R/optimizers.R

optimizer_adagrad

Adagrad optimizer.

Description

Adagrad optimizer as described in Adaptive Subgradient Methods for Online Learning and Stochastic Optimization.

Usage

 
optimizer_adagrad( 
  learning_rate = 0.01, 
  epsilon = NULL, 
  decay = 0, 
  clipnorm = NULL, 
  clipvalue = NULL, 
  ... 
) 

Arguments

Arguments Description
learning_rate float >= 0. Learning rate.
epsilon float >= 0. Fuzz factor. If NULL, defaults to k_epsilon().
decay float >= 0. Learning rate decay over each update.
clipnorm Gradients will be clipped when their L2 norm exceeds this value.
clipvalue Gradients will be clipped when their absolute value exceeds this value.
Unused, present only for backwards compatability

Note

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

See Also

Other optimizers: optimizer_adadelta(), optimizer_adamax(), optimizer_adam(), optimizer_nadam(), optimizer_rmsprop(), optimizer_sgd()