nums.models.lbfgs module

class nums.models.lbfgs.BackTrackingLineSearch(model)[source]

Bases: object

execute(X, y, theta, grad, p, rho=0.1, init_alpha=1.0, c=0.0001, min_alpha=1e-10)[source]

f(X, y, theta_prime)[source]

class nums.models.lbfgs.LBFGS(model, m=10, max_iter=100, thresh=0.0001, dtype=<class 'numpy.float64'>)[source]

Bases: object

converged(g)[source]

execute(X, y, theta)[source]

get_H()[source]

get_p(H, g)[source]

class nums.models.lbfgs.LBFGSMemory(k, s, y)[source]: Bases: object