REGISTERED_MODELS

Lists details about registered machine learning models in the database.

Lists details about registered machine learning models in the database. The table lists only registered models for which the caller has USAGE privileges.

For a table that lists all models, registered and unregistered, see MODELS.

Column Name Data Type Description
REGISTERED_NAME VARCHAR The abstract name to which the model is registered. This REGISTERED_NAME can represent a group of models for a higher-level application, where each model in the group has a unique version number.
REGISTERED_VERSION INTEGER The unique version number of the model under its specified REGISTERED_NAME.
STATUS INTEGER

The status of the registered model, one of the following:

  • under_review: Status assigned to newly registered models.

  • staging: Model is targeted for A/B testing against the model currently in production.

  • production: Model is in production for its specified application. Only one model can be in production for a given registered_name at one time.

  • archived: Status of models that were previously in production. Archived models can be returned to production at any time.

  • declined: Model is no longer in consideration for production.

  • unregistered: Model is removed from the versioning environment. The model does not appear in the REGISTERED_MODELS system table.

REGISTERED_TIME VARCHAR The time at which the model was registered.
MODEL_ID INTEGER The model's internal ID.
SCHEMA_NAME VARCHAR The name of the schema that contains the model.
MODEL_NAME VARCHAR The name of the model. [schema_name.]model_name can be used to uniquely identify a model, as can the combination of its REGISTERED_NAME and REGISTERED_VERSION.
MODEL_TYPE VARCHAR The type of algorithm used to create the model.
CATEGORY VARCHAR

The category of the model, one of the following:

  • VERTICA_MODELS

  • PMML

  • TENSORFLOW

By default, models created in Vertica are assigned to the VERTICA_MODELS category.

Example

If a user with the MLSUPERVISOR role queries REGISTERED_MODELS, all registered models are listed:

=> SELECT * FROM REGISTERED_MODELS;
  registered_name | registered_version |    status    |        registered_time        |      model_id     | schema_name |    model_name     |      model_type       |    category
------------------+--------------------+--------------+-------------------------------+-------------------+-------------+-------------------+-----------------------+----------------
 linear_reg_app   |                  2 | UNDER_REVIEW | 2023-01-29 09:09:00.082166-04 | 45035996273714020 | public      | linear_reg_spark1 | PMML_REGRESSION_MODEL | PMML
 linear_reg_app   |                  1 | PRODUCTION   | 2023-01-24 06:19:04.553102-05 | 45035996273850350 | public      | native_linear_reg | LINEAR_REGRESSION     | VERTICA_MODELS
 logistic_reg_app |                  2 | PRODUCTION   | 2023-01-25 08:45:11.279013-02 | 45035996273855542 | public      | log_reg_cgd       | LOGISTIC_REGRESSION   | VERTICA_MODELS
 logistic_reg_app |                  1 | ARCHIVED     | 2023-01-22 04:29:25.990626-02 | 45035996273853740 | public      | log_reg_bfgs      | LOGISTIC_REGRESSION   | VERTICA_MODELS
(4 rows)

See also