Rprop Optimizer

neualpy.optimizer.Rprop(learning_rate=0.01, etas=(0.5, 1.2), step_sizes=(1e-06, 50.0))

info

Rprop Optimizer is mostly stable and can be used for any project. In the future, any chance of breaking changes is very low.

Applies Rprop algorithm, Rprop:Resilient backpropagation algorithm

For more information, check this page.

Supported Arguments

• learning_rate=0.01: (Float) Learning Rate for the optimizer
• etas=(0.5,1.2): (Tuple) pair of (etaminus, etaplis), that are multiplicative increase and decrease factors
• step_sizes=(le-06,50): (Tuple) a pair of minimal and maximal allowed step sizes for the optimizer

Code Example

from neuralpy.models import Sequential

from neuralpy.optimizer import Rprop

...

# Rest of the imports

...

model = Sequential()

...

# Rest of the architecture

...

model.compile(optimizer=Rprop(), ...))