This example shows you how to use Azure Machine Learning, GitHub Actions and Event Grid to create a CI/CD flow. The intent of this project is to provide references on how you can implement this with your own projects.
It's recommended that you have a basic understanding of Azure Machine Learning, Event Grid and GitHub Actions. Make sure you have at least contributor access to an Azure Machine Learning workspace. Here are some features and concepts the repo will be using:
- MLOps on Azure Machine Learning
- Azure Machine Learning Pipelines
- Azure Event Grid integration with Azure Machine Learning
You can start with either folder. In order to run this project end to end, complete the set up from both folders.
This folder contains example code to build an image classifaction model to distinguish betweens tacos and burritos. You can replace any of the training code with your own. Additionally this folder also provides a build and deploy yaml in the subfolder workflows that you will trigger as actions.
This folder contains code for a simple Azure function that gets triggered by a model registration event (via event grid). This is important when automating continuous model deployment.