Optimizers

This module contains the optimizers

 gradient_descent (x:numpy.ndarray, y:numpy.ndarray, pini:dict, ll:dict,
                   niter=1000, eta=0.001)

	Type	Default	Details
x	ndarray		x data of N elements
y	ndarray		y data of N elements
pini	dict		initial parameters of the loss function
ll	dict		dictionary with the loss (‘loss’), the gradients (‘grads’) and the function (‘fun’) for the regression/classification
niter	int	1000	number of iterations
eta	float	0.001	learning rate
Returns	dict		dictionary containing the vector of the losses (‘loss’) and the parameters (following the keys of pini)

 sgd_epoch (x:numpy.ndarray, y:numpy.ndarray, pini:dict, ll:dict, bs=10,
            eta=0.001)

	Type	Default	Details
x	ndarray		x data of N elements
y	ndarray		y data of N elements
pini	dict		initial parameters of the loss function
ll	dict		dictionary with the loss (‘loss’), the gradients (‘grads’) and the function (‘fun’) for the regression/classification
bs	int	10	Batch size
eta	float	0.001	learning rate
Returns	dict		dictionary containing the vector of the losses (‘loss’) and the parameters (following the keys of pini)

 sgd (x:numpy.ndarray, y:numpy.ndarray, pini:dict, ll:dict, bs=10,
      eta=0.001, niter=1000)

	Type	Default	Details
x	ndarray		x data of N elements
y	ndarray		y data of N elements
pini	dict		initial parameters of the loss function
ll	dict		dictionary with the loss (‘loss’), the gradients (‘grads’) and the function (‘fun’) for the regression/classification
bs	int	10	Batch size
eta	float	0.001	learning rate
niter	int	1000	number of epochs
Returns	dict		dictionary containing the vector of the losses (‘loss’) and the parameters (following the keys of pini)