aml-acceleration-template's Introduction

Azure Machine Learning Acceleration Template

This repo features an Azure Machine Learning (AML) Acceleration template which enables you to quickly onboard your existing Machine Learning code to AML. The template enables a smooth ML development process between your local machine and the Azure Cloud. Furthermore, it includes simple examples for running your model's training and batch inferecing as Machine Learning Pipelines for automation.

If you want to follow a guided approach to use this repo, start with migrating your first workload to AML and walk through the individual sections.

Getting Started

We recommend you to start with migrating your first workload to AML as it covers all prerequisites and outlines a simple and proven step-by-step approach.

This repo follows a pre-defined structure for storing your model code, pipelines, etc.

File/folder	Description
`automation`	Azure DevOps based CI/CD pipelines for MLOps
`instructions\`	A step-by-step guide on how to onboard your first workload to AML
`sample-data\`	Some small sample data used for the template example
`src\`	Model(s) code and other required code assets
`src\model1`	A full end-to-end example for training, real-time and batch inferencing and automation
`pipelines-yaml\`	A set of YAML-based ML pipelines
`pipelines-py\`	A set of Python-based ML pipelines

Authors

Clemens Siebler, AI Technical Specialist GBB EMEA
Erik Zwiefel, AI Principal Technical Specialist GBB Americas
Alan Weaver, AI Senior Technical Specialist GBB EMEA
Alexander Zeltov, AI Principal Technical Specialist GBB Americas

Contributing

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact [email protected] with any additional questions or comments.

aml-acceleration-template's People

Contributors

Stargazers

Watchers

aml-acceleration-template's Issues

Add code examples on how to publish pipelines and test them

Add Python pipeline examples

Add YAML pipeline examples

Include Postman or rest-client example for testing deployed model

Show examples / instructions for a Dev/Test strategy

When you are checking in new code and want to test out the pipeline execution - without deploying a new model, where/how should you do that.

My initial thoughts are:

Everything is done via PR to master
A CI/CD Test Pipeline is established and linked to a dev/test AML workspace
When a new PR is opened, the CI/CD Test Pipeline is a necessary step in the approval process
In that PR, we'll need to run the pipeline and verify the outputs