[Feature] Add gpqa prompt from simple_evals, openai #353
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: pr_run_test | |
| on: | |
| pull_request: | |
| paths-ignore: | |
| - 'README.md' | |
| - 'README_zh-CN.md' | |
| - 'docs/**' | |
| - 'configs/**' | |
| - 'tools/**' | |
| workflow_dispatch: | |
| schedule: | |
| - cron: '56 22 * * *' | |
| concurrency: | |
| group: ${{ github.workflow }}-${{ github.ref }} | |
| cancel-in-progress: true | |
| env: | |
| CONDA_ENV: opencompass_base | |
| USERSPACE_PREFIX: /cpfs01/user/qa-llm-cicd | |
| HF_CACHE_PATH: /cpfs01/shared/public/public_hdd/llmeval/model_weights/hf_hub | |
| jobs: | |
| pr_run_test: | |
| runs-on: self-hosted | |
| environment: 'prod' | |
| timeout-minutes: 30 | |
| steps: | |
| - name: Clone repository | |
| uses: actions/checkout@v2 | |
| - name: Prepare - Install opencompass | |
| run: | | |
| eval "$(conda shell.bash hook)" | |
| conda activate ${{env.CONDA_ENV}} | |
| python3 -m pip uninstall opencompass -y | |
| python3 -m pip install -e . --cache-dir ${{env.USERSPACE_PREFIX}}/.cache/pip | |
| conda info --envs | |
| - name: Prepare - prepare data and hf model | |
| run: | | |
| cp -r ${{env.USERSPACE_PREFIX}}/data . | |
| rm -rf ~/.cache/huggingface/hub -f && mkdir ~/.cache -p && mkdir ~/.cache/huggingface -p | |
| ln -s ${{env.HF_CACHE_PATH}} ~/.cache/huggingface/hub | |
| export HF_DATASETS_OFFLINE=1; export TRANSFORMERS_OFFLINE=1; | |
| - name: Run test | |
| run: | | |
| eval "$(conda shell.bash hook)" | |
| conda activate ${{env.CONDA_ENV}} | |
| conda info --envs | |
| rm -rf regression_result | |
| python3 run.py --models hf_internlm_chat_7b --datasets siqa_gen --work-dir regression_result --debug | |
| - name: Get result | |
| run: | | |
| score=$(sed -n '$p' regression_result/*/summary/*.csv | awk -F ',' '{print $NF}') | |
| if (( ${score%.*} >= 70 && ${score%.*} <= 75 )); then | |
| echo "score is $score between 70 and 75" | |
| else | |
| echo "score is $score not between 70 and 75" | |
| exit 1 | |
| fi | |
| rm -rf regression_result | |
| - name: Uninstall opencompass | |
| if: always() | |
| run: | | |
| eval "$(conda shell.bash hook)" | |
| conda activate ${{env.CONDA_ENV}} | |
| python3 -m pip uninstall opencompass -y | |
| conda info --envs | |
| notify_to_feishu: | |
| if: ${{ always() && !cancelled() && contains(needs.*.result, 'failure') && (github.ref_name == 'develop' || github.ref_name == 'main') }} | |
| needs: [pr_run_test] | |
| environment: 'prod' | |
| timeout-minutes: 5 | |
| runs-on: self-hosted | |
| steps: | |
| - name: notify | |
| run: | | |
| curl -X POST -H "Content-Type: application/json" -d '{"msg_type":"post","content":{"post":{"zh_cn":{"title":"Opencompass- pr test failed","content":[[{"tag":"text","text":"branch: ${{github.ref_name}}, run action: ${{github.workflow}} failed. "},{"tag":"a","text":"Please click here for details ","href":"https://github.com/'${{ github.repository }}'/actions/runs/'${GITHUB_RUN_ID}'"},{"tag":"at","user_id":"'${{ secrets.USER_ID }}'"}]]}}}}' ${{ secrets.WEBHOOK_URL }} |