Removing models

This feature is in the beta tier. For more information on feature tiers, see API Tiers.

If we no longer need a trained model and want to free up memory, we can remove the model from the catalog.

1. Syntax

Remove a model from the catalog:

CALL gds.beta.model.drop(modelName: String, failIfMissing: Boolean)
YIELD
    modelInfo: Map,
    trainConfig: Map,
    graphSchema: Map,
    loaded: Boolean,
    stored: Boolean,
    creationTime: DateTime,
    shared: Boolean

Table 1. Parameters
Name	Type	Default	Optional	Description
modelName	String	`n/a`	no	The name of a model stored in the catalog.
failIfMissing	Boolean	`true`	yes	By default, the library will raise an error when trying to remove a non-existing model. When set to `false`, the procedure returns an empty result.

Table 2. Results
Name	Type	Description
modelInfo	Map	Detailed information about the trained model. Always includes the `modelName` and `modelType`, e.g., GraphSAGE. Dependent on the model type, there are additional fields.
trainConfig	Map	The configuration used for training the model.
graphSchema	Map	The schema of the graph on which the model was trained.
loaded	Boolean	True, if the model is loaded in the in-memory model catalog.
stored	Boolean	True, if the model is stored on disk.
creationTime	Datetime	Time when the model was created.
shared	Boolean	True, if the model is shared between users.

2. Example

In this section we are going to demonstrate the usage of gds.beta.model.drop. Assume we trained a model by running train on one of our Machine learning algorithms.

Remove a model from the catalog:

CALL gds.beta.model.drop('my-model')
YIELD modelInfo, loaded, shared, stored
RETURN modelInfo.modelName AS modelName, loaded, shared, stored

Table 3. Results
modelName	loaded	shared	stored
`"my-model"`	`true`	`false`	`false`

In this example, the removed my-model was of the imaginary type some-model-type. The model was loaded in-memory, but neither stored on disk nor published.

If the model name does not exist, an error will be raised.

Was this page helpful?