How to change advanced configuration for model training?

This guide will elaborate on, and show how users can change the advanced configurations for models prior to training.

NOTE: This specifically applies to classification & regression problem types

To assist users in the train new model process, our Data Scientists have come up with recommended configurations for each model template as the default.

However, experienced users may prefer to manually make changes to the model's configurations for training:

  • Model Name
  • Training Configuration
  • Hyperparameter Search Grid

To make changes, select "Advanced Configuration".

Changing the model name

Change the model name in the input field. Model name must be unique and must not contain any special character.

Selecting the training configuration

For users on higher-tier subscriptions, they have the option to use more powerful hardware for model training.

Changing the hyperparameter search grid

The Engine currently supports Grid Search for hyperparameter optimization, and users can manually configure the search grid if they desire.