Apache SINGA
A distributed deep learning platform .
 All Classes Namespaces Files Functions Variables Typedefs Macros
Public Member Functions | List of all members
singa::RBMVisLayer Class Reference

RBM visible layer. More...

#include <neuron_layer.h>

Inheritance diagram for singa::RBMVisLayer:
singa::RBMLayer singa::Layer

Public Member Functions

void Setup (const LayerProto &proto, int npartitions) override
 Setup layer properties. More...
 
void ComputeFeature (int flag, Metric *perf) override
 Compute features of this layer based on connected layers. More...
 
void ComputeGradient (int flag, Metric *perf) override
 Compute gradients for parameters and connected layers. More...
 
- Public Member Functions inherited from singa::RBMLayer
const Blob< float > & neg_data (const Layer *layer)
 
Blob< float > * mutable_neg_data (const Layer *layer)
 
const std::vector< Param * > GetParams () const override
 Layers that have paramters must override this function. More...
 
virtual Blob< float > * Sample (int flat)
 
- Public Member Functions inherited from singa::Layer
virtual ConnectionType src_neuron_connection (int k) const
 Return the connection type between one neuron of this layer and its source layer. More...
 
virtual ConnectionType dst_layer_connection () const
 Return the connection type of this layer and all dst layers. More...
 
virtual const std::string DebugString (int step, int flag)
 For print debug info about each layer, e.g., norm of feature vector, norm of parameters. More...
 
int partition_dim () const
 
int partition_id () const
 
int type () const
 
const std::string & name () const
 Return name of this layer.
 
virtual const Blob< float > & data (const Layer *from) const
 
virtual Blob< float > * mutable_data (const Layer *from)
 
virtual const Blob< float > & grad (const Layer *from) const
 
virtual Blob< float > * mutable_grad (const Layer *from)
 
const std::vector< Layer * > srclayers () const
 return LayerS that connected to this layer
 
const std::vector< Layer * > dstlayers () const
 return LayerS that this layer connected to
 
int srclayers_size () const
 
int dstlayers_size () const
 
void clear_dstlayers ()
 
void clear_srclayers ()
 
void add_srclayer (Layer *src)
 
void add_dstlayer (Layer *dst)
 

Additional Inherited Members

- Static Public Member Functions inherited from singa::Layer
static LayerCreate (const LayerProto &proto)
 
- Protected Attributes inherited from singa::RBMLayer
bool gaussian_
 if ture, sampling according to guassian distribution
 
int hdim_
 dimension of the hidden layer
 
int vdim_
 dimension of the visible layer
 
int batchsize_
 
bool first_gibbs_
 
Paramweight_
 
Parambias_
 
Blob< float > neg_data_
 
Blob< float > neg_sample_
 
Blob< float > sample_
 
- Protected Attributes inherited from singa::Layer
LayerProto layer_proto_
 
Blob< float > data_
 
Blob< float > grad_
 
std::vector< Layer * > srclayers_
 
std::vector< Layer * > dstlayers_
 

Detailed Description

RBM visible layer.

Member Function Documentation

void singa::RBMVisLayer::ComputeFeature ( int  flag,
Metric perf 
)
overridevirtual

Compute features of this layer based on connected layers.

Parameters
perfpointer to Metric obj for collect and aggregate performance

Implements singa::Layer.

void singa::RBMVisLayer::ComputeGradient ( int  flag,
Metric perf 
)
overridevirtual

Compute gradients for parameters and connected layers.

Parameters
flagused to get the calling phase, e.g., forward of training (kForward | kTrain)
flagused to get the calling phase, e.g., forward of training

Implements singa::Layer.

void singa::RBMVisLayer::Setup ( const LayerProto &  proto,
int  npartitions 
)
overridevirtual

Setup layer properties.

Setup the shapes for data and parameters, also setup some properties based on the layer configuration and connected layers.

Parameters
protolayer configuration.
npartitionsnum of total partitions of the original layer. This layer should be setup as one partition.

Reimplemented from singa::RBMLayer.


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