Skip to content

Add performance statistics for speculative decoding #4909

Add performance statistics for speculative decoding

Add performance statistics for speculative decoding #4909

Workflow file for this run

name: Windows (VS 2019, Python 3.11)
on:
workflow_dispatch:
pull_request:
merge_group:
push:
branches:
- master
- 'releases/**'
permissions: read-all # Required by https://github.com/ossf/scorecard/blob/e23b8ad91fd6a64a0a971ca4fc0a4d1650725615/docs/checks.md#token-permissions
concurrency:
# github.ref is not unique in post-commit
group: ${{ github.event_name == 'push' && github.run_id || github.ref }}-windows
cancel-in-progress: true
env:
PYTHON_VERSION: '3.11'
OV_BRANCH: 0080d90974ca84f9a6d359da3388a2a18a93b753
OV_TARBALL: ''
jobs:
openvino_download:
name: Download OpenVINO package
outputs:
status: ${{ steps.openvino_download.outcome }}
timeout-minutes: 10
defaults:
run:
shell: bash
runs-on: ubuntu-20.04
steps:
- name: Download OpenVINO build
id: openvino_download
run: |
wget ${{ env.OV_TARBALL}} --progress=bar:force:noscroll -O openvino_package.zip
unzip -l openvino_package.zip
continue-on-error: true
#
# Upload to artifacts
#
- name: Upload openvino package
if: steps.openvino_download.outcome == 'success'
uses: actions/upload-artifact@v4
with:
name: openvino_package
path: openvino_package.zip
if-no-files-found: 'error'
openvino_build:
name: Build OpenVINO package
needs: [openvino_download]
if: needs.openvino_download.outputs.status != 'success'
timeout-minutes: 150
defaults:
run:
shell: pwsh
runs-on: windows-2019-16-core
env:
CMAKE_BUILD_TYPE: 'Release'
CMAKE_GENERATOR: 'Ninja Multi-Config'
CMAKE_CXX_COMPILER_LAUNCHER: ccache
CMAKE_C_COMPILER_LAUNCHER: ccache
OPENVINO_REPO: ${{ github.workspace }}\\openvino
INSTALL_DIR: ${{ github.workspace }}\\openvino\\install
BUILD_DIR: ${{ github.workspace }}\\openvino\\build
steps:
- name: git configuration
run: git config --system core.longpaths true
- name: Clone OpenVINO
uses: actions/checkout@v4
with:
repository: 'openvinotoolkit/openvino'
path: ${{ env.OPENVINO_REPO }}
submodules: 'true'
ref: ${{ env.OV_BRANCH }}
#
# Dependencies
#
- name: Setup Python ${{ env.PYTHON_VERSION }}
uses: actions/setup-python@v5
with:
python-version: ${{ env.PYTHON_VERSION }}
cache: 'pip'
- name: Install python dependencies
run: |
# For Python API: build and wheel packaging
python3 -m pip install -r ${env:OPENVINO_REPO}/src/bindings/python/wheel/requirements-dev.txt
- name: Install build dependencies
run: |
Invoke-WebRequest https://github.com/ninja-build/ninja/releases/download/v1.11.1/ninja-win.zip -OutFile ninja-win.zip -MaximumRetryCount 10
Expand-Archive -Force ninja-win.zip
# Add it to the GitHub Path so it would be available in the subsequent steps
Add-Content -Path $env:GITHUB_PATH -Value "${{ github.workspace }}/ninja-win"
#
# Build
#
- name: Setup ccache
uses: hendrikmuhs/[email protected]
with:
max-size: "2000M"
# Should save cache only if run in the master branch of the base repo
# github.ref_name is 'ref/PR_#' in case of the PR, and 'branch_name' when executed on push
save: ${{ github.ref_name == 'master' && 'true' || 'false' }}
verbose: 2
key: ccache-windows
restore-keys: |
ccache-windows
- name: Configure Developer Command Prompt for Microsoft Visual C++
uses: ilammy/msvc-dev-cmd@0b201ec74fa43914dc39ae48a89fd1d8cb592756 # v1.13.0
- name: CMake configure - OpenVINO
run: |
cmake -G "${{ env.CMAKE_GENERATOR }}" `
-DENABLE_CPPLINT=OFF `
-DBUILD_nvidia_plugin=OFF `
-DBUILD_SHARED_LIBS=ON `
-DENABLE_TESTS=OFF `
-DCMAKE_COMPILE_WARNING_AS_ERROR=OFF `
-DENABLE_STRICT_DEPENDENCIES=OFF `
-DENABLE_PYTHON=ON `
-DENABLE_WHEEL=ON `
-DENABLE_JS=OFF `
-DENABLE_SAMPLES=ON `
-DENABLE_OV_ONNX_FRONTEND=OFF `
-DENABLE_OV_PADDLE_FRONTEND=OFF `
-DENABLE_OV_PYTORCH_FRONTEND=ON `
-DENABLE_OV_TF_FRONTEND=ON `
-DENABLE_OV_TF_LITE_FRONTEND=OFF `
-DENABLE_INTEL_GPU=OFF `
-DENABLE_INTEL_NPU=ON `
-DCMAKE_DISABLE_FIND_PACKAGE_PkgConfig=ON `
-S ${{ env.OPENVINO_REPO }} `
-B ${{ env.BUILD_DIR }}
- name: Clean ccache stats
run: ccache --zero-stats --show-config
- name: Cmake build - OpenVINO
run: cmake --build ${{ env.BUILD_DIR }} --parallel --config ${{ env.CMAKE_BUILD_TYPE }} --verbose
- name: Show ccache stats
run: ccache --show-stats
- name: Cmake install - OpenVINO
run: |
cmake -DCMAKE_INSTALL_PREFIX=${{ env.INSTALL_DIR }}/ov_package -P ${{ env.BUILD_DIR }}/cmake_install.cmake
cmake -DCMAKE_INSTALL_PREFIX=${{ env.INSTALL_DIR }}/ov_package -DCOMPONENT=python_wheels -P ${{ env.BUILD_DIR }}/cmake_install.cmake
- name: Pack Artifacts
run: |
$file=Get-ChildItem -Path "${{ env.INSTALL_DIR }}"
$compress = @{
Path = $file
CompressionLevel = "Optimal"
DestinationPath = "${{ env.BUILD_DIR }}/openvino_package.zip"
}
Compress-Archive @compress
#
# Upload build artifacts and logs
#
- name: Upload openvino package
if: ${{ always() }}
uses: actions/upload-artifact@v4
with:
name: openvino_package
path: ${{ env.BUILD_DIR }}/openvino_package.zip
if-no-files-found: 'error'
genai_python_lib:
name: OpenVINO genai extension (cmake + wheel)
needs: [ openvino_download, openvino_build ]
if: |
always() &&
(needs.openvino_download.outputs.status == 'success' || needs.openvino_build.result == 'success')
timeout-minutes: 90
defaults:
run:
shell: pwsh
runs-on: windows-2019
env:
OV_INSTALL_DIR: ${{ github.workspace }}\\ov
CMAKE_BUILD_PARALLEL_LEVEL: null
steps:
- name: Clone openvino.genai
uses: actions/checkout@v4
with:
submodules: recursive
- name: Setup Python ${{ env.PYTHON_VERSION }}
uses: actions/setup-python@v5
with:
python-version: ${{ env.PYTHON_VERSION }}
cache: 'pip'
- name: Download OpenVINO package
uses: actions/download-artifact@v4
with:
name: openvino_package
path: ${{ env.OV_INSTALL_DIR }}
- name: Extract OpenVINO packages
run: |
pushd ${{ env.OV_INSTALL_DIR }}
Expand-Archive openvino_package.zip -DestinationPath ./tmp
mv ./tmp/*/* .
popd
- name: Configure Developer Command Prompt for Microsoft Visual C++
uses: ilammy/msvc-dev-cmd@0b201ec74fa43914dc39ae48a89fd1d8cb592756 # v1.13.0
- name: Build genai libs
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
cmake -DCMAKE_BUILD_TYPE=Release -S ./ -B ./build/
cmake --build ./build/ --config Release -j
- name: Test bindings
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
python -m pip install ./thirdparty/openvino_tokenizers/[transformers] -r ./tests/python_tests/requirements.txt --find-links ${env:OV_INSTALL_DIR}/wheels
python -m pytest -v ./tests/python_tests/test_chat_generate_api.py::test_set_chat_template
env:
PYTHONPATH: "./build/" # cmd evaluates variables in a different way. Setting PYTHONPATH before setupvars.bat instead of doing that after solves that.
- name: Test bindings (wheel)
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
python -m pip install . --verbose --find-links ${env:OV_INSTALL_DIR}/wheels
python -m pip install ./tools/who_what_benchmark --find-links ${env:OV_INSTALL_DIR}/wheels
python -m pytest -v ./tests/python_tests/ --ignore ./tests/python_tests/test_whisper_generate_api.py --ignore ./tests/python_tests/test_vlm_api.py -k "not test_set_chat_template"
genai_python_lib_whisper:
name: OpenVINO genai extension whisper tests (cmake + wheel)
needs: [ openvino_download, openvino_build ]
if: |
always() &&
(needs.openvino_download.outputs.status == 'success' || needs.openvino_build.result == 'success')
timeout-minutes: 90
defaults:
run:
shell: pwsh
runs-on: windows-2019
env:
OV_INSTALL_DIR: ${{ github.workspace }}\\ov
CMAKE_BUILD_PARALLEL_LEVEL: null
steps:
- name: Clone openvino.genai
uses: actions/checkout@v4
with:
submodules: recursive
- name: Setup Python ${{ env.PYTHON_VERSION }}
uses: actions/setup-python@v5
with:
python-version: ${{ env.PYTHON_VERSION }}
cache: 'pip'
- name: Download OpenVINO package
uses: actions/download-artifact@v4
with:
name: openvino_package
path: ${{ env.OV_INSTALL_DIR }}
- name: Extract OpenVINO packages
run: |
pushd ${{ env.OV_INSTALL_DIR }}
Expand-Archive openvino_package.zip -DestinationPath ./tmp
mv ./tmp/*/* .
popd
- name: Configure Developer Command Prompt for Microsoft Visual C++
uses: ilammy/msvc-dev-cmd@0b201ec74fa43914dc39ae48a89fd1d8cb592756 # v1.13.0
- name: Build genai libs
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
cmake -DCMAKE_BUILD_TYPE=Release -S ./ -B ./build/
cmake --build ./build/ --config Release --target py_openvino_genai -j
- name: Test bindings
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
python -m pip install ./thirdparty/openvino_tokenizers/[transformers] -r ./tests/python_tests/requirements.txt --find-links ${env:OV_INSTALL_DIR}/wheels
python -m pytest -v ./tests/python_tests/test_whisper_generate_api.py -k test_smoke
env:
PYTHONPATH: "./build/" # cmd evaluates variables in a different way. Setting PYTHONPATH before setupvars.bat instead of doing that after solves that.
- name: Test bindings (wheel)
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
python -m pip install . --verbose --find-links ${env:OV_INSTALL_DIR}/wheels
python -m pip install ./tools/who_what_benchmark --find-links ${env:OV_INSTALL_DIR}/wheels
python -m pytest -v ./tests/python_tests/test_whisper_generate_api.py -k "not test_smoke"
genai_python_lib_vlm:
name: OpenVINO genai VLM tests (cmake + wheel)
needs: [ openvino_download, openvino_build ]
if: |
always() &&
(needs.openvino_download.outputs.status == 'success' || needs.openvino_build.result == 'success')
timeout-minutes: 90
defaults:
run:
shell: pwsh
runs-on: windows-2019-16-core
env:
OV_INSTALL_DIR: ${{ github.workspace }}\\ov
CMAKE_BUILD_PARALLEL_LEVEL: null
steps:
- name: Clone openvino.genai
uses: actions/checkout@v4
with:
submodules: recursive
- name: Setup Python ${{ env.PYTHON_VERSION }}
uses: actions/setup-python@v5
with:
python-version: ${{ env.PYTHON_VERSION }}
cache: 'pip'
- name: Download OpenVINO package
uses: actions/download-artifact@v4
with:
name: openvino_package
path: ${{ env.OV_INSTALL_DIR }}
- name: Extract OpenVINO packages
run: |
pushd ${{ env.OV_INSTALL_DIR }}
Expand-Archive openvino_package.zip -DestinationPath ./tmp
mv ./tmp/*/* .
popd
- name: Configure Developer Command Prompt for Microsoft Visual C++
uses: ilammy/msvc-dev-cmd@0b201ec74fa43914dc39ae48a89fd1d8cb592756 # v1.13.0
- name: Build genai libs
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
cmake -DCMAKE_BUILD_TYPE=Release -S ./ -B ./build/
cmake --build ./build/ --config Release --target py_openvino_genai -j
- name: Test bindings
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
python -m pip install ./thirdparty/openvino_tokenizers/[transformers] -r ./tests/python_tests/requirements.txt --find-links ${env:OV_INSTALL_DIR}/wheels
python -m pytest -v ./tests/python_tests/test_vlm_api.py
env:
PYTHONPATH: "./build/" # cmd evaluates variables in a different way. Setting PYTHONPATH before setupvars.bat instead of doing that after solves that.
genai_package:
name: OpenVINO genai extension (install to OpenVINO package)
strategy:
matrix:
build-type: [Release, Debug]
needs: [ openvino_download, openvino_build ]
if: |
always() &&
(needs.openvino_download.outputs.status == 'success' || needs.openvino_build.result == 'success')
timeout-minutes: 60
defaults:
run:
shell: pwsh
runs-on: windows-2019
env:
OV_INSTALL_DIR: ${{ github.workspace }}\\ov
CMAKE_BUILD_PARALLEL_LEVEL: null
steps:
- name: Clone openvino.genai
uses: actions/checkout@v4
with:
submodules: recursive
- name: Setup Python ${{ env.PYTHON_VERSION }}
uses: actions/setup-python@v5
with:
python-version: ${{ env.PYTHON_VERSION }}
cache: 'pip'
- name: Download OpenVINO package
uses: actions/download-artifact@v4
with:
name: openvino_package
path: ${{ env.OV_INSTALL_DIR }}
- name: Extract OpenVINO packages
run: |
pushd ${{ env.OV_INSTALL_DIR }}
Expand-Archive openvino_package.zip -DestinationPath ./tmp
mv ./tmp/*/* .
popd
- name: Configure Developer Command Prompt for Microsoft Visual C++
uses: ilammy/msvc-dev-cmd@0b201ec74fa43914dc39ae48a89fd1d8cb592756 # v1.13.0
- name: Build genai libs
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
cmake -DCMAKE_BUILD_TYPE=${{ matrix.build-type }} -S ./ -B ./build/
cmake --build ./build/ --config ${{ matrix.build-type }} --target package -j
- name: Build and Install dependencies
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
python -m pip install ./thirdparty/openvino_tokenizers/[transformers] --find-links ${env:OV_INSTALL_DIR}/wheels
python -m pip install -r ./samples/requirements.txt --find-links ${env:OV_INSTALL_DIR}/wheels
optimum-cli export openvino --trust-remote-code --weight-format fp16 --model TinyLlama/TinyLlama-1.1B-Chat-v1.0 TinyLlama-1.1B-Chat-v1.0
optimum-cli export openvino --trust-remote-code --model openai/whisper-tiny whisper-tiny
- name: Install samples
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
cmake --install ./build/ --config ${{ matrix.build-type }} --prefix ${{ env.OV_INSTALL_DIR }}
- name: Build samples (Release)
if: ${{ 'Release' == matrix.build-type }} # build_samples enforces Release build
run: |
& ${{ env.OV_INSTALL_DIR }}\samples\cpp\build_samples.ps1 -i ${{ github.workspace }}/samples_install
- name: Build samples (Debug)
if: ${{ 'Release' != matrix.build-type }}
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
cmake -DCMAKE_BUILD_TYPE= -S ./ov/samples/cpp/ -B "samples build"
cmake --build "samples build" --config ${{ matrix.build-type }} -j
cmake --install "samples build" --config ${{ matrix.build-type }} --component samples_bin --prefix samples_install
- name: Test C++ samples (greedy_causal_lm)
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
Start-Process -FilePath "${{ github.workspace }}/samples_install/samples_bin/greedy_causal_lm.exe" -ArgumentList "TinyLlama-1.1B-Chat-v1.0 ''"
- name: Test C++ samples (whisper_speech_recognition)
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
curl -o how_are_you_doing_today.wav https://storage.openvinotoolkit.org/models_contrib/speech/2021.2/librispeech_s5/how_are_you_doing_today.wav
Start-Process -FilePath "${{ github.workspace }}/samples_install/samples_bin/whisper_speech_recognition.exe" -ArgumentList "whisper-tiny how_are_you_doing_today.wav"
- name: Test python samples (multinomial_causal_lm)
if: ${{ 'Release' == matrix.build-type }} # Python bindings can be built in Release only
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
python ${{ env.OV_INSTALL_DIR }}\samples\python\multinomial_causal_lm\multinomial_causal_lm.py TinyLlama-1.1B-Chat-v1.0 0
- name: Test python samples (whisper_speech_recognition)
if: ${{ 'Release' == matrix.build-type }} # Python bindings can be built in Release only
run: |
. "${{ env.OV_INSTALL_DIR }}/setupvars.ps1"
python ${{ env.OV_INSTALL_DIR }}\samples\python\whisper_speech_recognition\whisper_speech_recognition.py whisper-tiny how_are_you_doing_today.wav
Overall_Status:
name: ci/gha_overall_status_windows
needs: [openvino_download, openvino_build, genai_python_lib, genai_package, genai_python_lib_vlm, genai_python_lib_whisper]
if: ${{ always() }}
runs-on: ubuntu-latest
steps:
- name: Check status of all jobs
if: >-
${{
contains(needs.*.result, 'failure') ||
contains(needs.*.result, 'cancelled')
}}
run: exit 1