Move Azure ML designer to pipeline

Question

Hi pipeline in Desinger in Azure Machine Learning.

How can I move those pipeline to different workspace for different env.

Answer 1

Hi @RSingh

Thank you for using the Microsoft Q&A forum.

To move a pipeline created in Azure Machine Learning Designer to a different workspace for a different environment, there are two options available.

Option 1: Use Azure Machine Learning registry to share models, components, and environments across workspaces. First, create an Azure Machine Learning registry and a workspace for the new environment. Then, create an environment and component in the registry. You can use the same environment and component across multiple workspaces. Next, run the pipeline job in the original workspace using the component from the registry. The compute resources and training data will be local to the workspace. Finally, create a model in the registry from the output of the pipeline job and deploy the model to an online endpoint in the new workspace for inference.

Option 2: Use the Azure Machine Learning workspace move feature to migrate the workspace and its contents, including experiment jobs and environments, as a single, automated step. However, it's essential to note that Azure Machine Learning workspace move is currently in public preview and not recommended for production workloads. For detailed steps and prerequisites, please refer to the Azure documentation on "Move Azure Machine Learning workspaces between subscriptions (preview)".

Please keep in mind the limitations and considerations mentioned in the documentation.

I hope the information provided will assist you in making an informed decision for your specific use case.

Thank you.