|
| 1 | +name: amlarc-examples-test |
| 2 | +on: |
| 3 | + schedule: |
| 4 | + - cron: "0 23 * * *" |
| 5 | + workflow_dispatch: |
| 6 | + inputs: |
| 7 | + AMLARC_TEST_REGION: |
| 8 | + description: 'Resource Region' |
| 9 | + required: false |
| 10 | + default: 'eastus' |
| 11 | + AMLARC_RELEASE_TRAIN: |
| 12 | + description: 'Release version: experimental, staging or stable' |
| 13 | + required: false |
| 14 | + default: 'staging' |
| 15 | + CLEANUP_CLUSTER: |
| 16 | + description: 'Whether to delete cluster: true or false' |
| 17 | + required: false |
| 18 | + default: 'false' |
| 19 | + CLEANUP_WORKSPACE: |
| 20 | + description: 'Whether to delete workspace: true or false' |
| 21 | + required: false |
| 22 | + default: 'false' |
| 23 | + FILE_TICKET: |
| 24 | + description: 'Whether to file icm ticket: true or false' |
| 25 | + required: false |
| 26 | + default: 'false' |
| 27 | +jobs: |
| 28 | + test: |
| 29 | + runs-on: ubuntu-latest |
| 30 | + env: |
| 31 | + SUBSCRIPTION: 6560575d-fa06-4e7d-95fb-f962e74efd7a |
| 32 | + RESOURCE_GROUP: azureml-examples-rg |
| 33 | + KEY_VAULT_NAME: amlarcgithubworkflowkv |
| 34 | + KEY_VAULT_SUB: dd94e4df-b58f-4e81-9c5e-bd72a4b2ef9e |
| 35 | + KEY_VAULT_RG: amlarc-github-workflow |
| 36 | + LOCATION: ${{ github.event.inputs.AMLARC_TEST_REGION }} |
| 37 | + RELEASE_TRAIN: ${{ github.event.inputs.AMLARC_RELEASE_TRAIN }} |
| 38 | + VM_SKU: STANDARD_NC12 |
| 39 | + WORKSPACE: amlarc-githubtest-ws |
| 40 | + |
| 41 | + MIN_SUCCESS_NUM: 3 |
| 42 | + SEVERITY: 3 |
| 43 | + |
| 44 | + GITHUB_REPO: https://github.com/Azure/azureml-examples |
| 45 | + WORKFLOW_URL: https://github.com/Azure/azureml-examples/actions/workflows/amlarc-cli-train-test.yml |
| 46 | + TSG_ID: https://microsoft.sharepoint.com/teams/Vienna/_layouts/OneNote.aspx?id=%2Fteams%2FVienna%2FSiteAssets%2FVienna%20Notebook&wd=target%28AMLArcTSG.one%7C69E8BDB1-5734-4B07-967A-5A50A91CF040%2FTroubleshooting%20github%20workflow%7C53A1A232-F9F0-4192-B7D2-0474848EBB18%2F%29 |
| 47 | + |
| 48 | + steps: |
| 49 | + - name: check out repo |
| 50 | + uses: actions/checkout@v2 |
| 51 | + - name: setup python |
| 52 | + uses: actions/setup-python@v2 |
| 53 | + with: |
| 54 | + python-version: "3.8" |
| 55 | + - name: install tools |
| 56 | + run: bash .github/amlarc-tool.sh install_tools |
| 57 | + timeout-minutes: 30 |
| 58 | + - name: azure login |
| 59 | + uses: azure/login@v1 |
| 60 | + with: |
| 61 | + creds: ${{secrets.AZ_AE_CREDS}} |
| 62 | + timeout-minutes: 30 |
| 63 | + # provision resources |
| 64 | + - name: setup_aks |
| 65 | + run: bash .github/amlarc-tool.sh setup_aks |
| 66 | + timeout-minutes: 30 |
| 67 | + - name: connect_arc |
| 68 | + run: bash .github/amlarc-tool.sh connect_arc |
| 69 | + timeout-minutes: 30 |
| 70 | + - name: install_extension |
| 71 | + run: | |
| 72 | + export EXTENSION_SETTINGS="enableTraining=True enableInference=True allowInsecureConnections=True installNvidiaDevicePlugin=True installDcgmExporter=True" |
| 73 | + bash .github/amlarc-tool.sh install_extension |
| 74 | + timeout-minutes: 30 |
| 75 | + - name: setup_workspace |
| 76 | + run: bash .github/amlarc-tool.sh setup_workspace |
| 77 | + timeout-minutes: 30 |
| 78 | + - name: setup_compute |
| 79 | + run: | |
| 80 | + bash .github/amlarc-tool.sh setup_compute |
| 81 | + COMPUTE="gpu-cluster" bash .github/amlarc-tool.sh setup_compute |
| 82 | + COMPUTE="cpu-cluster" bash .github/amlarc-tool.sh setup_compute |
| 83 | + timeout-minutes: 30 |
| 84 | + - name: setup_instance_type |
| 85 | + run: | |
| 86 | + bash .github/amlarc-tool.sh setup_instance_type defaultinstancetype 4 32Gi |
| 87 | + bash .github/amlarc-tool.sh setup_instance_type cpu 4 32Gi |
| 88 | + bash .github/amlarc-tool.sh setup_instance_type gpu 4 32Gi 2 |
| 89 | + bash -c "echo Waiting 30 seconds for instance type to take effect ...; sleep 30" |
| 90 | + timeout-minutes: 30 |
| 91 | + |
| 92 | + ## setup repo |
| 93 | + - name: create dataset |
| 94 | + run: | |
| 95 | + az account set --subscription $SUBSCRIPTION |
| 96 | + az configure --defaults group=$RESOURCE_GROUP workspace=$WORKSPACE location=$LOCATION |
| 97 | + bash create-datasets.sh |
| 98 | + working-directory: setup-repo |
| 99 | + timeout-minutes: 30 |
| 100 | + |
| 101 | + ## run training cli test cases |
| 102 | + ## single step |
| 103 | + # dask |
| 104 | + - name: run cli/jobs/single-step/dask/nyctaxi/job.yml |
| 105 | + run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/dask/nyctaxi/job.yml compute=azureml:githubtest resources.instance_type=cpu environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true |
| 106 | + continue-on-error: true |
| 107 | + timeout-minutes: 40 |
| 108 | + # java |
| 109 | + # - name: run cli/jobs/single-step/java/iris/job-template.yml |
| 110 | + # run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/java/iris/job-template.yml |
| 111 | + # continue-on-error: true |
| 112 | + # timeout-minutes: 40 |
| 113 | + # julia |
| 114 | + - name: run cli/jobs/single-step/julia/iris/job.yml |
| 115 | + run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/julia/iris/job.yml |
| 116 | + continue-on-error: true |
| 117 | + timeout-minutes: 40 |
| 118 | + # lightgbm |
| 119 | + - name: run cli/jobs/single-step/lightgbm/iris/job.yml |
| 120 | + run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/lightgbm/iris/job.yml |
| 121 | + continue-on-error: true |
| 122 | + timeout-minutes: 40 |
| 123 | + # pytorch |
| 124 | + - name: run cli/jobs/single-step/pytorch/cifar-distributed/job.yml |
| 125 | + run: INSTANCE_TYPE_NAME=gpu bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/pytorch/cifar-distributed/job.yml compute=azureml:githubtest resources.instance_type=gpu environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true |
| 126 | + continue-on-error: true |
| 127 | + timeout-minutes: 40 |
| 128 | + - name: run cli/jobs/single-step/pytorch/iris/job.yml |
| 129 | + run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/pytorch/iris/job.yml |
| 130 | + continue-on-error: true |
| 131 | + timeout-minutes: 40 |
| 132 | + - name: run cli/jobs/single-step/pytorch/word-language-model/job.yml |
| 133 | + run: INSTANCE_TYPE_NAME=gpu bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/pytorch/word-language-model/job.yml |
| 134 | + continue-on-error: true |
| 135 | + timeout-minutes: 40 |
| 136 | + # r |
| 137 | + - name: run cli/jobs/single-step/r/accidents/job.yml |
| 138 | + run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/r/accidents/job.yml compute=azureml:githubtest resources.instance_type=cpu environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true |
| 139 | + continue-on-error: true |
| 140 | + timeout-minutes: 40 |
| 141 | + - name: run cli/jobs/single-step/r/iris/job.yml |
| 142 | + run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/r/iris/job.yml compute=azureml:githubtest resources.instance_type=cpu environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true |
| 143 | + continue-on-error: true |
| 144 | + timeout-minutes: 40 |
| 145 | + # scikit-learn |
| 146 | + - name: run cli/jobs/single-step/scikit-learn/diabetes/job.yml |
| 147 | + run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/scikit-learn/diabetes/job.yml |
| 148 | + continue-on-error: true |
| 149 | + timeout-minutes: 40 |
| 150 | + - name: run cli/jobs/single-step/scikit-learn/iris-notebook/job.yml |
| 151 | + run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/scikit-learn/iris-notebook/job.yml compute=azureml:githubtest resources.instance_type=cpu environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true |
| 152 | + continue-on-error: true |
| 153 | + timeout-minutes: 40 |
| 154 | + - name: run cli/jobs/single-step/scikit-learn/iris/job.yml |
| 155 | + run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/scikit-learn/iris/job.yml |
| 156 | + continue-on-error: true |
| 157 | + timeout-minutes: 40 |
| 158 | + # spark |
| 159 | + #- name: run cli/jobs/single-step/spark/nyctaxi/job.yml |
| 160 | + # run: bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/spark/nyctaxi/job.yml compute=azureml:githubtest resources.instance_type=cpu environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true |
| 161 | + # continue-on-error: true |
| 162 | + # timeout-minutes: 40 |
| 163 | + # tensorflow |
| 164 | + - name: run cli/jobs/single-step/tensorflow/mnist/job.yml |
| 165 | + run: INSTANCE_TYPE_NAME=gpu bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/tensorflow/mnist/job.yml |
| 166 | + continue-on-error: true |
| 167 | + timeout-minutes: 40 |
| 168 | + - name: run cli/jobs/single-step/tensorflow/mnist-distributed/job.yml |
| 169 | + run: INSTANCE_TYPE_NAME=gpu bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/tensorflow/mnist-distributed/job.yml |
| 170 | + continue-on-error: true |
| 171 | + timeout-minutes: 40 |
| 172 | + - name: run cli/jobs/single-step/tensorflow/mnist-distributed-horovod/job.yml |
| 173 | + run: INSTANCE_TYPE_NAME=gpu bash .github/amlarc-tool.sh run_cli_job cli/jobs/single-step/tensorflow/mnist-distributed-horovod/job.yml |
| 174 | + continue-on-error: true |
| 175 | + timeout-minutes: 40 |
| 176 | + |
| 177 | + ## pipeline |
| 178 | + # cifar-10 |
| 179 | + - name: run cli/jobs/pipelines/cifar-10/job.yml |
| 180 | + run: | |
| 181 | + bash .github/amlarc-tool.sh run_cli_job cli/jobs/pipelines/cifar-10/pipeline.yml \ |
| 182 | + jobs.get-data.environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true \ |
| 183 | + jobs.train-model.environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true \ |
| 184 | + jobs.eval-model.environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true |
| 185 | + continue-on-error: true |
| 186 | + timeout-minutes: 40 |
| 187 | + # nyc-taxi |
| 188 | + - name: run cli/jobs/pipelines/nyc-taxi/job.yml |
| 189 | + run: | |
| 190 | + bash .github/amlarc-tool.sh run_cli_job cli/jobs/pipelines/nyc-taxi/pipeline.yml \ |
| 191 | + jobs.prep-job.environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true \ |
| 192 | + jobs.transform-job.environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true \ |
| 193 | + jobs.train-job.environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true \ |
| 194 | + jobs.predict-job.environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true \ |
| 195 | + jobs.score-job.environment_variables.AZUREML_COMPUTE_USE_COMMON_RUNTIME=true |
| 196 | + continue-on-error: true |
| 197 | + timeout-minutes: 40 |
| 198 | + |
| 199 | + # check result |
| 200 | + - name: count_result |
| 201 | + if: ${{ always() }} |
| 202 | + run: | |
| 203 | + bash .github/amlarc-tool.sh count_result |
| 204 | + timeout-minutes: 30 |
| 205 | + - name: file_icm |
| 206 | + if: ${{ always() }} |
| 207 | + run: | |
| 208 | + bash .github/amlarc-tool.sh count_result || ret=$? |
| 209 | + if [ "$ret" != "0" ] && [ "$ret" != "" ] && [ "${{ github.event.inputs.FILE_TICKET }}" != "false" ] ; then |
| 210 | + # download certificates |
| 211 | + export ICM_HOST_NAME=ICM-HOST-AML-EXAMPLES |
| 212 | + export ICM_CONNECTOR_ID_NAME=ICM-CONNECTOR-ID-AML-EXAMPLES |
| 213 | + export ICM_ROUTING_ID_NAME=ICM-ROUTING-ID-AML-EXAMPLES |
| 214 | + bash .github/amlarc-tool.sh download_icm_cert |
| 215 | + export ICM_HOST=$(cat icm_host) |
| 216 | + export CONNECTOR_ID=$(cat icm_connector_id) |
| 217 | + export ROUTING_ID=$(cat icm_routing_id) |
| 218 | + export SUMMARY=$(bash .github/amlarc-tool.sh gen_summary_for_github_test) |
| 219 | + bash .github/amlarc-tool.sh file_icm |
| 220 | + fi |
| 221 | + timeout-minutes: 30 |
| 222 | + |
| 223 | + - name: clean up resources |
| 224 | + if: ${{ always() }} |
| 225 | + run: | |
| 226 | + set +e |
| 227 | + bash .github/amlarc-tool.sh delete_endpoints |
| 228 | + bash .github/amlarc-tool.sh delete_compute |
| 229 | + bash .github/amlarc-tool.sh delete_extension |
| 230 | + if [ "${{ github.event.inputs.CLEANUP_CLUSTER }}" == "true" ] ; then |
| 231 | + bash .github/amlarc-tool.sh delete_arc |
| 232 | + bash .github/amlarc-tool.sh delete_aks |
| 233 | + fi |
| 234 | + if [ "${{ github.event.inputs.CLEANUP_WORKSPACE }}" == "true" ] ; then |
| 235 | + bash .github/amlarc-tool.sh delete_workspace |
| 236 | + fi |
| 237 | + timeout-minutes: 30 |
0 commit comments