bartorch.optim.FISTA

class bartorch.optim.FISTA(regularizers=None, *, maxiter=30, step=0.95, eigen=False, hogwild=False, pqr=None, cclambda=0.0, precond=None)

Fast iterative soft thresholding (pics --fista).

Parameters:

• regularizers (Regularizer or iterable of Regularizer, optional) – Terms from bartorch.prox.

• maxiter (int)

• step (float) – Step size (pics -s); 0.95 is what pics uses when none is given.

• eigen (bool) – Scale the step by the largest eigenvalue of the normal operator, estimated with 30 power iterations (pics -e).

• hogwild (bool) – BART’s hogwild setting (pics -H).

• pqr (tuple of float, optional) – Acceleration parameters (p, q, r) (pics --fista_pqr); None keeps BART’s.

• cclambda (float) – Weight of an identity added to the normal operator (pics -q).

__init__(regularizers=None, *, maxiter=30, step=0.95, eigen=False, hogwild=False, pqr=None, cclambda=0.0, precond=None)

Methods

__init__([regularizers, maxiter, step, ...])
fixed_point(image_shape, *[, trainable])	Refused: the ravine step is not the same map twice.
in_library(y, A[, x0])	Solve with BART's own loop, without crossing back into Python.
unrolled(image_shape, *[, trainable])	This solver as a network of maxiter steps, trained end to end.