGML_AVX512=ON -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_BLAS=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE='Release' -DCMAKE_CXX_COMPILER=cl -DCMAKE_C_COMPILER=cl -GNinja"
run-e2e: false
vulkan: false
sccache: false
sccache-conf-path: ""
- os: "windows"
name: "amd64-vulkan"
runs-on: "windows-2019"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DGGML_VULKAN=ON -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE='Release' -DCMAKE_CXX_COMPILER=cl -DCMAKE_C_COMPILER=cl -GNinja"
run-e2e: false
vulkan: true
sccache: false
sccache-conf-path: ""
- os: "windows"
name: "amd64-avx2-cuda-12-0"
runs-on: "windows-cuda-12-0"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE='Release' -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_CUDA_COMPILER_LAUNCHER=sccache -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: 'C:\sccache.conf'
- os: "windows"
name: "amd64-avx-cuda-12-0"
runs-on: "windows-cuda-12-0"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DGGML_AVX2=OFF -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE='Release' -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_CUDA_COMPILER_LAUNCHER=sccache -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: 'C:\sccache.conf'
- os: "windows"
name: "amd64-avx512-cuda-12-0"
runs-on: "windows-cuda-12-0"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DGGML_AVX512=ON -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE='Release' -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_CUDA_COMPILER_LAUNCHER=sccache -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: 'C:\sccache.conf'
- os: "windows"
name: "amd64-avx2-cuda-11-7"
runs-on: "windows-cuda-11-7"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE='Release' -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_CUDA_COMPILER_LAUNCHER=sccache -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: 'C:\sccache.conf'
- os: "windows"
name: "amd64-avx-cuda-11-7"
runs-on: "windows-cuda-11-7"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DGGML_AVX2=OFF -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE='Release' -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_CUDA_COMPILER_LAUNCHER=sccache -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: 'C:\sccache.conf'
- os: "windows"
name: "amd64-avx512-cuda-11-7"
runs-on: "windows-cuda-11-7"
cmake-flags: "-DCORTEXLLAMA_VERSION=${{needs.create-draft-release.outputs.version}} -DGGML_AVX512=ON -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_CUDA=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_BUILD_TYPE='Release' -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache -DCMAKE_CUDA_COMPILER_LAUNCHER=sccache -GNinja"
run-e2e: false
vulkan: false
sccache: true
sccache-conf-path: 'C:\sccache.conf'
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: create sccache.conf
if: ${{ matrix.sccache }}
run: |
echo "[cache.s3]" > ${{ matrix.sccache-conf-path }}
echo 'bucket = "${{ secrets.MINIO_BUCKET_NAME }}"' >> ${{ matrix.sccache-conf-path }}
echo 'endpoint = "${{ secrets.MINIO_ENDPOINT }}"' >> ${{ matrix.sccache-conf-path }}
echo 'key_prefix = "${{ matrix.os }}-${{ matrix.name }}"' >> ${{ matrix.sccache-conf-path }}
echo 'use_ssl = false' >> ${{ matrix.sccache-conf-path }}
echo 'server_side_encryption = false' >> ${{ matrix.sccache-conf-path }}
echo 'no_credentials = false' >> ${{ matrix.sccache-conf-path }}
- name: Get Cer for code signing
if: runner.os == 'macOS'
run: base64 -d <<< "$CODE_SIGN_P12_BASE64" > /tmp/codesign.p12
shell: bash
env:
CODE_SIGN_P12_BASE64: ${{ secrets.CODE_SIGN_P12_BASE64 }}
- uses: apple-actions/import-codesign-certs@v2
if: runner.os == 'macOS'
with:
p12-file-base64: ${{ secrets.CODE_SIGN_P12_BASE64 }}
p12-password: ${{ secrets.CODE_SIGN_P12_PASSWORD }}
- uses: actions/setup-dotnet@v3
if: runner.os == 'Windows'
with:
dotnet-version: "8.0.x"
- name: Install choco on Windows
if: runner.os == 'Windows'
run: |
choco install make sccache ninja -y
- name: Install ninja build on Linux
if: runner.os == 'Linux'
run: |
sudo apt install ninja-build -y
- name: Add msbuild to PATH
if: runner.os == 'Windows'
uses: ilammy/[email protected]
- name: start scache server
if: ${{ matrix.sccache }}
run: |
sccache --start-server
env:
SCCACHE_BUCKET: "${{ secrets.MINIO_BUCKET_NAME }}"
SCCACHE_REGION: "${{ secrets.MINIO_REGION }}"
SCCACHE_ENDPOINT: "${{ secrets.MINIO_ENDPOINT }}"
SCCACHE_S3_USE_SSL: "false"
SCCACHE_S3_SERVER_SIDE_ENCRYPTION: "false"
SCCACHE_S3_KEY_PREFIX: "${{ matrix.os }}-${{ matrix.name }}"
SCCACHE_LOG: "debug"
SCCACHE_CONF: '${{ matrix.sccache-conf-path }}'
AWS_ACCESS_KEY_ID: "${{ secrets.MINIO_ACCESS_KEY_ID }}"
AWS_SECRET_ACCESS_KEY: "${{ secrets.MINIO_SECRET_ACCESS_KEY }}"
SCCACHE_IDLE_TIMEOUT: "0"
- name: Build
run: |
make build-example-server CMAKE_EXTRA_FLAGS="${{ matrix.cmake-flags }}"
- name: Pre Package
run: |
make pre-package
- name: Code Signing macOS
if: runner.os == 'macOS'
run: |
make codesign CODE_SIGN=true DEVELOPER_ID="${{ secrets.DEVELOPER_ID }}"
- name: Code Signing Windows
if: runner.os == 'Windows'
shell: cmd
run: |
set PATH=%PATH%;%USERPROFILE%\.dotnet\tools
make codesign CODE_SIGN=true AZURE_KEY_VAULT_URI="${{ secrets.AZURE_KEY_VAULT_URI }}" AZURE_CLIENT_ID="${{ secrets.AZURE_CLIENT_ID }}" AZURE_TENANT_ID="${{ secrets.AZURE_TENANT_ID }}" AZURE_CLIENT_SECRET="${{ secrets.AZURE_CLIENT_SECRET }}" AZURE_CERT_NAME="${{ secrets.AZURE_CERT_NAME }}"
- name: Package
run: |
make package
- name: Run e2e testing
if: ${{ matrix.run-e2e }}
run: |
make run-e2e-test LLM_MODEL_URL=${{ 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
- uses: actions/[email protected]
if: github.event_name == 'push' && startsWith(github.ref, 'refs/tags/')
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
with:
upload_url: ${{ needs.create-draft-release.outputs.upload_url }}
asset_path: ./cortex.llamacpp.tar.gz
asset_name: cortex.llamacpp-${{ needs.create-draft-release.outputs.version }}-${{ matrix.os }}-${{ matrix.name }}.tar.gz
asset_content_type: application/gzip
- name: Clean
if: always()
continue-on-error: true
run: |
sccache --stop-server
rm ${{ matrix.sccache-conf-path }}