Instructions

Pipelines

Instructions

From within the directory of each pipeline, you can run it via:

az ml run submit-pipeline -y pipeline.yml -n <experiment_name>

train pipeline

Direct execution

You can submit the pipeline asynchronously via:

az ml run submit-pipeline -n training-pipeline -y pipeline.yml

Publishing

You can also publish the pipeline as a pipeline draft:

az ml pipeline create-draft -e training-pipline-draft -n training-pipeline -y pipeline.yml

And then submit that draft as an experiment for testing:

az ml pipeline submit-draft -i <pipeline_draft_id>

Once you confirmed that the pipeline draft works fine, you can fully publish the pipeline as an endpoint so that others can use it:

az ml pipeline publish-draft -i <pipeline_draft_id>

The pipeline will then show up under Endpoints --> Pipeline endpoints in the Azure Machine Learning studio.

batch-inference pipeline

You can also publish the pipeline as a pipeline draft:

az ml pipeline create-draft -e batch-inferencing-pipline-draft -n batch-inferencing-pipeline -y pipeline.yml

And then submit that draft as an experiment for testing:

az ml pipeline submit-draft -i <pipeline_draft_id>

Once you confirmed that the pipeline draft works fine, you can fully publish the pipeline as an endpoint so that others can use it:

az ml pipeline publish-draft -i <pipeline_draft_id>

The pipeline will then show up under Endpoints --> Pipeline endpoints in the Azure Machine Learning studio.