SGD

class vulkpy.nn.SGD

Bases: Optimizer

SGD Optimizer

Use constant learning rate

See also

vulkpy.nn.Adam

Adam optimizer

Methods Summary

init_state(shape)

Initialize Optimizer state

Methods Documentation

init_state(shape: Iterable[int]) → SGDState

Initialize Optimizer state

Parameters:

shape (iterable of ints) – Shape of parameter

Returns:

Optimizer state

Return type:

SGDState

Notes

Currently SGDState is empty, however, we might add some field like momentum in future.

__init__(lr: float)

Initialize Stachostic Gradient Decent (SGD) Optimizer

Parameters:

lr (float) – Learning rate