`scipy.optimize._minpack:_lmder`

Signature

built-in_lmder(fun, Dfun, x0, args, full_output, col_deriv, ftol, xtol, gtol, maxfev, factor, diag)

Levenberg-Marquardt algorithm with user-supplied Jacobian.

Minimizes sum(fun(x)**2) using analytical Jacobian.

fun : callable: Residual function. Takes x as first argument, returns array of residuals.
Dfun : callable: Jacobian of fun. Returns matrix of partial derivatives.
x0 : ndarray, shape (n,): Initial guess, float64.
args : tuple: Extra arguments passed to fun and Dfun.
full_output : int: Non-zero to return optional outputs.
col_deriv : int: Non-zero if Dfun returns column derivatives. Default: 1.
ftol : float: Relative tolerance for residual sum of squares. Default: 1.49012e-8.
xtol : float: Relative tolerance for solution. Default: 1.49012e-8.
gtol : float: Orthogonality tolerance. Default: 0.0.
maxfev : int: Maximum number of function evaluations. Default: 100*(n+1).
factor : float: Initial step bound. Default: 100.0.
diag : ndarray, shape (n,) or None: Scaling factors for variables, float64.

x : ndarray, shape (n,): Solution vector.
infodict : dict: Dictionary with keys: fvec (ndarray), nfev (int), njev (int), fjac (ndarray), ipvt (ndarray), qtf (ndarray). Only returned if full_output != 0.
info : int: Exit status. 1-4 indicate success.

If full_output is 0: returns (x, info) If full_output != 0: returns (x, infodict, info)