Performance Benchmark #58
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: Performance Benchmark | |
| on: | |
| workflow_dispatch: # run on request (no need for PR) | |
| inputs: | |
| model-category: | |
| type: choice | |
| description: Model category to run benchmark | |
| options: | |
| - speed | |
| - balance | |
| - accuracy | |
| - default # speed, balance, accuracy models only | |
| - all # default + other models | |
| default: all | |
| data-group: | |
| type: choice | |
| description: Data group to run benchmark | |
| options: | |
| - small | |
| - medium | |
| - large | |
| - all | |
| default: all | |
| num-repeat: | |
| description: Overrides default per-data-group number of repeat setting | |
| default: 0 | |
| num-epoch: | |
| description: Overrides default per-model number of epoch setting | |
| default: 0 | |
| eval-upto: | |
| type: choice | |
| description: The last operation to evaluate. 'optimize' means all. | |
| options: | |
| - train | |
| - export | |
| - optimize | |
| default: optimize | |
| pytest-args: | |
| type: string | |
| description: | | |
| Additional perf-benchmark pytest arguments. | |
| "-k detection" -> detection task only | |
| "--dry-run" -> print command w/o execution. | |
| data-root: | |
| type: string | |
| description: Root directory containing validation data in CI server. | |
| default: "/home/validation/data/v2/" | |
| otx-ref: | |
| type: string | |
| description: | | |
| Target OTX ref (tag / branch name / commit hash) on main repo to test. Defaults to the current branch. | |
| `pip install otx[full]@https://github.com/openvinotoolkit/training_extensions.git@{otx_ref}` will be executed before run, | |
| and reverted after run. Works only for v2.x assuming CLI compatibility. | |
| default: __CURRENT_BRANCH_COMMIT__ | |
| workflow_call: | |
| inputs: | |
| model-category: | |
| type: string | |
| description: Model category to run benchmark [speed, balance, accuracy, default, all] | |
| default: default | |
| data-group: | |
| type: string | |
| description: Data group to run benchmark [small, medium, large, all] | |
| default: all | |
| num-repeat: | |
| type: number | |
| description: Overrides default per-data-group number of repeat setting | |
| default: 0 | |
| num-epoch: | |
| type: number | |
| description: Overrides default per-model number of epoch setting | |
| default: 0 | |
| eval-upto: | |
| type: string | |
| description: The last operation to evaluate. 'optimize' means all. [train, export, optimize] | |
| default: optimize | |
| pytest-args: | |
| type: string | |
| description: | | |
| Additional perf-benchmark pytest arguments. | |
| "-k detection" -> detection task only | |
| "--dry-run" -> print command w/o execution. | |
| data-root: | |
| type: string | |
| description: Root directory containing validation data in CI server. | |
| default: "/home/validation/data/v2/" | |
| otx-ref: | |
| type: string | |
| description: | | |
| Target OTX ref (tag / branch name / commit hash) on main repo to test. Defaults to the current branch. | |
| `pip install otx[full]@https://github.com/openvinotoolkit/training_extensions.git@{otx_ref}` will be executed before run, | |
| and reverted after run. Works only for v2.x assuming CLI compatibility. | |
| default: __CURRENT_BRANCH_COMMIT__ | |
| # Declare default permissions as read only. | |
| permissions: read-all | |
| jobs: | |
| Perf-Benchmark-Summary: | |
| if: ${{ always() }} | |
| runs-on: ubuntu-latest | |
| steps: | |
| - name: Download benchmark results | |
| uses: actions/download-artifact@v4 | |
| with: | |
| run-id: 8699060888 | |
| github-token: ${{ secrets.GITHUB_TOKEN }} | |
| path: tests/perf/history/latest | |
| - name: Checkout repository | |
| uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1 | |
| - name: Install Python | |
| uses: actions/setup-python@0a5c61591373683505ea898e09a3ea4f39ef2b9c # v5.0.0 | |
| with: | |
| python-version: "3.10" | |
| - name: Install dependencies | |
| run: python -m pip install --upgrade pip pandas matplotlib nbconvert ipython ipykernel openpyxl | |
| - name: Summarize benchamrk results | |
| run: | | |
| python tests/perf/history/summary.py tests/perf/history ./perf-benchmark-summary --pattern "*raw*.csv" --normalize | |
| jupyter nbconvert --execute --to html --no-input tests/perf/history/summary.ipynb --output-dir ./perf-benchmark-summary --output perf-benchmark-summary | |
| - name: Upload benchmark summary | |
| uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3 # v4.3.1 | |
| with: | |
| name: perf-benchmark-summary | |
| path: perf-benchmark-summary | |
| # Use always() to always run this step to publish test results when there are test failures | |
| if: ${{ always() }} |