Vantage CI/CD training

The idea of this hands-on is to generate a CI/CD pipeline based on this dummy repository.

Setup

• Use your own Github account to fork this repository.
• Create a virtual environment
• You can train new models or change something in the code as you like
• Now create your CI/CD Pipeline
  • It should have linting
  • Testing
  • Deploying a new model

For the deployment of the Docker that has the predict endpoint I used Heroku.

1. Create an account on https://id.heroku.com/login.
2. Secondly, create a new app with the naming employee-predict-vantage-{YOUR_NAME}.
3. Then, install Heroku CLI using brew tap heroku/brew && brew install heroku for MAC.
4. Create an API key using heroku authorizations:create which can somehow be used in your CI/CD pipeline

Make sure you push the container to your own app name. Change this in app/app.py. Finally, you can check on Heroku logs if the container is deployed successfully and check if you can predict by running: streamlit run /app/app.py