MLPerf inference GPT-J #88
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
| # This workflow will install Python dependencies, run tests and lint with a variety of Python versions | |
| # For more information see: https://help.github.com/actions/language-and-framework-guides/using-python-with-github-actions | |
| name: MLPerf inference GPT-J | |
| on: | |
| schedule: | |
| - cron: "15 19 * * *" | |
| jobs: | |
| build: | |
| if: github.repository_owner == 'gateoverflow' | |
| runs-on: [ self-hosted, linux, x64, GO-spr ] | |
| strategy: | |
| fail-fast: false | |
| matrix: | |
| python-version: [ "3.12" ] | |
| backend: [ "pytorch" ] | |
| precision: [ "float16" ] | |
| steps: | |
| - name: Test MLPerf Inference GPTJ | |
| run: | | |
| source gh_action/bin/deactivate || python3 -m venv gh_action | |
| source gh_action/bin/activate | |
| export CM_REPOS=$HOME/GH_CM | |
| python3 -m pip install cm4mlops | |
| cm pull repo | |
| cm run script --tags=run-mlperf,inference,_submission,_short --submitter="MLCommons" --docker --model=gptj-99 --backend=${{ matrix.backend }} --device=cuda --scenario=Offline --test_query_count=1 --precision=${{ matrix.precision }} --target_qps=1 --quiet --docker_it=no --docker_cm_repo=gateoverflow@cm4mlops --adr.compiler.tags=gcc --beam_size=1 --hw_name=gh_action --docker_dt=yes --results_dir=$HOME/gh_action_results --submission_dir=$HOME/gh_action_submissions --get_platform_details=yes --implementation=reference --clean | |
| cm run script --tags=push,github,mlperf,inference,submission --repo_url=https://github.com/gateoverflow/mlperf_inference_test_submissions_v5.0 --repo_branch=main --commit_message="Results from self hosted Github actions - NVIDIARTX4090" --quiet --submission_dir=$HOME/gh_action_submissions | |