Optimizer

class Optimizer

Bases: object

Base optimizer class that all other optimizers inherit

Example:

class CustomOptimizer(Optimizer):
    def build(self, net: 'NeuralNet') -> None:
        self.net = net

        ... Any aditional variables you wish to initialize

    def step(self) -> None:
        for param, grad in self.net.get_params_and_grads():
            ... Update params and grads

__init__()

Initialize self. See help(type(self)) for accurate signature.

Methods

__init__()	Initialize self.
build(net)	Set aditional parameters needed for step.
step()	At each step the neural net is updated.

build(net: NeuralNet) → None

Set aditional parameters needed for step.

Raises

NotImplementedError –

step() → None

At each step the neural net is updated.

Raises

NotImplementedError –