Fortinet white logo
Fortinet white logo

CLI Reference

mlflow

mlflow

Use this command to manually train the machine learning models, exclude assets, and validate artifacts.

For more information about the Machine Learning Anomaly Detection feature, see the FortiAnalyzer Administration Guide.

Syntax

execute mlflow cancel <artifact_name>

execute mlflow delete <artifact_name>

execute mlflow deploy <artifact_name>

execute mlflow exclusion-list {artifact | model} {add | list | remove} <asset>

execute mlflow test <artifact_name> <start_time> <end_time>

execute mlflow train <model_type> <start_time> <end_time> [train-assets-limit]

execute mlflow undeploy <artifact_name>

Variable

Description

cancel <artifact_name>

Cancel artifact training.

delete <artifact_name>

Permanently delete artifact.

deploy <artifact_name>

Deploy artifact for active inference.

exclusion-list {artifact | model} {add | list | remove} <asset>

Manage ML exclusion-lists.

  • artifact: Artifact-level exclusion-list operations.

    • add: Add asset to artifact exclusion-list.

    • list: List exclusion-listed assets for artifact.

    • remove: Remove asset from artifact exclusion-list.

  • model: Model-level exclusion-list operations.

    • add: Add asset to model exclusion-list.

    • list: List exclusion-listed assets for model type.

    • remove: Remove asset from model exclusion-list.

test <artifact_name> <start_time> <end_time>

Test/validate artifact on specified time range.

  • <start_time> <end_time>: Start time and end time can be entered in the following format: YYYY-MM-DD HH:MM.

train <model_type> <start_time> <end_time> [train-assets-limit]

Train the machine learning model on specified time range.

  • <model_type>: Model type includes login-anomaly, traffic-download-anomaly, traffic-upload-anomaly.

  • <start_time> <end_time>: Start time and end time can be entered in the following format: YYYY-MM-DD HH:MM.

  • [train-assets-limit]: Optionally, set the maximum number of assets to train (1 - 100000).

Example:

execute mlflow train login-anomaly '2026-02-05' '2026-05-05' 

undeploy <artifact_name>

Undeploy artifact from active inference.

mlflow

mlflow

Use this command to manually train the machine learning models, exclude assets, and validate artifacts.

For more information about the Machine Learning Anomaly Detection feature, see the FortiAnalyzer Administration Guide.

Syntax

execute mlflow cancel <artifact_name>

execute mlflow delete <artifact_name>

execute mlflow deploy <artifact_name>

execute mlflow exclusion-list {artifact | model} {add | list | remove} <asset>

execute mlflow test <artifact_name> <start_time> <end_time>

execute mlflow train <model_type> <start_time> <end_time> [train-assets-limit]

execute mlflow undeploy <artifact_name>

Variable

Description

cancel <artifact_name>

Cancel artifact training.

delete <artifact_name>

Permanently delete artifact.

deploy <artifact_name>

Deploy artifact for active inference.

exclusion-list {artifact | model} {add | list | remove} <asset>

Manage ML exclusion-lists.

  • artifact: Artifact-level exclusion-list operations.

    • add: Add asset to artifact exclusion-list.

    • list: List exclusion-listed assets for artifact.

    • remove: Remove asset from artifact exclusion-list.

  • model: Model-level exclusion-list operations.

    • add: Add asset to model exclusion-list.

    • list: List exclusion-listed assets for model type.

    • remove: Remove asset from model exclusion-list.

test <artifact_name> <start_time> <end_time>

Test/validate artifact on specified time range.

  • <start_time> <end_time>: Start time and end time can be entered in the following format: YYYY-MM-DD HH:MM.

train <model_type> <start_time> <end_time> [train-assets-limit]

Train the machine learning model on specified time range.

  • <model_type>: Model type includes login-anomaly, traffic-download-anomaly, traffic-upload-anomaly.

  • <start_time> <end_time>: Start time and end time can be entered in the following format: YYYY-MM-DD HH:MM.

  • [train-assets-limit]: Optionally, set the maximum number of assets to train (1 - 100000).

Example:

execute mlflow train login-anomaly '2026-02-05' '2026-05-05' 

undeploy <artifact_name>

Undeploy artifact from active inference.