a header file for internal Layers structure to describe layers information More...
#include <algorithm>
#include <cctype>
#include <iterator>
#include <limits>
#include <map>
#include <memory>
#include <string>
#include <vector>
#include "ie_blob.h"
#include "ie_common.h"
#include "ie_data.h"
#include "ie_layers_property.hpp"
Go to the source code of this file.
Data Structures | |
struct | PortMap |
struct | Body |
Describes a tensor iterator body. 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. | |
Enumerations | |
enum | PoolType |
Defines available pooling types. | |
enum | eBinaryConvolutionMode |
Defines possible modes of binary convolution operation. | |
enum | eOperation |
Defines possible operations that can be used. | |
enum | CellType |
Direct type of recurrent cell (including subtypes) Description of particular cell semantics is in LSTMCell, GRUCell, RNNCell. | |
enum | Direction |
Direction of iteration through sequence dimension. | |
enum | ePadMode |
Defines possible modes of pad operation. | |
Variables | |
struct { | |
std::string InferenceEngine::name | |
Layer name. | |
std::string InferenceEngine::type | |
Layer type. | |
Precision InferenceEngine::precision | |
Layer precision. | |
}; | |
This is an internal common Layer parameter parsing arguments. More... | |
class { | |
using | Ptr = std::shared_ptr< CNNLayer > |
A shared pointer to CNNLayer. | |
std::string InferenceEngine::name | |
Layer name. | |
std::string InferenceEngine::type | |
Layer type. | |
Precision InferenceEngine::precision | |
Layer base operating precision. | |
std::vector< DataPtr > InferenceEngine::outData | |
A vector of pointers to the output data elements of this layer in the di-graph (order matters) | |
std::vector< DataWeakPtr > InferenceEngine::insData | |
A vector of weak pointers to the input data elements of this layer in the di-graph (order matters) | |
Ptr InferenceEngine::_fusedWith | |
If suggested to fuse - a pointer to the layer which needs to be fused with this layer. | |
UserValue InferenceEngine::userValue | |
Convenience user values to store in this object as extra data. | |
std::string InferenceEngine::affinity | |
Layer affinity set by user. | |
std::map< std::string, std::string > InferenceEngine::params | |
Map of pairs: (parameter name, parameter value) | |
std::map< std::string, Blob::Ptr > InferenceEngine::blobs | |
Map of pairs: (name, weights/biases blob) | |
std::shared_ptr< ngraph::Node > node | |
}; | |
This is a base abstraction Layer - all DNN Layers inherit from this class. More... | |
Blob::Ptr | InferenceEngine::_weights |
A pointer to a weights blob. | |
Blob::Ptr | InferenceEngine::_biases |
A pointer to a biases blob. | |
PropertyVector< unsigned int > | InferenceEngine::_kernel |
A convolution kernel array [X, Y, Z, ...]. More... | |
unsigned int & | InferenceEngine::_kernel_x = _kernel .at(X_AXIS) |
unsigned int & | InferenceEngine::_kernel_y = _kernel .at(Y_AXIS) |
PropertyVector< unsigned int > | InferenceEngine::_padding |
A convolution paddings begin array [X, Y, Z, ...]. More... | |
unsigned int & | InferenceEngine::_padding_x = _padding .at(X_AXIS) |
unsigned int & | InferenceEngine::_padding_y = _padding .at(Y_AXIS) |
PropertyVector< unsigned int > | InferenceEngine::_pads_end |
A convolution paddings end array [X, Y, Z, ...]. More... | |
PropertyVector< unsigned int > | InferenceEngine::_stride |
A convolution strides array [X, Y, Z, ...]. More... | |
unsigned int & | InferenceEngine::_stride_x = _stride .at(X_AXIS) |
unsigned int & | InferenceEngine::_stride_y = _stride .at(Y_AXIS) |
PropertyVector< unsigned int > | InferenceEngine::_dilation |
A convolution dilations array [X, Y, Z, ...]. | |
unsigned int & | InferenceEngine::_dilation_x = _dilation .at(X_AXIS) |
unsigned int & | InferenceEngine::_dilation_y = _dilation .at(Y_AXIS) |
unsigned int | InferenceEngine::_out_depth = 0u |
A number of output feature maps (size) generating the 3'rd output dimension. | |
unsigned int | InferenceEngine::_group = 1u |
Number of groups. | |
std::string | InferenceEngine::_auto_pad |
Auto padding type. | |
unsigned int | InferenceEngine::_deformable_group = 1u |
Number of deformable groups. | |
PoolType | InferenceEngine::_type = MAX |
A pooling type. | |
bool | InferenceEngine::_exclude_pad = false |
A flag that indicates if padding is excluded or not. | |
eBinaryConvolutionMode | InferenceEngine::_mode = xnor_popcount |
Mode of binary convolution operation. | |
unsigned int | InferenceEngine::_in_depth = 0u |
A number of input feature maps (size) generating the 3'rd input dimension. | |
float | InferenceEngine::_pad_value = 0.0f |
A pad value which is used to fill pad area. | |
unsigned int | InferenceEngine::_out_num = 0 |
A size of output. | |
unsigned int | InferenceEngine::_axis = 1 |
An axis on which concatenation operation is performed. More... | |
unsigned int | InferenceEngine::_size = 0 |
Response size. | |
unsigned int | InferenceEngine::_k = 1 |
K. | |
float | InferenceEngine::_alpha = 0 |
Alpha coefficient. | |
float | InferenceEngine::_beta = 0 |
Beta coefficient. | |
bool | InferenceEngine::_isAcrossMaps = false |
Flag to specify normalization across feature maps (true) or across channels. | |
int | InferenceEngine::axis = 1 |
Axis number for a softmax operation. More... | |
float | InferenceEngine::bias = 0.f |
Bias for squares sum. | |
int | InferenceEngine::across_channels = 0 |
Indicate that mean value is calculated across channels. | |
int | InferenceEngine::normalize = 1 |
Indicate that the result needs to be normalized. | |
float | InferenceEngine::negative_slope = 0.0f |
Negative slope is used to takle negative inputs instead of setting them to 0. | |
float | InferenceEngine::min_value = 0.0f |
A minimum value. | |
float | InferenceEngine::max_value = 1.0f |
A maximum value. | |
eOperation | InferenceEngine::_operation = Sum |
A type of the operation to use. | |
std::vector< float > | InferenceEngine::coeff |
A vector of coefficients to scale the operands. | |
std::vector< int > | InferenceEngine::dim |
A vector of dimensions to be preserved. | |
std::vector< int > | InferenceEngine::offset = 0.f |
A vector of offsets for each dimension. More... | |
std::vector< int > | InferenceEngine::shape |
A vector of sizes of the shape. | |
int | InferenceEngine::num_axes = -1 |
A number of first axises to be taken for a reshape. | |
int | InferenceEngine::tiles = -1 |
A number of copies to be made. | |
unsigned int | InferenceEngine::_broadcast = 0 |
A flag that indicates if the same value is used for all the features. If false, the value is used pixel wise. | |
std::vector< PortMap > | InferenceEngine::input_port_map |
Input ports map. | |
std::vector< PortMap > | InferenceEngine::output_port_map |
Output ports map. | |
std::vector< PortMap > | InferenceEngine::back_edges |
Back edges map. | |
Body | InferenceEngine::body |
A Tensor Iterator body. | |
CellType | InferenceEngine::cellType = LSTM |
Direct type of recurrent cell (including subtypes) Description of particular cell semantics is in LSTMCell, GRUCell, RNNCell. | |
int | InferenceEngine::hidden_size = 0 |
Size of hidden state data. More... | |
float | InferenceEngine::clip = 0.0f |
Clip data into range [-clip, clip] on input of activations. More... | |
std::vector< std::string > | InferenceEngine::activations |
Activations used inside recurrent cell. More... | |
std::vector< float > | InferenceEngine::activation_alpha |
Alpha parameters of activations. More... | |
std::vector< float > | InferenceEngine::activation_beta |
Beta parameters of activations. More... | |
Direction | InferenceEngine::direction = FWD |
Direction of iteration through sequence dimension. | |
bool | InferenceEngine::_channel_shared = false |
A flag that indicates if the same negative_slope value is used for all the features. If false, the value is used pixel wise. | |
float | InferenceEngine::power = 1.f |
An exponent value. | |
float | InferenceEngine::scale = 1.f |
A scale factor. | |
float | InferenceEngine::epsilon = 1e-3f |
A small value to add to the variance estimate to avoid division by zero. | |
float | InferenceEngine::alpha = 1.f |
A scale factor of src1 matrix. | |
float | InferenceEngine::beta = 1.f |
A scale factor of src3 matrix. | |
bool | InferenceEngine::transpose_a = false |
A flag that indicates if the src1 matrix is to be transposed. | |
bool | InferenceEngine::transpose_b = false |
A flag that indicates if the src2 matrix is to be transposed. | |
PropertyVector< unsigned int > | InferenceEngine::pads_begin |
Size of padding in the beginning of each axis. | |
PropertyVector< unsigned int > | InferenceEngine::pads_end |
Size of padding in the end of each axis. | |
ePadMode | InferenceEngine::pad_mode = Constant |
Mode of pad operation. | |
float | InferenceEngine::pad_value = 0.0f |
A pad value which is used for filling in Constant mode. | |
std::string | InferenceEngine::begin_mask |
The begin_mask is a bitmask where bit i being 0 means to ignore the begin value and instead use the default value. | |
std::string | InferenceEngine::end_mask |
Analogous to begin_mask. | |
std::string | InferenceEngine::ellipsis_mask |
The ellipsis_mask is a bitmask where bit i being 1 means the i-th is actually an ellipsis. | |
std::string | InferenceEngine::new_axis_mask |
The new_axis_mask_ is a bitmask where bit i being 1 means the i-th position creates a new 1 dimension shape. | |
std::string | InferenceEngine::shrink_axis_mask |
The shrink_axis_mask is a bitmask where bit i being 1 means the i-th position shrinks the dimensionality. | |
unsigned int | InferenceEngine::group = 1 |
The group of output shuffled channels. | |
unsigned int | InferenceEngine::block_size = 1 |
The group of output shuffled channels. More... | |
std::vector< size_t > | InferenceEngine::_block_shape |
Spatial dimensions blocks sizes. | |
std::vector< size_t > | InferenceEngine::_pads_begin |
Size of padding in the beginning of each axis. | |
std::vector< size_t > | InferenceEngine::_crops_begin |
It specifies how many elements to crop from the intermediate result across the spatial dimensions. | |
std::vector< size_t > | InferenceEngine::_crops_end |
It specifies how many elements to crop from the intermediate result across the spatial dimensions. | |
bool | InferenceEngine::with_right_bound = true |
Indicates whether the intervals include the right or the left bucket edge. | |
int | InferenceEngine::seq_axis = 1 |
The seq_axis dimension in tensor which is partially reversed. | |
int | InferenceEngine::batch_axis = 0 |
The batch_axis dimension in tensor along which reversal is performed. | |
unsigned int | InferenceEngine::depth = 0 |
A depth of representation. | |
float | InferenceEngine::on_value = 1.f |
The locations represented by indices in input take value on_value. | |
float | InferenceEngine::off_value = 0.f |
The locations not represented by indices in input take value off_value. | |
int | InferenceEngine::levels = 1 |
The number of quantization levels. | |
bool | InferenceEngine::keep_dims = true |
The keep_dims dimension in tensor which is partially reversed. | |
std::string | InferenceEngine::mode |
The mode could be 'max' or 'min'. | |
std::string | InferenceEngine::sort |
top K values sort mode could be 'value' or 'index' | |
bool | InferenceEngine::sorted |
A flag indicating whether to sort unique elements. | |
bool | InferenceEngine::return_inverse |
A flag indicating whether to return indices of input data elements in the output of uniques. | |
bool | InferenceEngine::return_counts |
A flag indicating whether to return a number of occurences for each unique element. | |
bool | InferenceEngine::center_point_box = false |
The 'center_point_box' indicates the format of the box data. | |
bool | InferenceEngine::sort_result_descending = true |
The 'sort_result_descending' indicates that result will sort descending by score through all batches and classes. | |
int | InferenceEngine::flatten = 1 |
flatten value | |
int | InferenceEngine::grid_w = 0 |
Value of grid width. | |
int | InferenceEngine::grid_h = 0 |
Value of grid height. | |
float | InferenceEngine::stride_w = 0.f |
Value of width step between grid cells. | |
float | InferenceEngine::stride_h = 0.f |
Value of height step between grid cells. | |
int | InferenceEngine::max_rois = 0 |
The maximum number of output rois. | |
float | InferenceEngine::min_size = 0.f |
Minimium width and height for boxes. | |
float | InferenceEngine::nms_threshold = 0.7f |
Non max suppression threshold. | |
int | InferenceEngine::pre_nms_topn = 1000 |
Maximum number of anchors selected before nms. | |
int | InferenceEngine::post_nms_topn = 1000 |
Maximum number of anchors selected after nms. | |
a header file for internal Layers structure to describe layers information
#define DEFINE_PROP | ( | prop_name | ) |
convinenent way to declare property with backward compatibility to 2D members
struct { ... } |
This is an internal common Layer parameter parsing arguments.
class { ... } |
This is a base abstraction Layer - all DNN Layers inherit from this class.
unsigned int InferenceEngine::_axis = 1 |
An axis on which concatenation operation is performed.
An axis on which split operation is performed.
PropertyVector<unsigned int> InferenceEngine::_kernel |
A convolution kernel array [X, Y, Z, ...].
Pooling kernel array [X, Y, Z, ...].
PropertyVector<unsigned int> InferenceEngine::_padding |
A convolution paddings begin array [X, Y, Z, ...].
Pooling paddings begin array [X, Y, Z, ...].
std::vector<size_t> InferenceEngine::_pads_end |
A convolution paddings end array [X, Y, Z, ...].
Size of padding in the end of each axis.
Pooling paddings end array [X, Y, Z, ...].
PropertyVector<unsigned int> InferenceEngine::_stride |
A convolution strides array [X, Y, Z, ...].
Pooling strides array [X, Y, Z, ...].
std::vector<float> InferenceEngine::activation_alpha |
Alpha parameters of activations.
Respective to activation list.
std::vector<float> InferenceEngine::activation_beta |
Beta parameters of activations.
Respective to activation list.
std::vector<std::string> InferenceEngine::activations |
Activations used inside recurrent cell.
Valid values: sigmoid, tanh, relu
int InferenceEngine::axis = 1 |
Axis number for a softmax operation.
The axis dimension in tensor which is top K values are picked.
Define the shape of output tensor.
The axis in tensor to shuffle channels.
The axis in Dictionary to gather Indexes from.
An axis by which iteration is performed.
An index of the axis to tile.
A number of axis to be taken for a reshape.
A vector of dimensions for cropping.
axis=0 means first input/output data blob dimension is sequence axis=1 means first input/output data blob dimension is batch
unsigned int InferenceEngine::block_size = 1 |
The group of output shuffled channels.
The group of output Space To Depth.
float InferenceEngine::clip = 0.0f |
Clip data into range [-clip, clip] on input of activations.
clip==0.0f means no clipping
int InferenceEngine::hidden_size = 0 |
Size of hidden state data.
In case of batch output state tensor will have shape [N, hidden_size]
float InferenceEngine::offset = 0.f |
A vector of offsets for each dimension.
An offset value.
Precision InferenceEngine::precision |
Layer precision.
Layer base operating precision.