1893 lines
50 KiB
C++
1893 lines
50 KiB
C++
// Copyright (C) 2018-2019 Intel Corporation
|
|
// SPDX-License-Identifier: Apache-2.0
|
|
//
|
|
|
|
/**
|
|
* @brief a header file for internal Layers structure to describe layers information
|
|
* @file ie_layers.h
|
|
*/
|
|
#pragma once
|
|
|
|
#include <memory>
|
|
#include <string>
|
|
#include <vector>
|
|
#include <algorithm>
|
|
#include <map>
|
|
#include <iterator>
|
|
#include <limits>
|
|
#include <cctype>
|
|
#include "ie_common.h"
|
|
#include "ie_data.h"
|
|
#include "ie_blob.h"
|
|
#include "ie_device.hpp"
|
|
#include "ie_layers_property.hpp"
|
|
|
|
namespace InferenceEngine {
|
|
/**
|
|
* @brief This is an internal common Layer parameter parsing arguments
|
|
*/
|
|
struct LayerParams {
|
|
/// @brief Layer name
|
|
std::string name;
|
|
/// @brief Layer type
|
|
std::string type;
|
|
/// @brief Layer precision
|
|
Precision precision;
|
|
};
|
|
|
|
/**
|
|
* @brief This is a base abstraction Layer - all DNN Layers inherit from this class
|
|
*/
|
|
class CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief A shared pointer to CNNLayer
|
|
*/
|
|
using Ptr = std::shared_ptr<CNNLayer>;
|
|
|
|
/**
|
|
* @brief Layer name
|
|
*/
|
|
std::string name;
|
|
/**
|
|
* @brief Layer type
|
|
*/
|
|
std::string type;
|
|
/**
|
|
* @brief Layer base operating precision
|
|
*/
|
|
Precision precision;
|
|
/**
|
|
* @brief A vector of pointers to the output data elements of this layer in the di-graph (order matters)
|
|
*/
|
|
std::vector<DataPtr> outData;
|
|
/**
|
|
* @brief A vector of weak pointers to the input data elements of this layer in the di-graph (order matters)
|
|
*/
|
|
std::vector<DataWeakPtr> insData;
|
|
/**
|
|
* @brief If suggested to fuse - a pointer to the layer which needs to be fused with this layer
|
|
*/
|
|
Ptr _fusedWith;
|
|
/**
|
|
* @brief Convenience user values to store in this object as extra data
|
|
*/
|
|
UserValue userValue;
|
|
/**
|
|
* @brief Layer affinity set by user.
|
|
*/
|
|
std::string affinity;
|
|
|
|
/**
|
|
* @brief A constructor. Creates a new CNNLayer instance and initializes layer parameters with the given values.
|
|
* @param prms Basic common parsing parameters
|
|
*/
|
|
explicit CNNLayer(const LayerParams &prms) : name(prms.name), type(prms.type),
|
|
precision(prms.precision), userValue({0}) {
|
|
}
|
|
|
|
/**
|
|
* @brief A virtual destructor
|
|
*/
|
|
virtual ~CNNLayer() = default;
|
|
|
|
/**
|
|
* @brief Sets a layer to be fused with
|
|
* @param layer Reference to the layer to be fused with
|
|
*/
|
|
void fuse(Ptr &layer) {
|
|
_fusedWith = layer;
|
|
}
|
|
|
|
/**
|
|
* @brief Returns the first element of the input data for this layer
|
|
* @return A smart pointer to the input data element
|
|
*/
|
|
virtual const DataPtr input() const {
|
|
if (insData.empty()) {
|
|
THROW_IE_EXCEPTION << "Internal error: input data is empty";
|
|
}
|
|
auto lockedFirstInsData = insData[0].lock();
|
|
if (!lockedFirstInsData) {
|
|
THROW_IE_EXCEPTION << "Internal error: unable to lock weak_ptr\n";
|
|
}
|
|
return lockedFirstInsData;
|
|
}
|
|
|
|
/**
|
|
* @brief Checks if the input data and layer data are legitimate
|
|
*/
|
|
INFERENCE_ENGINE_API_CPP(void) validateLayer();
|
|
|
|
/**
|
|
* @brief Parse string with float in accordance with IE rules
|
|
* @param str input string with float value
|
|
* @return float value if parsing was successful
|
|
* @throws InferenceEngineException in case of parsing error
|
|
*/
|
|
static float ie_parse_float(const std::string &str) {
|
|
if (str == "-inf") {
|
|
return -std::numeric_limits<float>::infinity();
|
|
} else if (str == "inf") {
|
|
return std::numeric_limits<float>::infinity();
|
|
} else {
|
|
float res;
|
|
std::stringstream val_stream(str);
|
|
val_stream.imbue(std::locale("C"));
|
|
val_stream >> res;
|
|
if (!val_stream.eof()) THROW_IE_EXCEPTION;
|
|
return res;
|
|
}
|
|
}
|
|
/**
|
|
* @brief serialize float with c_locale formating
|
|
* used for default values serializing
|
|
*/
|
|
static std::string ie_serialize_float(float value) {
|
|
std::stringstream val_stream;
|
|
val_stream.imbue(std::locale("C"));
|
|
val_stream << value;
|
|
return val_stream.str();
|
|
}
|
|
|
|
/**
|
|
* @brief Gets float value for the given parameter
|
|
* @param param name of the parameter to find
|
|
* @param def default value of the parameter if not found
|
|
* @return float value
|
|
*/
|
|
float GetParamAsFloat(const char* param, float def) const {
|
|
std::string val = GetParamAsString(param, ie_serialize_float(def).c_str());
|
|
try {
|
|
return ie_parse_float(val);
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << "Cannot parse parameter " << param << " from IR for layer " << name
|
|
<< ". Value " << val << " cannot be casted to float.";
|
|
}
|
|
}
|
|
|
|
/**
|
|
* @brief Returns a float value for the given layer parameter
|
|
* @param param Name of the layer parameter
|
|
* @return A float value for the specified parameter
|
|
*/
|
|
float GetParamAsFloat(const char *param) const {
|
|
std::string val = GetParamAsString(param);
|
|
try {
|
|
return ie_parse_float(val);
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << "Cannot parse parameter " << param << " from IR for layer " << name
|
|
<< ". Value " << val << " cannot be casted to float.";
|
|
}
|
|
}
|
|
|
|
/**
|
|
* @brief Returns a vector of float values for the given parameter or returns the default value
|
|
* @param param Name of the layer parameter
|
|
* @param def Default value of the parameter if not found
|
|
* @return vector of float values
|
|
*/
|
|
std::vector<float> GetParamAsFloats(const char *param, std::vector<float> def) const {
|
|
std::string vals = GetParamAsString(param, "");
|
|
std::vector<float> result;
|
|
std::istringstream stream(vals);
|
|
std::string str;
|
|
if (vals.empty())
|
|
return def;
|
|
while (getline(stream, str, ',')) {
|
|
try {
|
|
float val = ie_parse_float(str);
|
|
result.push_back(val);
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << "Cannot parse parameter " << param << " " << str << " from IR for layer " << name
|
|
<< ". Value " << vals << " cannot be casted to floats.";
|
|
}
|
|
}
|
|
return result;
|
|
}
|
|
|
|
/**
|
|
* @brief Returns a vector of float values for the given parameter
|
|
* @param param Name of the layer parameter
|
|
* @return vector of float values
|
|
*/
|
|
std::vector<float> GetParamAsFloats(const char *param) const {
|
|
std::string vals = GetParamAsString(param);
|
|
std::vector<float> result;
|
|
std::istringstream stream(vals);
|
|
std::string str;
|
|
while (getline(stream, str, ',')) {
|
|
try {
|
|
float val = ie_parse_float(str);
|
|
result.push_back(val);
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << "Cannot parse parameter " << param << " " << str << " from IR for layer " << name
|
|
<< ". Value " << vals << " cannot be casted to floats.";
|
|
}
|
|
}
|
|
return result;
|
|
}
|
|
|
|
/**
|
|
* @brief Returns an integer value for the given parameter or returns the default value
|
|
* @param param Name of the layer parameter
|
|
* @param def Default value of the parameter if not found
|
|
* @return An int value for the specified parameter
|
|
*/
|
|
int GetParamAsInt(const char *param, int def) const {
|
|
std::string val = GetParamAsString(param, std::to_string(def).c_str());
|
|
try {
|
|
return std::stoi(val);
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << "Cannot parse parameter " << param << " from IR for layer " << name
|
|
<< ". Value " << val << " cannot be casted to int.";
|
|
}
|
|
}
|
|
|
|
/**
|
|
* @brief Returns an integer value for the given parameter
|
|
* @param param Name of the layer parameter
|
|
* @return An int value for the specified parameter
|
|
*/
|
|
int GetParamAsInt(const char *param) const {
|
|
std::string val = GetParamAsString(param);
|
|
try {
|
|
return std::stoi(val);
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << "Cannot parse parameter " << param << " from IR for layer " << name
|
|
<< ". Value " << val << " cannot be casted to int.";
|
|
}
|
|
}
|
|
|
|
|
|
/**
|
|
* @brief Returns a vector of int values for the given parameter or returns the default value
|
|
* @param param Name of the layer parameter
|
|
* @param def Default value of the parameter if not found
|
|
* @return vector of int values
|
|
*/
|
|
std::vector<int> GetParamAsInts(const char *param, std::vector<int> def) const {
|
|
std::string vals = GetParamAsString(param, "");
|
|
std::vector<int> result;
|
|
std::istringstream stream(vals);
|
|
std::string str;
|
|
if (vals.empty())
|
|
return def;
|
|
while (getline(stream, str, ',')) {
|
|
try {
|
|
result.push_back(std::stoi(str));
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << "Cannot parse parameter " << param << " " << str << " from IR for layer " << name
|
|
<< ". Value " << vals << " cannot be casted to int.";
|
|
}
|
|
}
|
|
return result;
|
|
}
|
|
|
|
/**
|
|
* @brief Returns a vector of int values for the given parameter
|
|
* @param param Name of the layer parameter
|
|
* @return vector of int values
|
|
*/
|
|
std::vector<int> GetParamAsInts(const char *param) const {
|
|
std::string vals = GetParamAsString(param);
|
|
std::vector<int> result;
|
|
std::istringstream stream(vals);
|
|
std::string str;
|
|
while (getline(stream, str, ',')) {
|
|
try {
|
|
result.push_back(std::stoi(str));
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << "Cannot parse parameter " << param << " " << str << " from IR for layer " << name
|
|
<< ". Value " << vals << " cannot be casted to int.";
|
|
}
|
|
}
|
|
return result;
|
|
}
|
|
/**
|
|
* @brief Returns an unsigned integer value for the given parameter or returns the default value
|
|
* @param param Name of the layer parameter
|
|
* @param def Default value of the parameter if not found
|
|
* @return An unsigned integer value for the specified parameter
|
|
*/
|
|
unsigned int GetParamAsUInt(const char *param, unsigned int def) const {
|
|
std::string val = GetParamAsString(param, std::to_string(def).c_str());
|
|
std::string message = "Cannot parse parameter " + std::string(param) + " from IR for layer " + name
|
|
+ ". Value " + val + " cannot be casted to int.";
|
|
try {
|
|
int value = std::stoi(val);
|
|
if (value < 0) {
|
|
THROW_IE_EXCEPTION << message;
|
|
}
|
|
return static_cast<unsigned int>(value);
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << message;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* @brief Returns an unsigned integer value for the given parameter
|
|
* @param param Name of the layer parameter
|
|
* @return An unsigned integer value for the specified parameter
|
|
*/
|
|
unsigned int GetParamAsUInt(const char *param) const {
|
|
std::string val = GetParamAsString(param);
|
|
std::string message = "Cannot parse parameter " + std::string(param) + " from IR for layer " + name
|
|
+ ". Value " + val + " cannot be casted to unsigned int.";
|
|
try {
|
|
int value = std::stoi(val);
|
|
if (value < 0) {
|
|
THROW_IE_EXCEPTION << message;
|
|
}
|
|
return static_cast<unsigned int>(value);
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << message;
|
|
}
|
|
}
|
|
|
|
|
|
/**
|
|
* @brief Returns a vector of unsigned int values for the given parameter or returns the default value
|
|
* @param param Name of the layer parameter
|
|
* @param def Default value of the parameter if not found
|
|
* @return vector of unsigned int values
|
|
*/
|
|
std::vector<unsigned int> GetParamAsUInts(const char *param, std::vector<unsigned int> def) const {
|
|
std::string vals = GetParamAsString(param, "");
|
|
std::vector<unsigned int> result;
|
|
std::istringstream stream(vals);
|
|
std::string str;
|
|
std::string message = "Cannot parse parameter " + std::string(param) + " " + str + " from IR for layer " + name
|
|
+ ". Value " + vals + " cannot be casted to unsigned int.";
|
|
if (vals.empty())
|
|
return def;
|
|
while (getline(stream, str, ',')) {
|
|
try {
|
|
int value = std::stoi(str);
|
|
if (value < 0) {
|
|
THROW_IE_EXCEPTION << message;
|
|
}
|
|
result.push_back(static_cast<unsigned int>(value));
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << message;
|
|
}
|
|
}
|
|
return result;
|
|
}
|
|
|
|
/**
|
|
* @brief Returns a vector of unsigned int values for the given parameter
|
|
* @param param Name of the layer parameter
|
|
* @return vector of unsigned int values
|
|
*/
|
|
std::vector<unsigned int> GetParamAsUInts(const char *param) const {
|
|
std::string vals = GetParamAsString(param);
|
|
std::vector<unsigned int> result;
|
|
std::istringstream stream(vals);
|
|
std::string str;
|
|
std::string message = "Cannot parse parameter " + std::string(param) + " " + str + " from IR for layer " + name
|
|
+ ". Value " + vals + " cannot be casted to int.";
|
|
while (getline(stream, str, ',')) {
|
|
try {
|
|
int value = std::stoi(str);
|
|
if (value < 0) {
|
|
THROW_IE_EXCEPTION << message;
|
|
}
|
|
result.push_back(static_cast<unsigned int>(value));
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << message;
|
|
}
|
|
}
|
|
return result;
|
|
}
|
|
/**
|
|
* @brief Returns a boolean value for the given parameter.
|
|
* The valid values are (true, false, 1, 0).
|
|
* @param param Name of the layer parameter
|
|
* @param def Default value of the parameter if not found
|
|
* @return A bool value for the specified parameter
|
|
*/
|
|
bool GetParamAsBool(const char *param, bool def) const {
|
|
std::string val = GetParamAsString(param, std::to_string(def).c_str());
|
|
std::string loweredCaseValue;
|
|
std::transform(val.begin(), val.end(), std::back_inserter(loweredCaseValue), [](char value) {
|
|
return std::tolower(value);
|
|
});
|
|
|
|
bool result = false;
|
|
|
|
if (!(std::istringstream(loweredCaseValue) >> std::boolalpha >> result)) {
|
|
// attempting parse using non alpha bool
|
|
return (GetParamAsInt(param, def) != 0);
|
|
}
|
|
|
|
return result;
|
|
}
|
|
/**
|
|
* @brief Returns a boolean value for the given parameter
|
|
* @param param Name of the layer parameter
|
|
* @return A bool value for the specified parameter
|
|
*/
|
|
bool GetParamAsBool(const char *param) const {
|
|
std::string val = GetParamAsString(param);
|
|
std::string loweredCaseValue;
|
|
std::transform(val.begin(), val.end(), std::back_inserter(loweredCaseValue), [](char value) {
|
|
return std::tolower(value);
|
|
});
|
|
|
|
bool result = false;
|
|
|
|
if (!(std::istringstream(loweredCaseValue) >> std::boolalpha >> result)) {
|
|
// attempting parse using non alpha bool
|
|
return (GetParamAsInt(param) != 0);
|
|
}
|
|
|
|
return result;
|
|
}
|
|
|
|
/**
|
|
* @deprecated Use GetParamAsBool function for that functionality
|
|
*/
|
|
INFERENCE_ENGINE_DEPRECATED
|
|
bool GetParamsAsBool(const char *param, bool def) const {
|
|
return GetParamAsBool(param, def);
|
|
}
|
|
|
|
/**
|
|
* @brief Returns a string value for the given parameter or returns the default one
|
|
* @param param Name of the layer parameter
|
|
* @param def Default value of the parameter if not found
|
|
* @return A string value
|
|
*/
|
|
std::string GetParamAsString(const char *param, const char *def) const {
|
|
auto it = params.find(param);
|
|
if (it == params.end() || it->second.empty()) {
|
|
return def;
|
|
}
|
|
return (*it).second;
|
|
}
|
|
|
|
/**
|
|
* @brief Checks the param presence in the layer
|
|
* @param param Name of the layer parameter
|
|
* @return a bool depending param presence
|
|
*/
|
|
bool CheckParamPresence(const char *param) const {
|
|
auto it = params.find(param);
|
|
if (it == params.end()) {
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
/**
|
|
* @brief Returns a string value for the given parameter.
|
|
* Throws exception if parameter was not found.
|
|
* @param param Name of the layer parameter
|
|
* @return A string value
|
|
*/
|
|
std::string GetParamAsString(const char *param) const {
|
|
auto it = params.find(param);
|
|
if (it == params.end()) {
|
|
THROW_IE_EXCEPTION << "No such parameter name '" << param << "' for layer " << name;
|
|
}
|
|
return (*it).second;
|
|
}
|
|
|
|
std::vector<std::string> GetParamAsStrings(const char *param, std::vector<std::string> def) const {
|
|
std::string vals = GetParamAsString(param, "");
|
|
std::vector<std::string> result;
|
|
std::istringstream stream(vals);
|
|
std::string str;
|
|
if (vals.empty())
|
|
return def;
|
|
while (getline(stream, str, ',')) {
|
|
try {
|
|
result.push_back(str);
|
|
} catch (...) {
|
|
THROW_IE_EXCEPTION << "Cannot parse parameter " << param << " from IR for layer " << name << ".";
|
|
}
|
|
}
|
|
return result;
|
|
}
|
|
|
|
/**
|
|
* @brief Map of pairs: (parameter name, parameter value)
|
|
*/
|
|
std::map<std::string, std::string> params;
|
|
|
|
/**
|
|
* @brief Map of pairs: (name, weights/biases blob)
|
|
*/
|
|
std::map<std::string, Blob::Ptr> blobs;
|
|
};
|
|
|
|
/**
|
|
* @brief Alias for CNNLayer object
|
|
*/
|
|
using GenericLayer = class CNNLayer;
|
|
|
|
/**
|
|
* @brief This class represents a layer with Weights and/or Biases (e.g. Convolution/Fully Connected, etc.)
|
|
*/
|
|
class WeightableLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief A default constructor. Constructs a WeightableLayer instance and initiates layer parameters with the given values
|
|
* @param prms Initial layer parameters
|
|
*/
|
|
explicit WeightableLayer(const LayerParams &prms) : CNNLayer(prms) {}
|
|
|
|
/**
|
|
* @brief A pointer to a weights blob
|
|
*/
|
|
Blob::Ptr _weights;
|
|
/**
|
|
* @brief A pointer to a biases blob
|
|
*/
|
|
Blob::Ptr _biases;
|
|
|
|
/**
|
|
* @brief Constructs a WeightableLayer instance and initiates layer parameters with the given values
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief convinenent way to declare property with backward compatibility to 2D members
|
|
*/
|
|
#define DEFINE_PROP(prop_name) \
|
|
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);\
|
|
|
|
/**
|
|
* @brief This class represents a standard 3D Convolution Layer
|
|
*/
|
|
class ConvolutionLayer : public WeightableLayer {
|
|
public:
|
|
/**
|
|
* @brief A convolution kernel array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_kernel);
|
|
/**
|
|
* @brief A convolution paddings begin array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_padding);
|
|
/**
|
|
* @brief A convolution paddings end array [X, Y, Z, ...]
|
|
*/
|
|
PropertyVector<unsigned int> _pads_end;
|
|
/**
|
|
* @brief A convolution strides array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_stride);
|
|
/**
|
|
* @brief A convolution dilations array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_dilation);
|
|
/**
|
|
* @brief A number of output feature maps (size) generating the 3'rd output dimension
|
|
*/
|
|
unsigned int _out_depth = 0u;
|
|
/**
|
|
* @brief Number of groups
|
|
*/
|
|
unsigned int _group = 1u;
|
|
/**
|
|
* @brief Auto padding type
|
|
*/
|
|
std::string _auto_pad;
|
|
|
|
/**
|
|
* @brief Creates a new ConvolutionLayer instance.
|
|
*/
|
|
explicit ConvolutionLayer(const LayerParams &p) : WeightableLayer(p),
|
|
_kernel(2, 0u), _padding(2, 0u), _stride(2, 1u), _dilation(2, 1u) {}
|
|
/**
|
|
* @brief assignment operator
|
|
*/
|
|
ConvolutionLayer & operator = (const ConvolutionLayer & that) {
|
|
if (&that != this) {
|
|
WeightableLayer::operator=(that);
|
|
_kernel = that._kernel;
|
|
_padding = that._padding;
|
|
_pads_end = that._pads_end;
|
|
_stride = that._stride;
|
|
_dilation = that._dilation;
|
|
_out_depth = that._out_depth;
|
|
_group = that._group;
|
|
}
|
|
return *this;
|
|
}
|
|
/**
|
|
* @brief copy constructor
|
|
*/
|
|
ConvolutionLayer(const ConvolutionLayer & that) : WeightableLayer(that) {
|
|
operator = (that);
|
|
}
|
|
/**
|
|
* @brief move constructor
|
|
*/
|
|
ConvolutionLayer(ConvolutionLayer &&) = default;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard deconvolution layer
|
|
*/
|
|
class DeconvolutionLayer : public ConvolutionLayer {
|
|
public:
|
|
using ConvolutionLayer::ConvolutionLayer;
|
|
using ConvolutionLayer::operator=;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard deformable convolution layer
|
|
*/
|
|
class DeformableConvolutionLayer : public ConvolutionLayer {
|
|
public:
|
|
using ConvolutionLayer::ConvolutionLayer;
|
|
using ConvolutionLayer::operator=;
|
|
|
|
/**
|
|
* @brief Number of deformable groups
|
|
*/
|
|
unsigned int _deformable_group = 1u;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard pooling layer
|
|
*/
|
|
class PoolingLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief Pooling kernel array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_kernel);
|
|
/**
|
|
* @brief Pooling paddings begin array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_padding);
|
|
/**
|
|
* @brief Pooling paddings end array [X, Y, Z, ...]
|
|
*/
|
|
PropertyVector<unsigned int> _pads_end;
|
|
/**
|
|
* @brief Pooling strides array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_stride);
|
|
|
|
/**
|
|
* @enum PoolType
|
|
* @brief Defines available pooling types
|
|
*/
|
|
enum PoolType {
|
|
MAX = 1,
|
|
AVG = 2,
|
|
STOCH = 3,
|
|
ROI = 4,
|
|
SPACIAL_PYRAMID = 5
|
|
};
|
|
|
|
/**
|
|
* @brief A pooling type
|
|
*/
|
|
PoolType _type = MAX;
|
|
|
|
/**
|
|
* @brief A flag that indicates if padding is excluded or not
|
|
*/
|
|
bool _exclude_pad = false;
|
|
/**
|
|
* @brief Auto padding type
|
|
*/
|
|
std::string _auto_pad;
|
|
|
|
/**
|
|
* @brief Creates a new PoolingLayer instance.
|
|
*/
|
|
explicit PoolingLayer(const LayerParams &p) : CNNLayer(p),
|
|
_kernel(2, 0u), _padding(2, 0u), _stride(2, 0u) {}
|
|
|
|
/**
|
|
* @brief assignment operator
|
|
*/
|
|
PoolingLayer & operator = (const PoolingLayer & that) {
|
|
if (&that != this) {
|
|
CNNLayer::operator=(that);
|
|
_kernel = that._kernel;
|
|
_padding = that._padding;
|
|
_pads_end = that._pads_end;
|
|
_stride = that._stride;
|
|
_type = that._type;
|
|
_exclude_pad = that._exclude_pad;
|
|
}
|
|
return *this;
|
|
}
|
|
/**
|
|
* @brief copy constructor
|
|
*/
|
|
PoolingLayer(const PoolingLayer & that) : CNNLayer(that) {
|
|
operator=(that);
|
|
}
|
|
|
|
/**
|
|
* @brief move constructor
|
|
*/
|
|
PoolingLayer(PoolingLayer &&) = default;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard binary convolution layer
|
|
*/
|
|
class BinaryConvolutionLayer : public WeightableLayer {
|
|
public:
|
|
/**
|
|
* @enum eBinaryConvolutionMode
|
|
* @brief Defines possible modes of binary convolution operation
|
|
*/
|
|
enum eBinaryConvolutionMode {
|
|
xnor_popcount = 0
|
|
};
|
|
|
|
/**
|
|
* @brief Mode of binary convolution operation
|
|
*/
|
|
eBinaryConvolutionMode _mode = xnor_popcount;
|
|
|
|
/**
|
|
* @brief A number of input feature maps (size) generating the 3'rd input dimension
|
|
*/
|
|
unsigned int _in_depth = 0u;
|
|
|
|
/**
|
|
* @brief A pad value which is used to fill pad area
|
|
*/
|
|
float _pad_value = 0.0f;
|
|
|
|
/**
|
|
* @brief A convolution kernel array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_kernel);
|
|
/**
|
|
* @brief A convolution paddings begin array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_padding);
|
|
/**
|
|
* @brief A convolution paddings end array [X, Y, Z, ...]
|
|
*/
|
|
PropertyVector<unsigned int> _pads_end;
|
|
/**
|
|
* @brief A convolution strides array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_stride);
|
|
/**
|
|
* @brief A convolution dilations array [X, Y, Z, ...]
|
|
*/
|
|
DEFINE_PROP(_dilation);
|
|
/**
|
|
* @brief A number of output feature maps (size) generating the 3'rd output dimension
|
|
*/
|
|
unsigned int _out_depth = 0u;
|
|
/**
|
|
* @brief Number of groups
|
|
*/
|
|
unsigned int _group = 1u;
|
|
/**
|
|
* @brief Auto padding type
|
|
*/
|
|
std::string _auto_pad;
|
|
|
|
/**
|
|
* @brief Creates a new BinaryConvolutionLayer instance.
|
|
*/
|
|
explicit BinaryConvolutionLayer(const LayerParams &p) : WeightableLayer(p),
|
|
_kernel(2, 0u), _padding(2, 0u), _stride(2, 1u), _dilation(2, 1u) {}
|
|
/**
|
|
* @brief assignment operator
|
|
*/
|
|
BinaryConvolutionLayer & operator = (const BinaryConvolutionLayer & that) {
|
|
if (&that != this) {
|
|
WeightableLayer::operator=(that);
|
|
_kernel = that._kernel;
|
|
_padding = that._padding;
|
|
_pads_end = that._pads_end;
|
|
_stride = that._stride;
|
|
_dilation = that._dilation;
|
|
_out_depth = that._out_depth;
|
|
_group = that._group;
|
|
_mode = that._mode;
|
|
_in_depth = that._in_depth;
|
|
_pad_value = that._pad_value;
|
|
}
|
|
return *this;
|
|
}
|
|
/**
|
|
* @brief copy constructor
|
|
*/
|
|
BinaryConvolutionLayer(const BinaryConvolutionLayer & that) : WeightableLayer(that) {
|
|
operator = (that);
|
|
}
|
|
/**
|
|
* @brief move constructor
|
|
*/
|
|
BinaryConvolutionLayer(BinaryConvolutionLayer &&) = default;
|
|
};
|
|
|
|
#undef DEFINE_PROP
|
|
|
|
/**
|
|
* @brief This class represents a fully connected layer
|
|
*/
|
|
class FullyConnectedLayer : public WeightableLayer {
|
|
public:
|
|
/**
|
|
* @brief A size of output
|
|
*/
|
|
unsigned int _out_num = 0;
|
|
|
|
/**
|
|
* @brief Creates a new FullyConnectedLayer instance and initializes layer parameters with the given values.
|
|
*/
|
|
using WeightableLayer::WeightableLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents concatenation layer
|
|
* Takes as input several data elements and merges them to one using the supplied axis
|
|
*/
|
|
class ConcatLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief An axis on which concatenation operation is performed
|
|
*/
|
|
unsigned int _axis = 1;
|
|
|
|
/**
|
|
* @brief Creates a new ConcatLayer instance and initializes layer parameters with the given values.
|
|
* If batch is used, then batch needs to be specified as an input dimension also
|
|
* In current implementation 1 means channels, 0 - batch
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a layer that evenly splits the input into the supplied outputs
|
|
*/
|
|
class SplitLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief An axis on which split operation is performed
|
|
*/
|
|
unsigned int _axis = 1;
|
|
|
|
/**
|
|
* @brief Creates a new SplitLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a Linear Response Normalization (LRN) Layer
|
|
*/
|
|
class NormLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief Response size
|
|
*/
|
|
unsigned int _size = 0;
|
|
/**
|
|
* @brief K
|
|
*/
|
|
unsigned int _k = 1;
|
|
/**
|
|
* @brief Alpha coefficient
|
|
*/
|
|
float _alpha = 0;
|
|
/**
|
|
* @brief Beta coefficient
|
|
*/
|
|
float _beta = 0;
|
|
/**
|
|
* @brief Flag to specify normalization across feature maps (true) or across channels
|
|
*/
|
|
bool _isAcrossMaps = false;
|
|
|
|
/**
|
|
* @brief Creates a new NormLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents standard softmax Layer
|
|
*/
|
|
class SoftMaxLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief Axis number for a softmax operation
|
|
*/
|
|
int axis = 1;
|
|
/**
|
|
* @brief Creates a new SoftMaxLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @class GRNLayer
|
|
* @brief This class represents standard GRN Layer
|
|
*/
|
|
class GRNLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief A default constructor. Creates a new GRNLayer instance and initializes layer parameters with the given values.
|
|
* @param prms Initial layer parameters
|
|
*/
|
|
explicit GRNLayer(const LayerParams &prms) : CNNLayer(prms), bias(0.f) {}
|
|
|
|
/**
|
|
* @brief Bias for squares sum
|
|
*/
|
|
float bias = 0.f;
|
|
};
|
|
|
|
/**
|
|
* @class MVNLayer
|
|
* @brief This class represents standard MVN Layer
|
|
*/
|
|
class MVNLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief A default constructor. Creates a new MVNLayer instance and initializes layer parameters with the given values.
|
|
* @param prms Initial layer parameters
|
|
*/
|
|
explicit MVNLayer(const LayerParams &prms) : CNNLayer(prms), across_channels(0), normalize(1) {}
|
|
|
|
/**
|
|
* @brief Indicate that mean value is calculated across channels
|
|
*/
|
|
int across_channels;
|
|
|
|
/**
|
|
* @brief Indicate that the result needs to be normalized
|
|
*/
|
|
int normalize = 1;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a Rectified Linear activation layer
|
|
*/
|
|
class ReLULayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief Negative slope is used to takle negative inputs instead of setting them to 0
|
|
*/
|
|
float negative_slope = 0.0f;
|
|
|
|
/**
|
|
* @brief Creates a new ReLULayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a Clamp activation layer
|
|
* Clamps all tensor elements into the range [min_value, max_value]
|
|
*/
|
|
class ClampLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief A minimum value
|
|
*/
|
|
float min_value = 0.0f;
|
|
|
|
/**
|
|
* @brief A maximum value
|
|
*/
|
|
float max_value = 1.0f;
|
|
/**
|
|
* @brief Creates a new ClampLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a ReLU6 activation layer
|
|
* Clamps all tensor elements into the range [0, 6.0]
|
|
*/
|
|
class ReLU6Layer : public ClampLayer {
|
|
public:
|
|
explicit ReLU6Layer(const LayerParams &prms) : ClampLayer(prms) {
|
|
max_value = 6.0f;
|
|
}
|
|
|
|
using ClampLayer::ClampLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents an element wise operation layer
|
|
*/
|
|
class EltwiseLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @enum eOperation
|
|
* @brief Defines possible operations that can be used
|
|
*/
|
|
enum eOperation {
|
|
Sum = 0, Prod, Max, Sub, Min, Div, Squared_diff, Floor_mod, Pow,
|
|
Equal, Not_equal, Less, Less_equal, Greater, Greater_equal,
|
|
Logical_AND, Logical_OR, Logical_XOR, Logical_NOT, Mean
|
|
};
|
|
|
|
/**
|
|
* @brief A type of the operation to use
|
|
*/
|
|
eOperation _operation = Sum;
|
|
|
|
/**
|
|
* @brief A vector of coefficients to scale the operands
|
|
*/
|
|
std::vector<float> coeff;
|
|
|
|
/**
|
|
* @brief Creates a new EltwiseLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard crop layer
|
|
*/
|
|
class CropLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief A vector of dimensions for cropping
|
|
*/
|
|
std::vector<int> axis;
|
|
/**
|
|
* @brief A vector of dimensions to be preserved
|
|
*/
|
|
std::vector<int> dim;
|
|
/**
|
|
* @brief A vector of offsets for each dimension
|
|
*/
|
|
std::vector<int> offset;
|
|
|
|
/**
|
|
* @brief Creates a new CropLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard reshape layer
|
|
*/
|
|
class ReshapeLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief A vector of sizes of the shape
|
|
*/
|
|
std::vector<int> shape;
|
|
/**
|
|
* @brief A number of axis to be taken for a reshape
|
|
*/
|
|
int axis = 0;
|
|
/**
|
|
* @brief A number of first axises to be taken for a reshape
|
|
*/
|
|
int num_axes = -1;
|
|
|
|
/**
|
|
* @brief Creates a new ReshapeLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard Tile Layer
|
|
*/
|
|
class TileLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief An index of the axis to tile
|
|
*/
|
|
int axis = -1;
|
|
/**
|
|
* @brief A number of copies to be made
|
|
*/
|
|
int tiles = -1;
|
|
|
|
/**
|
|
* @brief Creates a new TileLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a Layer which performs Scale and Shift
|
|
*/
|
|
class ScaleShiftLayer : public WeightableLayer {
|
|
public:
|
|
/**
|
|
* @brief A flag that indicates if the same value is used for all the features. If false, the value is used pixel wise
|
|
*/
|
|
unsigned int _broadcast = 0;
|
|
|
|
/**
|
|
* @brief Creates a new ScaleShiftLayer instance.
|
|
*/
|
|
using WeightableLayer::WeightableLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents TensorIterator layer
|
|
*/
|
|
class TensorIterator : public CNNLayer {
|
|
public:
|
|
struct PortMap {
|
|
// Data map rule
|
|
int from; /**< Index of exteral data from ins/outs fields of CNNLayer */
|
|
int to; /**< Index of internal data in iterator body */
|
|
|
|
// Iteration rule
|
|
int axis; /**< Axis to iterate throught */
|
|
int stride; /**< Stride to iterate throught */
|
|
int start; /**< Start index of iteration range */
|
|
int end; /**< Last index of iteration range */
|
|
int part_size; /**< Part size which will be transfered to body subnetwork */
|
|
};
|
|
|
|
struct Body {
|
|
std::vector<DataPtr> inputs;
|
|
std::vector<DataPtr> outputs;
|
|
};
|
|
|
|
std::vector<PortMap> input_port_map;
|
|
std::vector<PortMap> output_port_map;
|
|
std::vector<PortMap> back_edges;
|
|
|
|
Body body;
|
|
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief Base class for recurrent cell layers
|
|
*/
|
|
class RNNCellBase : public WeightableLayer {
|
|
public:
|
|
using WeightableLayer::WeightableLayer;
|
|
|
|
/**
|
|
* @brief Direct type of recurrent cell (including subtypes)
|
|
* Description of particular cell semantics is in LSTMCell, GRUCell, RNNCell.
|
|
*/
|
|
enum CellType {
|
|
LSTM, /**< Original LSTM cell */
|
|
GRU, /**< Original GRU cell */
|
|
RNN, /**< Original RNN cell */
|
|
GRU_LBR, /**< GRU cell modification. "Linear before reset" */
|
|
};
|
|
|
|
/** @copybrief CellType */
|
|
CellType cellType = LSTM;
|
|
|
|
/**
|
|
* @brief Size of hidden state data
|
|
*
|
|
* In case of batch output state tensor will have shape [N, hidden_size]
|
|
*/
|
|
int hidden_size = 0;
|
|
|
|
/**
|
|
* @brief Clip data into range [-clip, clip] on input of activations
|
|
*
|
|
* clip==0.0f means no clipping
|
|
*/
|
|
float clip = 0.0f;
|
|
/**
|
|
* @brief Activations used inside recurrent cell
|
|
*
|
|
* Valid values: sigmoid, tanh, relu
|
|
*/
|
|
std::vector<std::string> activations;
|
|
|
|
/**
|
|
* @brief Alpha parameters of activations
|
|
*
|
|
* Respective to activation list.
|
|
*/
|
|
std::vector<float> activation_alpha;
|
|
|
|
/**
|
|
* @brief Beta parameters of activations
|
|
*
|
|
* Respective to activation list.
|
|
*/
|
|
std::vector<float> activation_beta;
|
|
};
|
|
|
|
/**
|
|
* @brief 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)
|
|
*/
|
|
class LSTMCell : public RNNCellBase {
|
|
public:
|
|
using RNNCellBase::RNNCellBase;
|
|
using RNNCellBase::operator=;
|
|
};
|
|
|
|
/**
|
|
* @brief 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
|
|
*/
|
|
class GRUCell : public RNNCellBase {
|
|
public:
|
|
using RNNCellBase::RNNCellBase;
|
|
using RNNCellBase::operator=;
|
|
};
|
|
|
|
/**
|
|
* @brief 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)
|
|
*/
|
|
class RNNCell : public RNNCellBase {
|
|
public:
|
|
using RNNCellBase::RNNCellBase;
|
|
using RNNCellBase::operator=;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief Sequence of recurrent cells
|
|
*
|
|
* N - batch size
|
|
* T - sequence size
|
|
* S - state size (=hidden_size)
|
|
* NS - num of state tensors (LSTM=2, GRU/RNN=1)
|
|
* ND - num of direction (BDR=2, WFD/BWD=1)
|
|
*
|
|
* Inputs:
|
|
* [N,T,D] Xt - input data
|
|
* [ND,N,S] Ht-1 - initial hidden state
|
|
* [ND,N,S] Ct-1 - initial cell state // if NS==2
|
|
* [N] SL - sequence length (optional input)
|
|
*
|
|
* Outputs:
|
|
* [ND,N,T,S] Xt - input data
|
|
* [ND,N,S] Ht-1 - initial hidden state
|
|
* [ND,N,S] Ct-1 - initial cell state // if NS==2
|
|
*
|
|
* NB! if axis==0 batch and sequense dimensions are swapped (N <-> T) for input and output tensors
|
|
*
|
|
* Weights:
|
|
* - weights [ND,G,S,D+S]
|
|
* - biases [ND,G,S]
|
|
* NB! if ND==2 weights are concatenated cell weights [forward_cell_weights, backward_cell_weights]
|
|
*
|
|
*/
|
|
class RNNSequenceLayer : public RNNCellBase {
|
|
public:
|
|
using RNNCellBase::RNNCellBase;
|
|
|
|
/**
|
|
* @brief An axis by which iteration is performed
|
|
* axis=0 means first input/output data blob dimension is sequence
|
|
* axis=1 means first input/output data blob dimension is batch
|
|
*/
|
|
unsigned int axis = 1;
|
|
|
|
/**
|
|
* @brief Direction of iteration through sequence dimension
|
|
*/
|
|
enum Direction {
|
|
FWD, /**< Forward mode. Iterate starts from index 0 with step 1. */
|
|
BWD, /**< Backward mode. Iterate starts from last index with step -1. */
|
|
BDR /**< Bidirectional mode. First is forward pass, second is backward. */
|
|
};
|
|
|
|
/** @copybrief Direction */
|
|
Direction direction = FWD;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a Layer which performs Scale and Shift
|
|
*/
|
|
class PReLULayer : public WeightableLayer {
|
|
public:
|
|
/**
|
|
* @brief A flag that indicates if the same negative_slope value is used for all the features. If false, the value is used pixel wise
|
|
*/
|
|
bool _channel_shared;
|
|
|
|
public:
|
|
/**
|
|
* @brief A default constructor. Creates a new PReLULayer instance and initializes layer parameters with the given values.
|
|
* @param prms Initial layer parameters
|
|
*/
|
|
explicit PReLULayer(const LayerParams &prms) : WeightableLayer(prms), _channel_shared(false) {}
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard Power Layer
|
|
* Formula is: output = (offset + scale * input) ^ power
|
|
*/
|
|
class PowerLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief An exponent value
|
|
*/
|
|
float power = 1.f;
|
|
/**
|
|
* @brief A scale factor
|
|
*/
|
|
float scale = 1.f;
|
|
/**
|
|
* @brief An offset value
|
|
*/
|
|
float offset = 0.f;
|
|
|
|
/**
|
|
* @brief Creates a new PowerLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a Batch Normalization Layer
|
|
*/
|
|
class BatchNormalizationLayer : public WeightableLayer {
|
|
public:
|
|
/**
|
|
* @brief A small value to add to the variance estimate to avoid division by zero
|
|
*/
|
|
float epsilon = 1e-3f;
|
|
|
|
/**
|
|
* @brief Creates a new BatchNormalizationLayer instance.
|
|
*/
|
|
using WeightableLayer::WeightableLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a general matrix multiplication operation layer
|
|
* Formula is: dst := alpha*src1*src2 + beta*src3
|
|
*/
|
|
class GemmLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief A scale factor of src1 matrix
|
|
*/
|
|
float alpha = 1.f;
|
|
/**
|
|
* @brief A scale factor of src3 matrix
|
|
*/
|
|
float beta = 1.f;
|
|
/**
|
|
* @brief A flag that indicates if the src1 matrix is to be transposed
|
|
*/
|
|
bool transpose_a = false;
|
|
/**
|
|
* @brief A flag that indicates if the src2 matrix is to be transposed
|
|
*/
|
|
bool transpose_b = false;
|
|
/**
|
|
* @brief Creates a new GemmLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard Pad layer
|
|
* Adds paddings to input tensor
|
|
*/
|
|
class PadLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @enum ePadMode
|
|
* @brief Defines possible modes of pad operation
|
|
*/
|
|
enum ePadMode {
|
|
Constant = 0, Edge, Reflect, Symmetric
|
|
};
|
|
|
|
/**
|
|
* @brief Size of padding in the beginning of each axis
|
|
*/
|
|
PropertyVector<unsigned int> pads_begin;
|
|
/**
|
|
* @brief Size of padding in the end of each axis
|
|
*/
|
|
PropertyVector<unsigned int> pads_end;
|
|
/**
|
|
* @brief Mode of pad operation
|
|
*/
|
|
ePadMode pad_mode = Constant;
|
|
/**
|
|
* @brief A pad value which is used for filling in Constant mode
|
|
*/
|
|
float pad_value = 0.0f;
|
|
/**
|
|
* @brief Creates a new PadLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard Gather layer
|
|
* Gather slices from Dictionary according to Indexes
|
|
*/
|
|
class GatherLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief The axis in Dictionary to gather Indexes from
|
|
*/
|
|
int axis = 0;
|
|
/**
|
|
* @brief Creates a new GatherLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard Strided Slice layer
|
|
* Strided Slice picks from input tensor according parameters
|
|
*/
|
|
class StridedSliceLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief 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 begin_mask;
|
|
/**
|
|
* @brief Analogous to begin_mask
|
|
*/
|
|
std::string end_mask;
|
|
/**
|
|
* @brief The ellipsis_mask is a bitmask where bit i being 1 means
|
|
* the i-th is actually an ellipsis
|
|
*/
|
|
std::string ellipsis_mask;
|
|
/**
|
|
* @brief 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 new_axis_mask;
|
|
/**
|
|
* @brief The shrink_axis_mask is a bitmask where bit i being 1 means
|
|
* the i-th position shrinks the dimensionality
|
|
*/
|
|
std::string shrink_axis_mask;
|
|
|
|
/**
|
|
* @brief Creates a new StridedSliceLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a standard Shuffle Channels layer
|
|
* Shuffle Channels picks from input tensor according parameters
|
|
*/
|
|
class ShuffleChannelsLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief The axis in tensor to shuffle channels
|
|
*/
|
|
int axis = 1;
|
|
|
|
/**
|
|
* @brief The group of output shuffled channels
|
|
*/
|
|
unsigned int group = 1;
|
|
|
|
/**
|
|
* @brief Creates a new ShuffleChannelsLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard Depth To Space layer
|
|
* Depth To Space picks from input tensor according parameters
|
|
*/
|
|
class DepthToSpaceLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief The group of output shuffled channels
|
|
*/
|
|
unsigned int block_size = 1;
|
|
|
|
/**
|
|
* @brief Creates a new DepthToSpaceLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard Space To Depth layer
|
|
* Depth To Space picks from input tensor according parameters
|
|
*/
|
|
class SpaceToDepthLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief The group of output Space To Depth
|
|
*/
|
|
unsigned int block_size = 1;
|
|
|
|
/**
|
|
* @brief Creates a new SpaceToDepthLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents SparseFillEmptyRows layer
|
|
* SparseFillEmptyRows fills empty rows in a sparse tensor
|
|
*/
|
|
class SparseFillEmptyRowsLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief Creates a new SparseFillEmptyRowsLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard Reverse Sequence layer
|
|
* Reverse Sequence modifies input tensor according parameters
|
|
*/
|
|
class ReverseSequenceLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief The seq_axis dimension in tensor which is partially reversed
|
|
*/
|
|
int seq_axis = 1;
|
|
|
|
/**
|
|
* @brief The batch_axis dimension in tensor along which reversal is performed
|
|
*/
|
|
int batch_axis = 0;
|
|
|
|
/**
|
|
* @brief Creates a new ReverseSequence instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a OneHot layer
|
|
* Converts input into OneHot representation.
|
|
*/
|
|
class OneHotLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief A depth of representation
|
|
*/
|
|
unsigned int depth = 0;
|
|
|
|
/**
|
|
* @brief The locations represented by indices in input take value on_value
|
|
*/
|
|
float on_value = 1.f;
|
|
|
|
/**
|
|
* @brief The locations not represented by indices in input take value off_value
|
|
*/
|
|
float off_value = 0.f;
|
|
|
|
/**
|
|
* @brief Define the shape of output tensor
|
|
*/
|
|
int axis = -1;
|
|
|
|
/**
|
|
* @brief Creates a new OneHot instance
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard RangeLayer layer
|
|
* RangeLayer modifies input tensor dimensions according parameters
|
|
*/
|
|
class RangeLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief Creates a new RangeLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard Fill layer
|
|
* RFill modifies input tensor according parameters
|
|
*/
|
|
class FillLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief Creates a new Fill instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a SelectLayer layer
|
|
* SelectLayer layer takes elements from the second (“then”) or the third (“else”) input based on condition mask (“cond”) provided in the first input.
|
|
* The “cond” tensor is broadcasted to “then” and “else” tensors.
|
|
* The output tensor shape is equal to broadcasted shape of “cond”, “then” and “else”.
|
|
*/
|
|
class SelectLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief Creates a new SelectLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard Broadcast layer
|
|
* Broadcast modifies input tensor dimensions according parameters
|
|
*/
|
|
class BroadcastLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief Creates a new Broadcast instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
/**
|
|
* @brief This class represents a quantization operation layer
|
|
* Element-wise linear quantization of floating point input values into a descrete set of floating point values
|
|
*/
|
|
class QuantizeLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief The number of quantization levels
|
|
*/
|
|
int levels = 1;
|
|
|
|
/**
|
|
* @brief Creates a new QuantizeLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard Math layers
|
|
* Math modifies input tensor dimensions according parameters
|
|
*/
|
|
class MathLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief Creates a new Math instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard Reduce layers
|
|
* Reduce modifies input tensor according parameters
|
|
*/
|
|
class ReduceLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief The keep_dims dimension in tensor which is partially reversed
|
|
*/
|
|
bool keep_dims = true;
|
|
|
|
/**
|
|
* @brief Creates a new Reduce instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard TopK layer
|
|
* TopK picks top K values from input tensor according parameters
|
|
*/
|
|
class TopKLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief The mode could be 'max' or 'min'
|
|
*/
|
|
std::string mode;
|
|
/**
|
|
* @brief top K values sort mode could be 'value' or 'index'
|
|
*/
|
|
std::string sort;
|
|
/**
|
|
* @brief The axis dimension in tensor which is top K values are picked
|
|
*/
|
|
int axis = -1;
|
|
|
|
/**
|
|
* @brief Creates a new TopKLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents Unique layer.
|
|
* The Unique operation searches for unique elements in 1-D input
|
|
*/
|
|
class UniqueLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief A flag indicating whether to sort unique elements
|
|
*/
|
|
bool sorted;
|
|
/**
|
|
* @brief A flag indicating whether to return indices of input data elements in the output of uniques
|
|
*/
|
|
bool return_inverse;
|
|
/**
|
|
* @brief A flag indicating whether to return a number of occurences for each unique element
|
|
*/
|
|
bool return_counts;
|
|
|
|
/**
|
|
* @brief Creates a new UniqueLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard NonMaxSuppression layer
|
|
*/
|
|
class NonMaxSuppressionLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief The 'center_point_box' indicates the format of the box data
|
|
*/
|
|
bool center_point_box = false;
|
|
/**
|
|
* @brief Creates a new NonMaxSuppressionLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
|
|
/**
|
|
* @brief This class represents a standard Scatter layer
|
|
*/
|
|
class ScatterLayer : public CNNLayer {
|
|
public:
|
|
/**
|
|
* @brief The axis in Dictionary to scatter Indexes from
|
|
*/
|
|
int axis = 0;
|
|
/**
|
|
* @brief Creates a new ScatterLayer instance.
|
|
*/
|
|
using CNNLayer::CNNLayer;
|
|
};
|
|
|
|
} // namespace InferenceEngine
|