cloudscaleml / seer

MIT License
13 stars 15 forks source link

MLOps

This exercise is designed to add a CI/CD process to an existing machine learning project!

Uploading Some Data

To get this to work we need some data. This can be done by downloading a zip file and uploading it to the storage attached to the Azure Machine Learning workspace. Make sure to follow the instructions in order to set this up.

Build Pipeline

The goal behind the build pipeline is to create a process whereby every time code is changed in our repo, the machine learning pipeline is updated, run, and a model is created. The process is a bit involved but not impossible! The instructions are fairly detailed and should set you up for success.

Release Pipeline

The release pipeline is the delivery process of the CI/CD pipeline. It is often advantageous to seperate the two process in order to control how we deliver our machine learning model (i.e. one version might not work well so we want to avoid releasing it). The instructions to set this up are a bit involved but not impossible!

Tying it all together!

Now you should be able to make a change to the code and see all of the services work together to product a consumable endpoint which can be used in your software!