LMI Dist library Unit/Integration tests #20
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: LMI Dist library Unit/Integration tests | |
| on: | |
| workflow_dispatch: | |
| inputs: | |
| djl-version: | |
| description: 'The released version of DJL' | |
| required: false | |
| default: '' | |
| schedule: | |
| - cron: '0 4 * * *' | |
| jobs: | |
| create-runners: | |
| runs-on: [self-hosted, scheduler] | |
| steps: | |
| - name: Create new G5 instance | |
| id: create_gpu | |
| run: | | |
| cd /home/ubuntu/djl_benchmark_script/scripts | |
| token=$( curl -X POST -H "Authorization: token ${{ secrets.ACTION_RUNNER_PERSONAL_TOKEN }}" \ | |
| https://api.github.com/repos/deepjavalibrary/djl-serving/actions/runners/registration-token \ | |
| --fail \ | |
| | jq '.token' | tr -d '"' ) | |
| ./start_instance.sh action_g5 $token djl-serving | |
| outputs: | |
| gpu_instance_id: ${{ steps.create_gpu.outputs.action_g5_instance_id }} | |
| integration-test: | |
| runs-on: [ self-hosted, g5 ] | |
| timeout-minutes: 60 | |
| needs: create-runners | |
| steps: | |
| - uses: actions/checkout@v3 | |
| - name: Build container name | |
| run: ./serving/docker/scripts/docker_name_builder.sh deepspeed ${{ github.event.inputs.djl-version }} | |
| - name: Download docker | |
| working-directory: tests/integration | |
| run: | | |
| docker pull deepjavalibrary/djl-serving:$DJLSERVING_DOCKER_TAG | |
| # - name: Test gpt-neox-20b | |
| # working-directory: tests/integration | |
| # run: | | |
| # rm -rf models | |
| # python3 llm/prepare.py lmi_dist gpt-neox-20b | |
| # ./launch_container.sh deepjavalibrary/djl-serving:deepspeed-nightly $PWD/models deepspeed \ | |
| # serve -m test=file:/opt/ml/model/test/ | |
| # python3 llm/client.py lmi_dist gpt-neox-20b | |
| # docker rm -f $(docker ps -aq) | |
| # - name: Test falcon-7b | |
| # working-directory: tests/integration | |
| # run: | | |
| # rm -rf models | |
| # python3 llm/prepare.py lmi_dist falcon-7b | |
| # ./launch_container.sh deepjavalibrary/djl-serving:deepspeed-nightly $PWD/models deepspeed \ | |
| # serve -m test=file:/opt/ml/model/test/ | |
| # python3 llm/client.py lmi_dist falcon-7b | |
| # docker rm -f $(docker ps -aq) | |
| # - name: Test open-llama-7b | |
| # working-directory: tests/integration | |
| # run: | | |
| # rm -rf models | |
| # python3 llm/prepare.py lmi_dist open-llama-7b | |
| # ./launch_container.sh deepjavalibrary/djl-serving:deepspeed-nightly $PWD/models deepspeed \ | |
| # serve -m test=file:/opt/ml/model/test/ | |
| # python3 llm/client.py lmi_dist open-llama-7b | |
| # docker rm -f $(docker ps -aq) | |
| - name: Test flan-t5-base | |
| working-directory: tests/integration | |
| run: | | |
| rm -rf models | |
| python3 llm/prepare.py lmi_dist flan-t5-base | |
| ./launch_container.sh deepjavalibrary/djl-serving:deepspeed-nightly $PWD/models deepspeed \ | |
| serve -m test=file:/opt/ml/model/test/ | |
| python3 llm/client.py lmi_dist flan-t5-base | |
| docker rm -f $(docker ps -aq) | |
| - name: On fail step | |
| if: ${{ failure() }} | |
| working-directory: tests/integration | |
| run: | | |
| docker rm -f $(docker ps -aq) || true | |
| cat logs/serving.log | |
| - name: Upload test logs | |
| uses: actions/upload-artifact@v3 | |
| with: | |
| name: lmi-dist-logs | |
| path: tests/integration/logs/ | |
| stop-runners: | |
| if: always() | |
| runs-on: [ self-hosted, scheduler ] | |
| needs: [ create-runners, integration-test ] | |
| steps: | |
| - name: Stop all instances | |
| run: | | |
| cd /home/ubuntu/djl_benchmark_script/scripts | |
| instance_id=${{ needs.create-runners.outputs.gpu_instance_id }} | |
| ./stop_instance.sh $instance_id |