Full Model Evaluation #69
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: Full Model Evaluation | |
| on: | |
| workflow_dispatch: | |
| inputs: | |
| purge: | |
| description: 'Whether to purge the deployment after evaluation' | |
| required: true | |
| default: false | |
| type: boolean | |
| permissions: | |
| id-token: write | |
| contents: read | |
| jobs: | |
| deploy-azure: | |
| uses: ./.github/workflows/azure-dev.yml | |
| secrets: inherit | |
| evaluate-models: | |
| needs: deploy-azure | |
| runs-on: ubuntu-latest | |
| strategy: | |
| matrix: | |
| python_version: ["3.11"] | |
| env: | |
| # az/azd credentials | |
| AZURE_CLIENT_ID: ${{ vars.AZURE_CLIENT_ID }} | |
| AZURE_TENANT_ID: ${{ vars.AZURE_TENANT_ID }} | |
| AZURE_SUBSCRIPTION_ID: ${{ vars.AZURE_SUBSCRIPTION_ID }} | |
| AZURE_ENV_NAME: ${{ vars.AZURE_ENV_NAME }} | |
| AZURE_LOCATION: ${{ vars.AZURE_LOCATION }} | |
| steps: | |
| #--------------------------Setup-------------------------------- | |
| - name: Checkout | |
| uses: actions/checkout@v4 | |
| - name: Setup python | |
| uses: actions/setup-python@v5 | |
| with: | |
| python-version: ${{ matrix.python_version }} | |
| architecture: x64 | |
| - name: Install azd | |
| uses: Azure/[email protected] | |
| #--------------------------Azure Login-------------------------------- | |
| - name: Azure login | |
| uses: azure/login@v2 | |
| with: | |
| client-id: ${{ env.AZURE_CLIENT_ID }} | |
| tenant-id: ${{ env.AZURE_TENANT_ID }} | |
| subscription-id: ${{ env.AZURE_SUBSCRIPTION_ID }} | |
| #--------------------------Run Evaluation-------------------------------- | |
| - name: Info debug | |
| run: | | |
| echo "$GITHUB_ENV" | |
| - name: Install dependencies | |
| working-directory: ./evaluation | |
| run: | | |
| python -m pip install --upgrade pip | |
| pip install -r scripts/requirements.txt | |
| pip install pyrit | |
| - name: Run AI Rag Evaluations | |
| working-directory: ./evaluation | |
| run: | | |
| python -m scripts evaluate --config=./config.json --numquestions=10 | |
| echo "EVALUATION_RESULTS=evaluation/results/$(ls ./results/ | grep "experiment" | tail -n 1)" >> $GITHUB_ENV | |
| env: | |
| OPENAI_HOST: ${{ vars.OPENAI_HOST == '' && env.OPENAI_HOST || vars.OPENAI_HOST }} | |
| AZURE_OPENAI_SERVICE: ${{ vars.AZURE_OPENAI_SERVICE == '' && env.AZURE_OPENAI_SERVICE || vars.AZURE_OPENAI_SERVICE }} | |
| AZURE_SEARCH_SERVICE: ${{ vars.AZURE_SEARCH_SERVICE == '' && env.AZURE_SEARCH_SERVICE || vars.AZURE_SEARCH_SERVICE }} | |
| OPENAI_GPT_MODEL: ${{ vars.OPENAI_GPT_MODEL == '' && env.OPENAI_GPT_MODEL || vars.OPENAI_GPT_MODEL }} | |
| AZURE_OPENAI_EVAL_DEPLOYMENT: ${{ vars.AZURE_OPENAI_EVAL_DEPLOYMENT == '' && env.AZURE_OPENAI_EVAL_DEPLOYMENT || vars.AZURE_OPENAI_EVAL_DEPLOYMENT }} | |
| AZURE_SEARCH_INDEX: ${{ vars.AZURE_SEARCH_INDEX == '' && env.AZURE_SEARCH_INDEX || vars.AZURE_SEARCH_INDEX }} | |
| AZURE_PRINCIPAL_ID: ${{ secrets.AZURE_PRINCIPAL_ID == '' && env.AZURE_PRINCIPAL_ID || secrets.AZURE_PRINCIPAL_ID }} | |
| - name: Run Red Teaming Evaluations | |
| working-directory: ./evaluation | |
| run: | | |
| python -m scripts red-teaming | |
| env: | |
| OPENAI_HOST: ${{ vars.OPENAI_HOST == '' && env.OPENAI_HOST || vars.OPENAI_HOST }} | |
| AZURE_OPENAI_SERVICE: ${{ vars.AZURE_OPENAI_SERVICE == '' && env.AZURE_OPENAI_SERVICE || vars.AZURE_OPENAI_SERVICE }} | |
| AZURE_SEARCH_SERVICE: ${{ vars.AZURE_SEARCH_SERVICE == '' && env.AZURE_SEARCH_SERVICE || vars.AZURE_SEARCH_SERVICE }} | |
| OPENAI_GPT_MODEL: ${{ vars.OPENAI_GPT_MODEL == '' && env.OPENAI_GPT_MODEL || vars.OPENAI_GPT_MODEL }} | |
| AZURE_SEARCH_INDEX: ${{ vars.AZURE_SEARCH_INDEX == '' && env.AZURE_SEARCH_INDEX || vars.AZURE_SEARCH_INDEX }} | |
| AZURE_PRINCIPAL_ID: ${{ secrets.AZURE_PRINCIPAL_ID == '' && env.AZURE_PRINCIPAL_ID || secrets.AZURE_PRINCIPAL_ID }} | |
| AZURE_OPENAI_CHAT_ENDPOINT: ${{ vars.AZURE_OPENAI_CHAT_ENDPOINT == '' && env.AZURE_OPENAI_CHAT_ENDPOINT || vars.AZURE_OPENAI_CHAT_ENDPOINT }} | |
| AZURE_OPENAI_EVAL_ENDPOINT: ${{ vars.AZURE_OPENAI_EVAL_ENDPOINT == '' && env.AZURE_OPENAI_CHAT_ENDPOINT || vars.AZURE_OPENAI_EVAL_ENDPOINT }} | |
| AZURE_OPENAI_CHAT_DEPLOYMENT: ${{ vars.AZURE_OPENAI_CHAT_DEPLOYMENT == '' && env.AZURE_OPENAI_CHAT_DEPLOYMENT || vars.AZURE_OPENAI_CHAT_DEPLOYMENT }} | |
| AZURE_OPENAI_EVAL_DEPLOYMENT: ${{ vars.AZURE_OPENAI_EVAL_DEPLOYMENT == '' && env.AZURE_OPENAI_CHAT_DEPLOYMENT || vars.AZURE_OPENAI_EVAL_DEPLOYMENT }} | |
| - name: Dump results | |
| uses: actions/upload-artifact@v4 | |
| with: | |
| name: evaluation-results | |
| path: | | |
| ${{ env.EVALUATION_RESULTS }}/summary.json | |
| ${{ env.EVALUATION_RESULTS }}/eval_results.jsonl | |
| ${{ env.EVALUATION_RESULTS }}/config.json | |
| ${{ env.EVALUATION_RESULTS }}/evaluate_parameters.json | |
| evaluation/eval.png | |
| evaluation/mean_score.png | |
| evaluation/passing_rate.png | |
| #--------------------------Cleanup-------------------------------- | |
| - name: Cleanup | |
| if: ${{ inputs.purge == true }} | |
| run: azd down --force --purge |