Kaspersky Machine Learning for Anomaly Detection

Training of an ML model element completed with an error

Problem

Training of an ML model element completed with an error. A training error may appear both immediately after training start and after several training epochs are completed.

Solution

If the training of an ML model element completed with an error immediately after training start, make sure that there is data available for all ML model tags within the training interval specified in the Data selection interval setting, while considering markup (

learning indicator

A set of criteria used to determine the data time intervals on which the ML model performs the training.

If the training of an ML model element ended with an error after the training lasted for some time, this means that the selected architecture of the element, parameters of the element, or parameters of training the element are not suitable for accomplishing your task with the available data because the target quality metrics of the ML model cannot be achieved. Try to use an element of a different type or architecture, or change the element parameters or its training parameters.