opto.optimizers.optimizer.AbstractOptimizer#

class AbstractOptimizer[source]#

An optimizer is responsible for updating the parameters based on the feedback.

Attributes

propagator

Return a Propagator object that can be used to propagate feedback in backward.

Methods

step()	Update the parameters based on the feedback.
zero_feedback()	Reset the feedback.

__init__(parameters: List[ParameterNode], *args, **kwargs)[source]#

__new__(**kwargs)#