vastai-ci-integration

# Vast.ai CI Integration ## Overview Integrate Vast.ai GPU provisioning into CI/CD pipelines. Run GPU-accelerated tests, model validation, and benchmarks as part of your automated workflow using GitHub Actions with the Vast.ai CLI. ## Prerequisites - GitHub repository with Actions enabled - `VASTAI_API_KEY` stored as GitHub Actions secret - Docker image for GPU workload published to a registry ## Instructions ### Step 1: GitHub Actions Workflow ```yaml # .github/workflows/gpu-test.yml name: GPU Tests on: push: branches: [main] pull_request: jobs: gpu-test: runs-on: ubuntu-latest timeout-minutes: 30 steps: - uses: actions/checkout@v4 - name: Install Vast.ai CLI run: | pip install vastai vastai set api-key ${{ secrets.VASTAI_API_KEY }} - name: Provision GPU Instance id: provision run: | # Search for cheapest reliable GPU OFFER_ID=$(vastai search offers \ 'num_gpus=1 gpu_ram>=8 reliability>0.95 dph_total<=0.25' \ --order dph_total --raw --limit 1 \ | python3 -c "import sys,json; print(json.load(sys.stdin)[0]['id'])") # Create instance INSTANCE_ID=$(vastai create instance $OFFER_ID \ --image ghcr.io/${{ github.repository }}/gpu-test:latest \ --disk 20 --raw \ | python3 -c "import sys,json; print(json.load(sys.stdin)['new_contract'])") echo "instance_id=$INSTANCE_ID" >...