ott.neural.solvers.conjugate.FenchelConjugateLBFGS

ott.neural.solvers.conjugate.FenchelConjugateLBFGS#

class ott.neural.solvers.conjugate.FenchelConjugateLBFGS(gtol=0.001, max_iter=10, max_linesearch_iter=10, linesearch_type='backtracking', linesearch_init='increase', increase_factor=1.5)[source]#

Solve for the conjugate using LBFGS.

Parameters:
  • gtol (float) – gradient tolerance

  • max_iter (int) – maximum number of iterations

  • max_linesearch_iter (int) – maximum number of line search iterations

  • linesearch_type (Literal['zoom', 'backtracking', 'hager-zhang']) – type of line search

  • linesearch_init (Literal['increase', 'max', 'current']) – strategy for line search initialization

  • increase_factor (float) – factor by which to increase the step size during the line search

Methods

solve(f, y[, x_init])

Solve for the conjugate.

Attributes

gtol

increase_factor

linesearch_init

linesearch_type

max_iter

max_linesearch_iter