ie_layers.h File Reference

a header file for internal Layers structure to describe layers information More...

#include <memory>
#include <string>
#include <vector>
#include <algorithm>
#include <map>
#include <iterator>
#include <cctype>
#include "ie_common.h"
#include "ie_data.h"
#include "ie_blob.h"
#include "ie_device.hpp"
#include "ie_layers_property.hpp"

Go to the source code of this file.

Data Structures
struct	InferenceEngine::LayerParams
	This is an internal common Layer parameter parsing arguments. More...
class	InferenceEngine::CNNLayer
	This is a base abstraction Layer - all DNN Layers inherit from this class. More...
class	InferenceEngine::WeightableLayer
	This class represents a layer with Weights and/or Biases (e.g. Convolution/Fully Connected, etc.) More...
class	InferenceEngine::ConvolutionLayer
	This class represents a standard 3D Convolution Layer. More...
class	InferenceEngine::DeconvolutionLayer
	This class represents a standard deconvolution layer. More...
class	InferenceEngine::PoolingLayer
	This class represents a standard pooling layer. More...
class	InferenceEngine::BinaryConvolutionLayer
	This class represents a standard binary convolution layer. More...
class	InferenceEngine::FullyConnectedLayer
	This class represents a fully connected layer. More...
class	InferenceEngine::ConcatLayer
	This class represents concatenation layer Takes as input several data elements and merges them to one using the supplied axis. More...
class	InferenceEngine::SplitLayer
	This class represents a layer that evenly splits the input into the supplied outputs. More...
class	InferenceEngine::NormLayer
	This class represents a Linear Response Normalization (LRN) Layer. More...
class	InferenceEngine::SoftMaxLayer
	This class represents standard softmax Layer. More...
class	InferenceEngine::GRNLayer
	This class represents standard GRN Layer. More...
class	InferenceEngine::MVNLayer
	This class represents standard MVN Layer. More...
class	InferenceEngine::ReLULayer
	This class represents a Rectified Linear activation layer. More...
class	InferenceEngine::ClampLayer
	This class represents a Clamp activation layer Clamps all tensor elements into the range [min_value, max_value]. More...
class	InferenceEngine::ReLU6Layer
	This class represents a ReLU6 activation layer Clamps all tensor elements into the range [0, 6.0]. More...
class	InferenceEngine::EltwiseLayer
	This class represents an element wise operation layer. More...
class	InferenceEngine::CropLayer
	This class represents a standard crop layer. More...
class	InferenceEngine::ReshapeLayer
	This class represents a standard reshape layer. More...
class	InferenceEngine::TileLayer
	This class represents a standard Tile Layer. More...
class	InferenceEngine::ScaleShiftLayer
	This class represents a Layer which performs Scale and Shift. More...
class	InferenceEngine::TensorIterator
	This class represents TensorIterator layer. More...
struct	InferenceEngine::TensorIterator::PortMap
struct	InferenceEngine::TensorIterator::Body
class	InferenceEngine::RNNCellBase
	Base class for recurrent cell layers. More...
class	InferenceEngine::RNNSequenceLayer
	Sequence of recurrent cells. More...
class	InferenceEngine::PReLULayer
	This class represents a Layer which performs Scale and Shift. More...
class	InferenceEngine::PowerLayer
	This class represents a standard Power Layer Formula is: output = (offset + scale * input) ^ power. More...
class	InferenceEngine::BatchNormalizationLayer
	This class represents a Batch Normalization Layer. More...
class	InferenceEngine::GemmLayer
	This class represents a general matrix multiplication operation layer Formula is: dst := alpha*src1*src2 + beta*src3. More...
class	InferenceEngine::PadLayer
	This class represents a standard Pad layer Adds paddings to input tensor. More...
class	InferenceEngine::GatherLayer
	This class represents a standard Gather layer Gather slices from Dictionary according to Indexes. More...
class	InferenceEngine::StridedSliceLayer
	This class represents a standard Strided Slice layer Strided Slice picks from input tensor according parameters. More...
class	InferenceEngine::ShuffleChannelsLayer
	This class represents a standard Shuffle Channels layer Shuffle Channels picks from input tensor according parameters. More...
class	InferenceEngine::DepthToSpaceLayer
	This class represents a standard Depth To Space layer Depth To Space picks from input tensor according parameters. More...
class	InferenceEngine::SpaceToDepthLayer
	This class represents a standard Space To Depth layer Depth To Space picks from input tensor according parameters. More...
class	InferenceEngine::ReverseSequenceLayer
	This class represents a standard Reverse Sequence layer Reverse Sequence modifies input tensor according parameters. More...
class	InferenceEngine::SqueezeLayer
	This class represents a standard Squeeze layer Squeeze modifies input tensor dimensions according parameters. More...
class	InferenceEngine::UnsqueezeLayer
	This class represents a standard Unsqueeze layer Unsqueeze modifies input tensor dimensions according parameters. More...
class	InferenceEngine::RangeLayer
	This class represents a standard RangeLayer layer RangeLayer modifies input tensor dimensions according parameters. More...
class	InferenceEngine::FillLayer
	This class represents a standard Fill layer RFill modifies input tensor according parameters. More...
class	InferenceEngine::ExpandLayer
	This class represents a standard Expand layer Expand modifies input tensor dimensions according parameters. More...
class	InferenceEngine::QuantizeLayer
	This class represents a quantization operation layer Element-wise linear quantization of floating point input values into a descrete set of floating point values. More...

Macros
#define	DEFINE_PROP(prop_name)
	convinenent way to declare property with backward compatibility to 2D members More...

Typedefs
using	InferenceEngine::GenericLayer = class CNNLayer
	Alias for CNNLayer object.
using	InferenceEngine::LSTMCell = RNNCellBase
	LSTM Cell layer. More...
using	InferenceEngine::GRUCell = RNNCellBase
	GRU Cell layer. More...
using	InferenceEngine::RNNCell = RNNCellBase
	RNN Cell layer. More...

Detailed Description

a header file for internal Layers structure to describe layers information

Macro Definition Documentation

DEFINE_PROP

#define DEFINE_PROP

(

prop_name

)

Value:

PropertyVector<unsigned int> prop_name;\
unsigned int &prop_name##_x = prop_name.at(X_AXIS);\
unsigned int &prop_name##_y = prop_name.at(Y_AXIS);\

convinenent way to declare property with backward compatibility to 2D members

Typedef Documentation

GRUCell

using InferenceEngine::GRUCell = typedef RNNCellBase

GRU Cell layer.

G - number of gates (=3) N - batch size S - state size (=hidden_size)

Inputs: [N,D] Xt - input data [N,S] Ht-1 - initial hidden state

Outputs: [N,S] Ht - out hidden state

Weights:

• weights [G,S,D+S]
• biases [G,S] NB! gates order is ZRH {update, reset, output}

activations is {_f, _g} default: {_f=sigm, _g=tanh}

Equations:

• - matrix mult (.) - eltwise mult [,] - concatenation

zt = _f(Wz*[Ht-1, Xt] + Bz)

• rt = _f(Wr*[Ht-1, Xt] + Br)
• ht = _g(Wh*[rt (.) Ht-1, Xt] + Bh)
• Ht = (1 - zt) (.) ht + zt (.) Ht-1

LSTMCell

using InferenceEngine::LSTMCell = typedef RNNCellBase

LSTM Cell layer.

G - number of gates (=4) N - batch size S - state size (=hidden_size)

Inputs: [N,D] Xt - input data [N,S] Ht-1 - initial hidden state [N,S] Ct-1 - initial cell state

Outputs: [N,S] Ht - out hidden state [N,S] Ct - out cell state

Weights:

• weights [G,S,D+S]
• biases [G,S] NB! gates order is FICO {forget, input, candidate, output}

activations is {_f, _g, _h} default: {_f=sigm, _g=tanh, _h=tanh}

Equations:

• - matrix mult (.) - eltwise mult [,] - concatenation

ft = _f(Wf*[Ht-1, Xt] + Bf)

• it = _f(Wi*[Ht-1, Xt] + Bi)
• ct = _g(Wc*[Ht-1, Xt] + Bc)
• ot = _f(Wo*[Ht-1, Xt] + Bo)
• Ct = ft (.) Ct-1 + it (.) ct
• Ht = ot (.) _h(Ct)

RNNCell

using InferenceEngine::RNNCell = typedef RNNCellBase

RNN Cell layer.

G - number of gates (=1) N - batch size S - state size (=hidden_size)

Inputs: [N,D] Xt - input data [N,S] Ht-1 - initial hidden state

Outputs: [N,S] Ht - out hidden state

Weights:

• weights [G,S,D+S]
• biases [G,S]

activations is {_f} default: {_f=tanh}

Equations:

• - matrix mult [,] - concatenation

Ht = _f(Wi*[Ht-1, Xt] + Bi)