numdifftools.core.MinMaxStepGenerator

class numdifftools.core.MinMaxStepGenerator(step_min=None, step_max=None, num_steps=10, scale=None, num_extrap=0)[source][source]

Generates a sequence of steps

where
steps = logspace(log10(step_min), log10(step_max), num_steps)
Parameters:

step_min : float, array-like, optional

Defines the minimim step. Default value is:

EPS**(1/scale)*max(log(1+|x|), 1)

where x and scale are supplied at runtime through the __call__ method.

step_max : real scalar, optional

maximum step generated. Default value is:

exp(log(step_min) * scale / (scale + 1.5))

num_steps : scalar integer, optional

defines number of steps generated.

scale : real scalar, optional

scale used in base step. If set to a value it will override the scale supplied at runtime.

__init__(step_min=None, step_max=None, num_steps=10, scale=None, num_extrap=0)[source][source]

Methods

__init__([step_min, step_max, num_steps, ...])