E2E (NVIDIA L40S x4) #28
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
| # SPDX-License-Identifier: Apache-2.0 | |
| name: E2E (NVIDIA L40S x4) | |
| on: | |
| schedule: | |
| - cron: '0 16 * * *' # Runs at 4PM UTC every day | |
| workflow_dispatch: | |
| inputs: | |
| pr_or_branch: | |
| description: 'pull request number or branch name' | |
| required: true | |
| default: 'main' | |
| jobs: | |
| start-large-ec2-runner: | |
| runs-on: ubuntu-latest | |
| outputs: | |
| label: ${{ steps.start-ec2-runner.outputs.label }} | |
| ec2-instance-id: ${{ steps.start-ec2-runner.outputs.ec2-instance-id }} | |
| steps: | |
| - name: "Harden Runner" | |
| # v2.10.1 | |
| uses: step-security/harden-runner@91182cccc01eb5e619899d80e4e971d6181294a7 | |
| with: | |
| egress-policy: audit | |
| - name: Configure AWS credentials | |
| uses: aws-actions/configure-aws-credentials@e3dd6a429d7300a6a4c196c26e071d42e0343502 # v4.0.2 | |
| with: | |
| aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }} | |
| aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }} | |
| aws-region: ${{ vars.AWS_REGION }} | |
| - name: Start EC2 runner | |
| id: start-ec2-runner | |
| uses: machulav/ec2-github-runner@1827d6ca7544d7044ddbd2e9360564651b463da2 # v2.3.7 | |
| with: | |
| mode: start | |
| github-token: ${{ secrets.GH_PERSONAL_ACCESS_TOKEN }} | |
| ec2-image-id: ${{ vars.AWS_EC2_AMI }} | |
| ec2-instance-type: g6e.12xlarge | |
| subnet-id: subnet-024298cefa3bedd61 | |
| security-group-id: sg-06300447c4a5fbef3 | |
| iam-role-name: instructlab-ci-runner | |
| aws-resource-tags: > | |
| [ | |
| {"Key": "Name", "Value": "instructlab-ci-github-large-runner"}, | |
| {"Key": "GitHubRepository", "Value": "${{ github.repository }}"}, | |
| {"Key": "GitHubRef", "Value": "${{ github.ref }}"}, | |
| {"Key": "GitHubPR", "Value": "${{ github.event.number }}"} | |
| ] | |
| e2e-large-test: | |
| needs: | |
| - start-large-ec2-runner | |
| runs-on: ${{ needs.start-large-ec2-runner.outputs.label }} | |
| permissions: | |
| pull-requests: write | |
| steps: | |
| - name: "Harden Runner" | |
| # v2.10.1 | |
| uses: step-security/harden-runner@91182cccc01eb5e619899d80e4e971d6181294a7 | |
| with: | |
| egress-policy: audit | |
| - name: Install Packages | |
| run: | | |
| cat /etc/os-release | |
| sudo dnf install -y gcc gcc-c++ make git python3.11 python3.11-devel | |
| - name: Checkout instructlab/instructlab | |
| uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2 | |
| with: | |
| repository: "instructlab/instructlab" | |
| path: "instructlab" | |
| # https://github.com/actions/checkout/issues/249 | |
| fetch-depth: 0 | |
| - name: Checkout instructlab/training | |
| uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2 | |
| with: | |
| repository: "instructlab/training" | |
| path: "training" | |
| # https://github.com/actions/checkout/issues/249 | |
| fetch-depth: 0 | |
| - name: Determine if pr_or_branch is a PR number | |
| id: check_pr | |
| run: | | |
| PR_OR_BRANCH=${{ github.event.inputs.pr_or_branch || 'main' }} # Default to 'main' if not set | |
| if [[ "$PR_OR_BRANCH" =~ ^[0-9]+$ ]]; then | |
| echo "is_pr=true" >> "$GITHUB_OUTPUT" | |
| else | |
| echo "is_pr=false" >> "$GITHUB_OUTPUT" | |
| fi | |
| echo "pr_or_branch=$PR_OR_BRANCH" >> "$GITHUB_OUTPUT" | |
| - name: Check if gh cli is installed | |
| id: gh_cli | |
| run: | | |
| if command -v gh &> /dev/null ; then | |
| echo "gh_cli_installed=true" >> "$GITHUB_OUTPUT" | |
| else | |
| echo "gh_cli_installed=false" >> "$GITHUB_OUTPUT" | |
| fi | |
| - name: Install gh CLI | |
| if: steps.gh_cli.outputs.gh_cli_installed == 'false' | |
| run: | | |
| sudo dnf install 'dnf-command(config-manager)' -y | |
| sudo dnf config-manager --add-repo https://cli.github.com/packages/rpm/gh-cli.repo | |
| sudo dnf install gh --repo gh-cli -y | |
| - name: test gh CLI | |
| run: | | |
| gh --version | |
| - name: set default repo | |
| working-directory: ./training | |
| run: | | |
| gh repo set-default ${{ github.server_url }}/${{ github.repository }} | |
| env: | |
| GH_TOKEN: ${{ secrets.GITHUB_TOKEN }} | |
| - name: Add comment to PR | |
| if: steps.check_pr.outputs.is_pr == 'true' | |
| working-directory: ./training | |
| run: | | |
| gh pr comment "${{ steps.check_pr.outputs.pr_or_branch }}" -b "${{ github.workflow }} workflow launched on this PR: [View run](${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }})" | |
| env: | |
| GH_TOKEN: ${{ secrets.GITHUB_TOKEN }} | |
| - name: Fetch and checkout PR | |
| if: steps.check_pr.outputs.is_pr == 'true' | |
| working-directory: ./training | |
| run: | | |
| gh pr checkout ${{ steps.check_pr.outputs.pr_or_branch }} | |
| env: | |
| GH_TOKEN: ${{ secrets.GITHUB_TOKEN }} | |
| - name: Checkout branch | |
| if: steps.check_pr.outputs.is_pr == 'false' | |
| working-directory: ./training | |
| run: | | |
| git checkout ${{ steps.check_pr.outputs.pr_or_branch }} | |
| - name: Install ilab | |
| working-directory: ./instructlab | |
| run: | | |
| export CUDA_HOME="/usr/local/cuda" | |
| export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64" | |
| export PATH="$PATH:$CUDA_HOME/bin" | |
| python3.11 -m venv --upgrade-deps venv | |
| . venv/bin/activate | |
| nvidia-smi | |
| python3.11 -m pip cache remove llama_cpp_python | |
| CMAKE_ARGS="-DLLAMA_CUDA=on" python3.11 -m pip install . | |
| # https://github.com/instructlab/instructlab/issues/1821 | |
| # install with Torch and build dependencies installed | |
| python3.11 -m pip install packaging wheel setuptools-scm | |
| python3.11 -m pip install .[cuda] -r requirements-vllm-cuda.txt | |
| - name: Update instructlab-training library | |
| working-directory: ./training | |
| run: | | |
| . ../instructlab/venv/bin/activate | |
| pip install . | |
| pip install .[cuda] | |
| - name: Check disk before tests | |
| run: | | |
| df -h | |
| - name: Run e2e test | |
| working-directory: ./instructlab | |
| env: | |
| HF_TOKEN: ${{ secrets.HF_TOKEN }} | |
| run: | | |
| . venv/bin/activate | |
| # set preserve to true so we can retain the logs | |
| ./scripts/e2e-ci.sh -lp | |
| # we know that the file will be named something like f"/training_params_and_metrics_global{os.environ['RANK']}.jsonl" in python | |
| # and we know that it will be written into a directory created by `mktemp -d`. | |
| # Given this information, we can use the following command to find the file: | |
| log_files=$(find /tmp/ -name "training_params_and_metrics_global0.jsonl") | |
| phase_num=1; | |
| for log_file in $log_files; do | |
| mv "${log_file}" phase-${phase_num}-training-log.jsonl | |
| ((phase_num++)) | |
| done | |
| - name: Check disk after tests | |
| run: | | |
| df -h | |
| - name: Upload training logs Phase 1 | |
| uses: actions/upload-artifact@v4 | |
| with: | |
| name: phase-1-training-log.jsonl | |
| path: ./instructlab/phase-1-training-log.jsonl | |
| retention-days: 1 | |
| overwrite: true | |
| - name: Upload training logs Phase 2 | |
| uses: actions/upload-artifact@v4 | |
| with: | |
| name: phase-2-training-log.jsonl | |
| path: ./instructlab/phase-2-training-log.jsonl | |
| retention-days: 1 | |
| overwrite: true | |
| - name: Add comment to PR if the workflow failed | |
| if: failure() && steps.check_pr.outputs.is_pr == 'true' | |
| working-directory: ./training | |
| run: | | |
| gh pr comment "${{ steps.check_pr.outputs.pr_or_branch }}" -b "e2e workflow failed on this PR: [View run](${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}), please investigate." | |
| env: | |
| GH_TOKEN: ${{ secrets.GITHUB_TOKEN }} | |
| - name: Add comment to PR if the workflow succeeded | |
| if: success() && steps.check_pr.outputs.is_pr == 'true' | |
| working-directory: ./training | |
| run: | | |
| gh pr comment "${{ steps.check_pr.outputs.pr_or_branch }}" -b "e2e workflow succeeded on this PR: [View run](${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}), congrats!" | |
| env: | |
| GH_TOKEN: ${{ secrets.GITHUB_TOKEN }} | |
| - name: Post job results to Slack if the workflow failed | |
| if: failure() && steps.check_pr.outputs.is_pr == 'false' | |
| id: slack-report-failure | |
| uses: slackapi/slack-github-action@37ebaef184d7626c5f204ab8d3baff4262dd30f0 # v1.27.0 | |
| with: | |
| # Slack channel id, channel name, or user id to post message. | |
| # See also: https://api.slack.com/methods/chat.postMessage#channels | |
| # You can pass in multiple channels to post to by providing a comma-delimited list of channel IDs. | |
| channel-id: 'e2e-ci-results' | |
| slack-message: "*e2e-nvidia-l40s-x4* job in *${{ github.repository }}* running on branch `${{ steps.check_pr.outputs.pr_or_branch }}` completed *with failures* :meow_sad-rain: | ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}" | |
| env: | |
| SLACK_BOT_TOKEN: ${{ secrets.SON_OF_JEEVES_TOKEN }} | |
| - name: Post job results to Slack if the workflow succeeded | |
| if: success() && steps.check_pr.outputs.is_pr == 'false' | |
| id: slack-report-success | |
| uses: slackapi/slack-github-action@37ebaef184d7626c5f204ab8d3baff4262dd30f0 # v1.27.0 | |
| with: | |
| # Slack channel id, channel name, or user id to post message. | |
| # See also: https://api.slack.com/methods/chat.postMessage#channels | |
| # You can pass in multiple channels to post to by providing a comma-delimited list of channel IDs. | |
| channel-id: 'e2e-ci-results' | |
| slack-message: "*e2e-nvidia-l40s-x4* job in *${{ github.repository }}* running on branch `${{ steps.check_pr.outputs.pr_or_branch }}` completed *successfully* :meow_party: | ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}" | |
| env: | |
| SLACK_BOT_TOKEN: ${{ secrets.SON_OF_JEEVES_TOKEN }} | |
| stop-large-ec2-runner: | |
| needs: | |
| - start-large-ec2-runner | |
| - e2e-large-test | |
| runs-on: ubuntu-latest | |
| if: ${{ always() }} | |
| steps: | |
| - name: "Harden Runner" | |
| # v2.10.1 | |
| uses: step-security/harden-runner@91182cccc01eb5e619899d80e4e971d6181294a7 | |
| with: | |
| egress-policy: audit | |
| - name: Configure AWS credentials | |
| uses: aws-actions/configure-aws-credentials@e3dd6a429d7300a6a4c196c26e071d42e0343502 # v4.0.2 | |
| with: | |
| aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }} | |
| aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }} | |
| aws-region: ${{ vars.AWS_REGION }} | |
| - name: Stop EC2 runner | |
| uses: machulav/ec2-github-runner@1827d6ca7544d7044ddbd2e9360564651b463da2 # v2.3.7 | |
| with: | |
| mode: stop | |
| github-token: ${{ secrets.GH_PERSONAL_ACCESS_TOKEN }} | |
| label: ${{ needs.start-large-ec2-runner.outputs.label }} | |
| ec2-instance-id: ${{ needs.start-large-ec2-runner.outputs.ec2-instance-id }} | |
| loss-graphs: | |
| needs: | |
| - stop-large-ec2-runner | |
| runs-on: ubuntu-latest | |
| if: ${{ always() }} | |
| steps: | |
| - name: "Harden Runner" | |
| # v2.10.1 | |
| uses: step-security/harden-runner@91182cccc01eb5e619899d80e4e971d6181294a7 | |
| with: | |
| egress-policy: audit | |
| - name: Configure AWS credentials | |
| uses: aws-actions/configure-aws-credentials@e3dd6a429d7300a6a4c196c26e071d42e0343502 # v4.0.2 | |
| with: | |
| aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }} | |
| aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }} | |
| aws-region: ${{ vars.AWS_REGION }} | |
| - name: Download loss data Phase 1 | |
| id: phase-1-download-logs | |
| uses: actions/download-artifact@v4 | |
| with: | |
| name: phase-1-training-log.jsonl | |
| path: downloaded-data | |
| - name: Download loss data Phase 2 | |
| id: phase-2-download-logs | |
| uses: actions/download-artifact@v4 | |
| with: | |
| name: phase-2-training-log.jsonl | |
| path: downloaded-data | |
| - name: Checkout instructlab/training | |
| uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2 | |
| with: | |
| repository: "instructlab/training" | |
| path: "training" | |
| fetch-depth: 0 | |
| - name: Install dependencies | |
| working-directory: ./training | |
| run: | | |
| python -m pip install --upgrade pip | |
| pip install -r requirements-dev.txt | |
| - name: Try to upload Phase 1 to s3 | |
| id: phase-1-upload-s3 | |
| continue-on-error: true | |
| run: | | |
| python training/scripts/create-loss-graph.py \ | |
| --log-file "${{ steps.phase-1-download-logs.outputs.download-path }}/phase-1-training-log.jsonl" \ | |
| --output-file "./phase-1-test.md" \ | |
| --phase "1" \ | |
| --aws-region "${{ vars.AWS_REGION }}" \ | |
| --bucket-name "${{ vars.AWS_S3_LOSS_GRAPHS_BUCKET_NAME }}" \ | |
| --base-branch "${GITHUB_REF##*/}" \ | |
| --head-sha "${{ github.sha }}" \ | |
| --pr-number "${{ github.event.number }}" \ | |
| --origin-repository "${{ github.repository }}" | |
| - name: Try to upload Phase 2 to s3 | |
| id: phase-2-upload-s3 | |
| continue-on-error: true | |
| run: | | |
| python training/scripts/create-loss-graph.py \ | |
| --log-file "${{ steps.phase-2-download-logs.outputs.download-path }}/phase-2-training-log.jsonl" \ | |
| --output-file "./phase-2-test.md" \ | |
| --phase "2" \ | |
| --aws-region "${{ vars.AWS_REGION }}" \ | |
| --bucket-name "${{ vars.AWS_S3_LOSS_GRAPHS_BUCKET_NAME }}" \ | |
| --base-branch "${GITHUB_REF##*/}" \ | |
| --head-sha "${{ github.sha }}" \ | |
| --pr-number "${{ github.event.number }}" \ | |
| --origin-repository "${{ github.repository }}" | |
| - name: Check Phase 1 S3 upload status for success | |
| if: steps.phase-1-upload-s3.outcome == 'success' | |
| run: | | |
| echo "Uploaded Phase 1 loss graph to S3." | |
| cat ./phase-1-test.md >> "${GITHUB_STEP_SUMMARY}" | |
| - name: Check Phase 2 S3 upload status for success | |
| if: steps.phase-2-upload-s3.outcome == 'success' | |
| run: | | |
| echo "Uploaded Phase 2 loss graph to S3." | |
| cat ./phase-2-test.md >> "${GITHUB_STEP_SUMMARY}" | |
| - name: Check Phase 1 S3 upload status for failure | |
| if: steps.phase-1-upload-s3.outcome == 'failure' | |
| run: | | |
| echo "::warning::Failed to upload Phase 1 loss graph to S3. This won't block the workflow, but you may want to investigate." | |
| echo "Loss graph upload failed" >> "${GITHUB_STEP_SUMMARY}" | |
| - name: Check Phase 2 S3 upload status for failure | |
| if: steps.phase-2-upload-s3.outcome == 'failure' | |
| run: | | |
| echo "::warning::Failed to upload Phase 2 loss graph to S3. This won't block the workflow, but you may want to investigate." | |
| echo "Loss graph upload failed" >> "${GITHUB_STEP_SUMMARY}" |