lbfgsb.get_grad_projection_inf_norm#

lbfgsb.get_grad_projection_inf_norm(x: ndarray[tuple[Any, ...], dtype[float64]], grad: ndarray[tuple[Any, ...], dtype[float64]], lbounds: ndarray[tuple[Any, ...], dtype[float64]], ubounds: ndarray[tuple[Any, ...], dtype[float64]]) → float[source]#

Get the infinite norm of the projected gradient.

Parameters:

  • x (NDArrayFloat) – Parameter vector.

  • grad (NDArrayFloat) – Gradient of the parameter vector.

  • lbounds (NDArrayFloat) – Lower bounds.

  • ubounds (NDArrayFloat) – Upper bounds.

Returns:

The scaling factor.

Return type:

float