fnn_layer_dropout Module Reference

Module dedicated to the class dropoutlayer. More...

Data Types
type	dropoutlayer
	Implements a dropout layer. More...

Functions/Subroutines
type(dropoutlayer) function, public	dropout_layer_fromfile (batch_size, unit_num)
	Constructor for class dropoutlayer from a file. More...
subroutine	dropout_tofile (self, unit_num)
	Implements dropoutlayer::tofile. More...
subroutine	dropout_apply_forward (self, train, member, x, y)
	Implements dropoutlayer::apply_forward. More...
subroutine	dropout_apply_tangent_linear (self, member, dp, dx, dy)
	Implements dropoutlayer::apply_tangent_linear. More...
subroutine	dropout_apply_adjoint (self, member, dy, dp, dx)
	Implements dropoutlayer::apply_adjoint. More...

Detailed Description

Module dedicated to the class dropoutlayer.

Function/Subroutine Documentation

dropout_apply_adjoint()

subroutine fnn_layer_dropout::dropout_apply_adjoint ( class(dropoutlayer), intent(in)  self, integer(ik), intent(in)  member, real(rk), dimension(:), intent(inout)  dy, real(rk), dimension(:), intent(out)  dp, real(rk), dimension(:), intent(out)  dx  )

private

Implements dropoutlayer::apply_adjoint.

Applies the adjoint of the layer.

The adjoint operator reads

Note

This method should only be called after dropoutlayer::apply_forward.

Since there is no (trainable) parameters, the parameter perturbation should be an empty array.

The intent of dy is declared inout instead of in because of other subclasses of fnn_layer::layer.

Parameters

[in,out]	self	The layer.
[in]	member	The index inside the batch.
[in,out]	dy	The output perturbation.
[out]	dp	The parameter perturbation.
[out]	dx	The state perturbation.

dropout_apply_forward()

subroutine fnn_layer_dropout::dropout_apply_forward ( class(dropoutlayer), intent(inout)  self, logical, intent(in)  train, integer(ik), intent(in)  member, real(rk), dimension(:), intent(in)  x, real(rk), dimension(:), intent(out)  y  )

private

Implements dropoutlayer::apply_forward.

Applies and linearises the layer.

The forward function reads

where is the element-wise multiplication for vectors.

In training mode, with probability and with probability where is the dropout rate.

In testing mode, .

Note

Input parameter member should be less than layer::batch_size.

The linearisation is trivial and only requires to store the values of in dropoutlayer::draws.

Parameters

[in,out]	self	The layer.
[in]	train	Whether the model is used in training mode.
[in]	member	The index inside the batch.
[in]	x	The input of the layer.
[out]	y	The output of the layer.

dropout_apply_tangent_linear()

subroutine fnn_layer_dropout::dropout_apply_tangent_linear ( class(dropoutlayer), intent(in)  self, integer(ik), intent(in)  member, real(rk), dimension(:), intent(in)  dp, real(rk), dimension(:), intent(in)  dx, real(rk), dimension(:), intent(out)  dy  )

private

Implements dropoutlayer::apply_tangent_linear.

Applies the TL of the layer.

The TL operator reads

Note

This method should only be called after dropoutlayer::apply_forward.

Since there is no (trainable) parameters, the parameter perturbation should be an empty array.

Parameters

[in]	self	The layer.
[in]	member	The index inside the batch.
[in]	dp	The parameter perturbation.
[in]	dx	The state perturbation.
[out]	dy	The output perturbation.

dropout_layer_fromfile()

type(dropoutlayer) function, public fnn_layer_dropout::dropout_layer_fromfile	(	integer(ik), intent(in)	batch_size,
		integer(ik), intent(in)	unit_num
	)

Constructor for class dropoutlayer from a file.

Parameters

[in]	batch_size	The value for layer::batch_size.
[in]	unit_num	The unit number for the read statements.

Returns

The constructed layer.

dropout_tofile()

subroutine fnn_layer_dropout::dropout_tofile ( class(dropoutlayer), intent(in)  self, integer(ik), intent(in)  unit_num  )

private

Implements dropoutlayer::tofile.

Saves the layer.

Parameters

[in]	self	The layer.
[in]	unit_num	The unit number for the write statement.