question

JeremiahAdams-0775 avatar image
1 Vote"
JeremiahAdams-0775 asked JeremiahAdams-0775 commented

Azure ML Studio and Git/AzureDevOps

We have created an Azure ML Studio based proof of concept. I would like to get the "source code" into a source code repository. We use Git on AzureDevOps.

I'm at a loss where to begin since the designer has no source files with which to interact. All of the project is point-and-click via the Designer. This is a ML project built following this tutorial: https://docs.microsoft.com/en-us/azure/machine-learning/tutorial-designer-automobile-price-train-score

We've created our own models, data, etc and have a solution.

What artifacts from an Azure ML Studio project should be included in a version control system? Where are these files located?

Thanks for any insight.

-jeremiah

azure-machine-learning
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

ramr-msft avatar image
0 Votes"
ramr-msft answered JeremiahAdams-0775 commented

@JeremiahAdams-0775 Thanks for the question. Can you share a snippet of how you are uploading to azure Devops?. Please follow this document: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-deploy-model-designer. Basically you can register a trained model in Designer bring it out with SDK/CLI to deploy it. Just run az ml model download - that will get all of the files.

Also look at the MLOPs demo.


· 1
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

Thanks. This is what I was looking for.

0 Votes 0 ·
romungi-MSFT avatar image
0 Votes"
romungi-MSFT answered

@JeremiahAdams-0775 Azure ML designer is designed to simplify the user interface of using machine learning modules as a drag and drop interface where users can train and create models which can be deployed as a service. The interface's backend which creates connections between modules or the experiments are available to view in the ml portal from the designer interface or from the run ids which are under experiments tab. These run details can be from different sources like automl, designer or simply runs created by using the azure ml SDKs. The runs from designer are always set with a tag of azureml.Designer: true which makes it easy to find the runs from a particular designer experiment. All the runs contain details of the metrics, logs, steps, etc. which can be viewed and can be downloaded as a notebook file. For example:

87754-image.png

You can use these files under your version control but the run details are always available in your workspace under different run ids and you can add additional tags to these runs for managing them under your workspace.




image.png (62.6 KiB)
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.