A lightweight and (hopefully) easy-to-use template for Azure Machine Learning Services.
This repo serves as a template for creating an end-to-end solution on Azure Machine Learning Services. Its purpose is to catalyze ML projects and help them get faster to production.
Topic covered:
- setting up the environment
- exploratory data analysis
- creating a model training pipeline incl. hyperparameter optimization
- creating and deploying an inferencing webservice incl. authentication and logging.
AML's AutoML is not included yet but maybe in future.
The repo can be used with VS.Code, PyCharm, Jupyter Lab, Jupyter, ... whereby the experience with VS.Code is probably the best.
I am open and happy to accept pull requests. If you have suggestions I should integrate, let me know.
As always - feel free to use but don't blame me if things go wrong. You have been warned!
CAUTION
Some of the code contained is outdated because AML has new features meanwhile incl. managed endpoints. To get the latest and greatest, check Azure ML's documentation and the samples here.