DataPredict

API Reference - Models - GradientMethodBaseModel

The base model for all machine and deep learning models that uses gradient descent

Constructors

new()

Creates a new gradient method base model.

GradientMethodBaseModel.new(): GradientMethodBaseModelObject

Functions

setAreGradientsSaved()

GradientMethodBaseModel:setAreGradientSaved(option: boolean)

Parameters:

• option: Set whether or not to store gradients.

setAutoResetOptimizers()

Set if the optimizers resets at the end of iterations.

BaseModel:setAutoResetOptimizers(option: boolean)

Parameters:

• option: A boolean value that specifies if optimizers resets at the end of iterations.

getGradients()

GradientMethodBaseModel:getGradients(doNotDeepCopy: boolean): any

Parameters:

• doNotDeepCopy: Set whether or not to get a deep copy of the gradients.

Returns:

• Gradients: The gradient stored inside the model.

setGradients()

GradientMethodBaseModel:setGradients(Gradients: any, doNotDeepCopy: boolean)

Parameters:

• Gradients: The gradient to be stored by the model.

• doNotDeepCopy: Set whether or not to get a deep copy of the gradients.

clearGradients()

Clears stored gradients inside the gradient method base model.

GradientMethodBaseModel:clearGradients()