mic::neural_nets::optimization::AdamID< eT > Class Template Reference

AdamID - ADAM with integral and derivative coefficients. More...

#include <AdamID.hpp>

Inheritance diagram for mic::neural_nets::optimization::AdamID< eT >:

Collaboration diagram for mic::neural_nets::optimization::AdamID< eT >:

Public Member Functions
	AdamID (size_t rows_, size_t cols_, eT beta1_=0.9, eT beta2_=0.999, eT eps_=1e-8)
mic::types::MatrixPtr< eT >	calculateUpdate (mic::types::MatrixPtr< eT > x_, mic::types::MatrixPtr< eT > dx_, eT learning_rate_)
Public Member Functions inherited from mic::neural_nets::optimization::OptimizationFunction< eT >
	OptimizationFunction ()
virtual	~OptimizationFunction ()
	Virtual destructor - empty. More...
virtual void	update (mic::types::MatrixPtr< eT > p_, mic::types::MatrixPtr< eT > dp_, eT learning_rate_, eT decay_=0.0)
virtual void	update (mic::types::MatrixPtr< eT > p_, mic::types::MatrixPtr< eT > x_, mic::types::MatrixPtr< eT > y_, eT learning_rate_=0.001)

Protected Attributes
eT	beta1
	Decay rate 1 (momentum for past gradients). More...
eT	beta2
	Decay rate 2 (momentum for past squared gradients). More...
eT	eps
	Smoothing term that avoids division by zero. More...
eT	beta1_powt
	Decay rate 1 to the power of t - bias correction. More...
eT	beta2_powt
	Decay rate 2 to the power of t - bias correction. More...
mic::types::MatrixPtr< eT >	Edx
	Decaying average of gradients up to time t - E[g]. More...
mic::types::MatrixPtr< eT >	Edx2
	Decaying average of squared gradients up to time t - E[g^2]. More...
mic::types::MatrixPtr< eT >	dx_prev
	Previous value of gradients. More...
mic::types::MatrixPtr< eT >	delta
	Calculated update. More...

Detailed Description

template<typename eT = float>
class mic::neural_nets::optimization::AdamID< eT >

AdamID - ADAM with integral and derivative coefficients.

Author

tkornuta

Definition at line 39 of file AdamID.hpp.

Constructor & Destructor Documentation

template<typename eT = float>

mic::neural_nets::optimization::AdamID< eT >::AdamID ( size_t  rows_, size_t  cols_, eT  beta1_ = 0.9, eT  beta2_ = 0.999, eT  eps_ = 1e-8  )

inline

Constructor. Sets dimensions, values of decay (beta1=0.9 and beta2=0.999) and eps (default=1e-8).

Parameters

rows_	Number of rows of the updated matrix/its gradient.
cols_	Number of columns of the updated matrix/its gradient.

Definition at line 47 of file AdamID.hpp.

References mic::neural_nets::optimization::AdamID< eT >::beta1, mic::neural_nets::optimization::AdamID< eT >::beta1_powt, mic::neural_nets::optimization::AdamID< eT >::beta2, mic::neural_nets::optimization::AdamID< eT >::beta2_powt, mic::neural_nets::optimization::AdamID< eT >::delta, mic::neural_nets::optimization::AdamID< eT >::dx_prev, mic::neural_nets::optimization::AdamID< eT >::Edx, and mic::neural_nets::optimization::AdamID< eT >::Edx2.

Member Function Documentation

template<typename eT = float>

mic::types::MatrixPtr<eT> mic::neural_nets::optimization::AdamID< eT >::calculateUpdate ( mic::types::MatrixPtr< eT >  x_, mic::types::MatrixPtr< eT >  dx_, eT  learning_rate_  )

inlinevirtual

Calculates the update according to the AdamID update rule.

Parameters

x_	Pointer to the current matrix.
dx_	Pointer to current gradient of that matrix.
learning_rate_	Learning rate (default=0.001).

Implements mic::neural_nets::optimization::OptimizationFunction< eT >.

Definition at line 73 of file AdamID.hpp.

References mic::neural_nets::optimization::AdamID< eT >::beta1, mic::neural_nets::optimization::AdamID< eT >::beta1_powt, mic::neural_nets::optimization::AdamID< eT >::beta2, mic::neural_nets::optimization::AdamID< eT >::beta2_powt, mic::neural_nets::optimization::AdamID< eT >::delta, mic::neural_nets::optimization::AdamID< eT >::Edx, mic::neural_nets::optimization::AdamID< eT >::Edx2, and mic::neural_nets::optimization::AdamID< eT >::eps.

Member Data Documentation

template<typename eT = float>

eT mic::neural_nets::optimization::AdamID< eT >::beta1

protected

Decay rate 1 (momentum for past gradients).

Definition at line 118 of file AdamID.hpp.

Referenced by mic::neural_nets::optimization::AdamID< eT >::AdamID(), and mic::neural_nets::optimization::AdamID< eT >::calculateUpdate().

template<typename eT = float>

eT mic::neural_nets::optimization::AdamID< eT >::beta1_powt

protected

Decay rate 1 to the power of t - bias correction.

Definition at line 127 of file AdamID.hpp.

Referenced by mic::neural_nets::optimization::AdamID< eT >::AdamID(), and mic::neural_nets::optimization::AdamID< eT >::calculateUpdate().

template<typename eT = float>

eT mic::neural_nets::optimization::AdamID< eT >::beta2

protected

Decay rate 2 (momentum for past squared gradients).

Definition at line 121 of file AdamID.hpp.

Referenced by mic::neural_nets::optimization::AdamID< eT >::AdamID(), and mic::neural_nets::optimization::AdamID< eT >::calculateUpdate().

template<typename eT = float>

eT mic::neural_nets::optimization::AdamID< eT >::beta2_powt

protected

Decay rate 2 to the power of t - bias correction.

Definition at line 130 of file AdamID.hpp.

Referenced by mic::neural_nets::optimization::AdamID< eT >::AdamID(), and mic::neural_nets::optimization::AdamID< eT >::calculateUpdate().

template<typename eT = float>

mic::types::MatrixPtr<eT> mic::neural_nets::optimization::AdamID< eT >::delta

protected

Calculated update.

Definition at line 142 of file AdamID.hpp.

Referenced by mic::neural_nets::optimization::AdamID< eT >::AdamID(), and mic::neural_nets::optimization::AdamID< eT >::calculateUpdate().

template<typename eT = float>

mic::types::MatrixPtr<eT> mic::neural_nets::optimization::AdamID< eT >::dx_prev

protected

Previous value of gradients.

Definition at line 139 of file AdamID.hpp.

Referenced by mic::neural_nets::optimization::AdamID< eT >::AdamID().

template<typename eT = float>

mic::types::MatrixPtr<eT> mic::neural_nets::optimization::AdamID< eT >::Edx

protected

Decaying average of gradients up to time t - E[g].

Definition at line 133 of file AdamID.hpp.

Referenced by mic::neural_nets::optimization::AdamID< eT >::AdamID(), and mic::neural_nets::optimization::AdamID< eT >::calculateUpdate().

template<typename eT = float>

mic::types::MatrixPtr<eT> mic::neural_nets::optimization::AdamID< eT >::Edx2

protected

Decaying average of squared gradients up to time t - E[g^2].

Definition at line 136 of file AdamID.hpp.

Referenced by mic::neural_nets::optimization::AdamID< eT >::AdamID(), and mic::neural_nets::optimization::AdamID< eT >::calculateUpdate().

template<typename eT = float>

eT mic::neural_nets::optimization::AdamID< eT >::eps

protected

Smoothing term that avoids division by zero.

Definition at line 124 of file AdamID.hpp.

Referenced by mic::neural_nets::optimization::AdamID< eT >::calculateUpdate().

---

The documentation for this class was generated from the following file:

• src/optimization/AdamID.hpp