Module conjugate_gradient 

Nonlinear conjugate gradient (Polak-Ribiere+): steepest descent’s cost per iteration with far better search directions on ill- conditioned valleys, no matrix storage.

Functions

conjugate_gradient

Minimize f from x0 by Polak-Ribiere conjugate gradient with the beta >= 0 safeguard (automatic restart to steepest descent). grad falls back to central finite differences when absent.