Upgrade model management to SDK v2

  • Article

This article gives a comparison of scenario(s) in SDK v1 and SDK v2.

Create model

  • SDK v1

    import urllib.request
from azureml.core.model import Model

# Register model
model = Model.register(ws, model_name="local-file-example", model_path="mlflow-model/model.pkl")

  • SDK v2

    from azure.ai.ml.entities import Model
from azure.ai.ml.constants import AssetTypes

file_model = Model(
    path="mlflow-model/model.pkl",
    type=AssetTypes.CUSTOM_MODEL,
    name="local-file-example",
    description="Model created from local file."
)
ml_client.models.create_or_update(file_model)

Use model in an experiment/job

  • SDK v1

    model = run.register_model(model_name='run-model-example',
                           model_path='outputs/model/')
print(model.name, model.id, model.version, sep='\t')

  • SDK v2

    from azure.ai.ml.entities import Model
from azure.ai.ml.constants import AssetTypes

run_model = Model(
    path="azureml://jobs/$RUN_ID/outputs/artifacts/paths/model/",
    name="run-model-example",
    description="Model created from run.",
    type=AssetTypes.CUSTOM_MODEL
)

ml_client.models.create_or_update(run_model)

For more information about models, see Work with models in Azure Machine Learning.

Mapping of key functionality in SDK v1 and SDK v2

Functionality in SDK v1 Rough mapping in SDK v2
Model.register ml_client.models.create_or_update
run.register_model ml_client.models.create_or_update
Model.deploy ml_client.begin_create_or_update(blue_deployment)

Next steps

For more information, see the documentation here: