torch_numopt.algorithms.lbfgs

Limited-memory BFGS (L-BFGS) optimizers.

L-BFGS is a quasi-Newton method that approximates the inverse Hessian using a limited history of past updates (s, y pairs). It is memory-efficient and works well for medium-scale optimization problems. This module provides both a fixed learning rate variant and a line-search variant (recommended).

Classes

LBFGS(params[, lr_init, lr_method, memory_size])	Limited-memory BFGS optimizer with fixed learning rate.
LBFGSLS(params[, lr_init, lr_method, c1, ...])	L-BFGS with line search.
LBFGSMixin(*args[, memory_size])