Machine learning model DDL¶

The following DDL commands are used to create, view, and manage machine-learning models and their versions.

A model is a schema-level object that contains a machine learning model that has been trained and stored in the Snowpark ML Registry. Model commands let you create and manage models in SQL. You can also create and manage models in Python using the Snowpark ML Registry API.

Model monitors allow you to monitor the performance of machine learning models you have deployed in Snowflake.

Machine learning models¶

CREATE MODEL	Creates a new machine learning model in the current/specified schema or replaces an existing model.
ALTER MODEL	Modifies the properties for an existing model, including its name, tags, default version, or comment.
SHOW MODELS	Lists the machine learning models that you have privileges to access.
DROP MODEL	Removes a machine learning model from the current/specified schema.

Machine learning model versions¶

ALTER MODEL … ADD VERSION	Adds a new version to an existing model from an internal stage.
ALTER MODEL … DROP VERSION	Removes a version from an existing model.
ALTER MODEL … MODIFY VERSION	Modifies a version of a model, changing the version’s comment or metadata.
SHOW VERSIONS IN MODEL	Lists the versions in a machine learning model.

Machine learing model functions¶

SHOW FUNCTIONS IN MODEL

Shows the models (methods) attached to a machine learing model.

Machine learning model monitors¶

CREATE MODEL MONITOR	Create a new model monitor.
ALTER MODEL MONITOR	Modify the properties of an existing model monitor, including its refresh interval and warehouse, or suspend or resume it.
SHOW MODEL MONITORS	Lists the model monitors that you have privileges to access.
DESCRIBE MODEL MONITOR	Shows the properties of a model monitor.
DROP MODEL MONITOR	Removes a model monitor from the current/specified schema.