Python: module softMax

softMax

Helper module to compute an approximate maximum (or minimum) coordinate value in a chosen direction. <string>

The soft max value is given by z_softMax or z_logSumExp z_softMax = ∑_iz_ie^{z_i ⁄ γ} ⁄ ∑_ie^{z_i ⁄ γ} The logSumExp expression is z_logSumExp = γlog∑_ie^{z_i ⁄ γ} where the sum is over all chosen atoms, and z_i = p⋅q_i + r_i, p is a unit vector in the direction given by the projection member, q_i and r_i are the position and radius of atom i, respectively. [The actual exponential computation is relative to the maximum value - to avoid overflows.]

Thus, the soft max is an approximation to the bounds of a collection of coordinates which is smooth as the identity of the involved atoms changes. Constructor: SoftMax(sel, gamma=1, type="softMax") The first argument is a string or atomSel.AtomSel object describing the collection of atoms to describe. The type argument is either "softMax" or "logSumExp". Methods: value() - Return the soft max value. hardValue() - Return the hard (actual) maximum value. selection() - return the sel argument specified in the constructor. derivs(pref, - compute a gradient with respect to coordinates given derivs) a constant pref value, and a derivList.DerivList object. Accessors: gamma - the softness parameter in the above equation: the larger this value, the softer the function (and the worse the approximation is to the hardValue). radii - a vector of radii size len(sel). All radii are 0 by default. projection - direction in which to compute the maximum - the z direction (0,0,1) by default. isMinimum - If True, the minimum value in the given direction is computed - by appropriately setting gamma to a negative value. type - Specify the soft max expression, "softMax" or "logSumExp". # This file was automatically generated by SWIG (http://www.swig.org). # Version 4.0.2 # # Do not make changes to this file unless you know what you are doing--modify # the SWIG interface file instead.

Classes

builtins.object

SoftMax

class SoftMax(builtins.object)

SoftMax(*args, **kwargs)

Methods defined here:

__init__(self, *args, **kwargs): Initialize self. See help(type(self)) for accurate signature.

__repr__ = _swig_repr(self)

derivs(self, *args, **kwargs) -> 'void'

gamma(self, *args, **kwargs) -> 'float_type const'

hardValue(self, *args, **kwargs) -> 'float_type'

isMinimum(self, *args, **kwargs) -> 'bool const'

projection(self, *args, **kwargs) -> 'Vec3 const'

pyXplorHelp(self, *args, **kwargs) -> 'String'

radii(self, *args, **kwargs) -> 'CDSVector< float_type > const'

selection(self, *args, **kwargs) -> 'AtomSel'

setGamma(self, *args, **kwargs) -> 'void'

setIsMinimum(self, *args, **kwargs) -> 'void'

setProjection(self, *args, **kwargs) -> 'void'

setRadii(self, *args, **kwargs) -> 'void'

setType(self, *args, **kwargs) -> 'void'

type(self, *args, **kwargs) -> 'SoftMax::SoftMaxType const'

value(self, *args, **kwargs) -> 'float_type'

Static methods defined here:

__swig_destroy__ = delete_SoftMax(...)

Data descriptors defined here:

__dict__: dictionary for instance variables (if defined)

__weakref__: list of weak references to the object (if defined)

thisown: The membership flag

Data and other attributes defined here:

LOGSUMEXP = 1

SOFTMAX = 0

Functions
		pyXplorHelp(*args) -> 'String'