GenAIHub
EN | PT
← Back to Technical Section

GitHub Actions for GenAI

Automate CI/CD pipelines for training, testing, and deploying AI models

What is GitHub Actions?

GitHub Actions is a CI/CD platform that automates build, test, and deployment pipelines directly from your GitHub repository. For GenAI projects, it enables automated model training, testing, validation, and deployment workflows triggered by code changes, schedules, or manual triggers.

Key Innovation: GitHub Actions integrates natively with GitHub repositories, providing free compute minutes, matrix builds for testing across environments, and seamless secrets management for API keys.

2,000

Free minutes/month

15K+

Marketplace Actions

GPU

Self-hosted Runners

∞

Workflow Triggers

Core Concepts

Workflow Components

Workflow

YAML config file in .github/workflows/

Job

Set of steps running on same runner

Step

Individual task (run command/action)

Action

Reusable unit from Marketplace

Trigger push, PR, cron Job: test Step: checkout Step: run tests Job: build Step: build image Step: push registry Job: deploy Step: deploy prod

Basic Workflow Structure

Minimal CI Workflow

# .github/workflows/ci.yml
name: CI Pipeline

on:
  push:
    branches: [main]
  pull_request:
    branches: [main]

jobs:
  test:
    runs-on: ubuntu-latest
    
    steps:
      - name: Checkout code
        uses: actions/checkout@v4
      
      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: '3.11'
          cache: 'pip'
      
      - name: Install dependencies
        run: |
          pip install -r requirements.txt
      
      - name: Run tests
        run: |
          pytest tests/ -v --cov=src

GenAI Workflow Examples

Model Training Pipeline

# .github/workflows/train.yml
name: Train Model

on:
  workflow_dispatch:  # Manual trigger
    inputs:
      model_type:
        description: 'Model architecture'
        required: true
        default: 'bert-base'
      epochs:
        description: 'Number of epochs'
        required: true
        default: '10'

jobs:
  train:
    runs-on: self-hosted  # GPU runner
    
    steps:
      - uses: actions/checkout@v4
      
      - name: Configure AWS credentials
        uses: aws-actions/configure-aws-credentials@v4
        with:
          aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }}
          aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
          aws-region: us-east-1
      
      - name: Download training data
        run: aws s3 cp s3://my-bucket/data/ ./data/ --recursive
      
      - name: Train model
        env:
          WANDB_API_KEY: ${{ secrets.WANDB_API_KEY }}
        run: |
          python train.py \
            --model ${{ inputs.model_type }} \
            --epochs ${{ inputs.epochs }} \
            --output ./models/
      
      - name: Upload model artifacts
        uses: actions/upload-artifact@v4
        with:
          name: trained-model
          path: ./models/

Deploy to Cloud Run

# .github/workflows/deploy.yml
name: Deploy to Cloud Run

on:
  push:
    branches: [main]

env:
  PROJECT_ID: my-gcp-project
  SERVICE: genai-api
  REGION: us-central1

jobs:
  deploy:
    runs-on: ubuntu-latest
    
    steps:
      - uses: actions/checkout@v4
      
      - name: Authenticate to Google Cloud
        uses: google-github-actions/auth@v2
        with:
          credentials_json: ${{ secrets.GCP_SA_KEY }}
      
      - name: Set up Cloud SDK
        uses: google-github-actions/setup-gcloud@v2
      
      - name: Build and push Docker image
        run: |
          gcloud builds submit \
            --tag gcr.io/$PROJECT_ID/$SERVICE
      
      - name: Deploy to Cloud Run
        run: |
          gcloud run deploy $SERVICE \
            --image gcr.io/$PROJECT_ID/$SERVICE \
            --platform managed \
            --region $REGION \
            --allow-unauthenticated

Workflow Triggers

🔄 push

Trigger on commits to specific branches

on:
  push:
    branches: [main, dev]

🔀 pull_request

Trigger on PR events (opened, sync)

on:
  pull_request:
    types: [opened, synchronize]

⏰ schedule

Cron-based scheduled runs

on:
  schedule:
    - cron: '0 6 * * 1'  # Mon 6AM

🎯 workflow_dispatch

Manual trigger with inputs

on:
  workflow_dispatch:
    inputs:
      environment: 
        type: choice
        options: [dev, prod]

Secrets & Environment Variables

Security: Never hardcode API keys or credentials. Use GitHub Secrets (Settings → Secrets → Actions) and reference them as ${{ secrets.SECRET_NAME }}

Common Secrets for GenAI

Secret Name Purpose
OPENAI_API_KEY OpenAI API access
GCP_SA_KEY Google Cloud Service Account JSON
AWS_ACCESS_KEY_ID AWS authentication
WANDB_API_KEY Weights & Biases experiment tracking
DOCKER_PASSWORD Docker Hub registry access

Matrix Builds

Run tests across multiple Python versions, operating systems, or model configurations in parallel.

jobs:
  test:
    strategy:
      matrix:
        python-version: ['3.9', '3.10', '3.11']
        os: [ubuntu-latest, macos-latest]
        model: [gpt-4, claude-3]
    
    runs-on: ${{ matrix.os }}
    
    steps:
      - uses: actions/checkout@v4
      
      - name: Set up Python ${{ matrix.python-version }}
        uses: actions/setup-python@v5
        with:
          python-version: ${{ matrix.python-version }}
      
      - name: Test with ${{ matrix.model }}
        env:
          MODEL_NAME: ${{ matrix.model }}
        run: pytest tests/

Best Practices

Tip: Use actions/cache to cache pip packages, model weights, and dependencies between runs for faster builds.

  • Pin action versions: Use @v4 not @latest for reproducibility
  • Use caching: Cache pip, npm, model downloads to speed up workflows
  • Timeout limits: Set timeout-minutes to prevent runaway jobs
  • Concurrency groups: Cancel redundant runs on same branch
  • Reusable workflows: Extract common patterns to separate workflow files
  • Self-hosted runners: Use GPU runners for model training jobs
  • Split jobs: Separate test, build, deploy for parallel execution
  • Environment protection: Require approvals for production deploys

Popular Actions for GenAI

actions/checkout@v4

Clone repository code

actions/setup-python@v5

Install Python with caching

docker/build-push-action@v5

Build and push Docker images

google-github-actions/auth@v2

Authenticate to Google Cloud

aws-actions/configure-aws-credentials@v4

AWS authentication

actions/upload-artifact@v4

Store build outputs

Learn More

Essential Resources

Related Topics

Test Your Knowledge

Score 8/10 or higher to pass