keraflow.optimizers.Adagrad Class Reference

Adagrad optimizer. More...

Inheritance diagram for keraflow.optimizers.Adagrad:

Public Member Functions
def	__init__
	It is recommended to leave the parameters of this optimizer at their default values. More...
Public Member Functions inherited from keraflow.optimizers.Optimizer
def	__init__

Detailed Description

Adagrad optimizer.

• Reference: Adaptive Subgradient Methods for Online Learning and Stochastic Optimization

Constructor & Destructor Documentation

def keraflow.optimizers.Adagrad.__init__	(		self,
			lr = 0.01,
			epsilon = 1e-8,
			kwargs
	)

It is recommended to leave the parameters of this optimizer at their default values.

Parameters

lr	float >= 0. Learning rate.
epsilon	float >= 0.
kwargs	see Optimizer.__init__