You are here: Artificial Intelligence > Deep Learning > Deep Learning Tool Interface

Deep Learning Tool Interface

The Deep Learning Tool dialog contains of a number of panels — Model Overview, Model Editing, and Model Training — that you access to complete a model training workflow.

Choose Artificial Intelligence > Deep Learning Tool on the menu bar to open the Deep Learning Tool dialog, shown below.

Deep Learning Tool dialog

Deep Learning Tool dialog

You should note that the information presented on every panel in the dialog is associated with the currently selected model. The available panels are described briefly below.

Model Overview panel… Provides a list of all available deep learning models and detailed information about the selected model. Options to create new models, as well as to preview predictions of trained models, are also available on this panel (see Model Overview Panel).

Model Editing panel… Provides options to edit the network structure of a model (see Model Editing Panel).

Model Training panel… Provides inputs for training a model and options for setting the training parameters (see Model Training Panel).

Additional controls, which appear on multiple panels of the Deep Learning Tool dialog, are described in the following table.

Dialog options

 

Description

Back to Model Overview

Opens the Model Overview panel (see Model Overview Panel).

Go to Editing

Opens the Model Editing panel (see Model Editing Panel).

Go to Training/Train

Opens the Model Training panel (see Model Training Panel), or initiates training when on the Model Training panel.

Reset

Resets the model weights back to their default (untrained) settings.

Note Resetting a model might be necessary in cases in which training is "stuck" in a local minima and no escape is possible. You should also note that the current version of the model will not be overwritten by resetting.

Reload

Loads the last saved version of the model. In this case, changes made since the last save will be lost.

Save

Saves the currently selected model.

Models are automatically saved in a proprietary DRA file format (*.dra extension) with an accompanying JSON file (*.json extension) that contains the model metadata. Files with a .dra extension are known as Dragonfly Deep Learning Model files.

Note Model files are saved in the PythonPluginExtensions\DeepTrainer folder in corresponding folders in the directory pythonAllUsersExtensions . To directly open a model folder, right-click a listed model and then choose Open Model Folder in File Browser in the pop-up menu.

Close

Closes the Deep Learning Tool dialog.

 

Dragonfly Help Live Version