imate
C++/CUDA Reference
|
Defines the function. More...
#include <smoothstep.h>
Public Member Functions | |
SmoothStep (double alpha_) | |
Sets the default for the parameter alpha to 1.0 . More... | |
virtual float | function (const float lambda_) const |
virtual double | function (const double lambda_) const |
virtual long double | function (const long double lambda_) const |
Public Member Functions inherited from Function | |
virtual | ~Function () |
Default virtual destructor. More... | |
Public Attributes | |
double | alpha |
Defines the function.
\[ f: \lambda \mapsto \frac{1}{2} \left( 1 + \mathrm{tanh}(\alpha \lambda) \right) \]
where \( \alpha \) is a scale parameter and should be set by this->alpha
member.
The matrix function \( f: \mathbb{R}^{n \times n} \to \mathbb{R}^{n \times n} \) is used in
\[ \mathrm{trace} \left( f(\mathbf{A}) \right). \]
However, instead of a matrix function, the equivalent scalar function \( f: \mathbb{R} \to \mathbb{R} \) is defiend which acts on the eigenvalues of the matrix.
Definition at line 50 of file smoothstep.h.
SmoothStep::SmoothStep | ( | double | alpha_ | ) |
Sets the default for the parameter alpha
to 1.0
.
Definition at line 27 of file smoothstep.cpp.
References alpha.
|
virtual |
[in] | lambda_ | Eigenvalue (or singular value) of matrix. |
Implements Function.
Definition at line 55 of file smoothstep.cpp.
References alpha.
|
virtual |
[in] | lambda_ | Eigenvalue (or singular value) of matrix. |
Implements Function.
Definition at line 41 of file smoothstep.cpp.
References alpha.
|
virtual |
[in] | lambda_ | Eigenvalue (or singular value) of matrix. |
Implements Function.
Definition at line 69 of file smoothstep.cpp.
References alpha.
double SmoothStep::alpha |
Definition at line 57 of file smoothstep.h.
Referenced by function(), and SmoothStep().