机器学习模型 DDL¶

以下 DDL 命令用于创建、查看和管理机器学习模型及其版本。

模型是一个架构级对象，其中包含已在 Snowpark ML 注册表中训练和存储的机器学习模型。通过模型命令，您可以在 SQL 中创建和管理模型。您还可以通过 Snowpark ML 注册表 API 使用 Python 创建和管理模型。

模型监控器允许您监控在 Snowflake 中部署的机器学习模型的性能。

机器学习模型

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.

机器学习模型版本

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.

机器学习模型函数

SHOW FUNCTIONS IN MODEL

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

机器学习模型监控器

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.