monthly-tests #3
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: monthly-tests | |
| on: | |
| workflow_dispatch: | |
| schedule: | |
| - cron: '32 17 9 * *' | |
| env: | |
| WORKSPACE_PREFIX: $(echo $GITHUB_WORKSPACE |cut -d '/' -f 1-4) | |
| SLURM_PARTITION: llm_s | |
| jobs: | |
| training_tp_norm_layer_msp: | |
| runs-on: [t_cluster] | |
| steps: | |
| - name: mask env | |
| run: | | |
| echo "::add-mask::${{env.WORKSPACE_PREFIX}}" | |
| echo "::add-mask::$path_prefix" | |
| - uses: actions/checkout@v3 | |
| with: | |
| ref: ${{ github.event_name == 'schedule' && 'develop' || github.event_name == 'workflow_dispatch' && '' }} | |
| - name: training_tp_norm_layer_msp | |
| run: | | |
| source activate ${evo_env_torch21_flash2} | |
| jobname=${GITHUB_RUN_ID}-${GITHUB_JOB}-${GITHUB_RUN_ATTEMPT} | |
| srun -p ${SLURM_PARTITION} --exclusive --kill-on-bad-exit=1 --job-name=$jobname -N 1 -n 1 --gres=gpu:8 pytest -s -v --color=yes -m "check_norm_msp" ./tests/test_training/test_norm_weight.py | |
| exit_code=$? | |
| sh ./ci_scripts/common/check_slurm_cancled.sh $exit_code $jobname | |
| training_tp_norm_layer_fsp: | |
| runs-on: [t_cluster] | |
| steps: | |
| - name: mask env | |
| run: | | |
| echo "::add-mask::${{env.WORKSPACE_PREFIX}}" | |
| echo "::add-mask::$path_prefix" | |
| - uses: actions/checkout@v3 | |
| with: | |
| ref: ${{ github.event_name == 'schedule' && 'develop' || github.event_name == 'workflow_dispatch' && '' }} | |
| - name: training_tp_norm_layer_fsp | |
| run: | | |
| source activate ${evo_env_torch21_flash2} | |
| jobname=${GITHUB_RUN_ID}-${GITHUB_JOB}-${GITHUB_RUN_ATTEMPT} | |
| srun -p ${SLURM_PARTITION} --exclusive --kill-on-bad-exit=1 --job-name=$jobname -N 1 -n 1 --gres=gpu:8 pytest -s -v --color=yes -m "check_norm_fsp" ./tests/test_training/test_norm_weight.py | |
| exit_code=$? | |
| sh ./ci_scripts/common/check_slurm_cancled.sh $exit_code $jobname | |
| training_tp_norm_layer_isp: | |
| runs-on: [t_cluster] | |
| steps: | |
| - name: mask env | |
| run: | | |
| echo "::add-mask::${{env.WORKSPACE_PREFIX}}" | |
| echo "::add-mask::$path_prefix" | |
| - uses: actions/checkout@v3 | |
| with: | |
| ref: ${{ github.event_name == 'schedule' && 'develop' || github.event_name == 'workflow_dispatch' && '' }} | |
| - name: training_tp_norm_layer_isp | |
| run: | | |
| source activate ${evo_env_torch21_flash2} | |
| jobname=${GITHUB_RUN_ID}-${GITHUB_JOB}-${GITHUB_RUN_ATTEMPT} | |
| srun -p ${SLURM_PARTITION} --exclusive --kill-on-bad-exit=1 --job-name=$jobname -N 1 -n 1 --gres=gpu:8 pytest -s -v --color=yes -m "check_norm_isp" ./tests/test_training/test_norm_weight.py | |
| exit_code=$? | |
| sh ./ci_scripts/common/check_slurm_cancled.sh $exit_code $jobname | |
| notify_to_feishu: | |
| if: ${{ always() && !cancelled() && contains(needs.*.result, 'failure') && (github.ref_name == 'develop' || github.ref_name == 'main') }} | |
| needs: [ | |
| training_tp_norm_layer_msp, | |
| training_tp_norm_layer_fsp, | |
| training_tp_norm_layer_isp | |
| ] | |
| runs-on: [t_cluster] | |
| steps: | |
| - name: mask env | |
| run: | | |
| echo "::add-mask::${{env.WORKSPACE_PREFIX}}" | |
| echo "::add-mask::$path_prefix" | |
| - name: notify | |
| run: | | |
| curl -X POST -H "Content-Type: application/json" -d '{"msg_type":"post","content":{"post":{"zh_cn":{"title":"'${{ github.repository }}' GitHubAction Failed","content":[[{"tag":"text","text":""},{"tag":"a","text":"Please click here for details ","href":"https://github.com/'${{ github.repository }}'/actions/runs/'${GITHUB_RUN_ID}'"},{"tag":"at","user_id":"'${{ secrets.USER_ID }}'"}]]}}}}' ${{ secrets.WEBHOOK_URL }} |