edge::Model class
#include <model.hpp>

A base class for machine learning models.

Reference

This class serves as a base class for machine learning models. It provides common functionality such as loading a model, accessing inputs and outputs, applying a delegate, and executing the model.

Derived classes

class ModelImpl final: Implementation of the Model interface for QNN models.
class ModelImpl final: Implementation of the Model interface for TensorFlow Lite models.

Constructors, destructors, conversion operators

Model(const std::filesystem::path& modelPath) explicit: Constructor for the Model class.
Model() defaulted
Model(const Model&) defaulted
Model(Model&&) deleted
~Model() defaulted virtual: Virtual destructor for the Model class.

Public functions

auto operator=(const Model&) -> Model & defaulted
auto operator=(Model&&) -> Model & deleted
auto loadModel(const std::filesystem::path& modelPath) -> STATUS pure virtual: Pure virtual function to load a model from a file path.
auto loadModel(const nonstd::span<uint8_t>& modelBuffer) -> STATUS pure virtual: Pure virtual function to load a model from a file buffer.
auto getNumInputs() const -> size_t: Get the number of input tensors in the model.
auto getNumOutputs() const -> size_t: Get the number of output tensors in the model.
auto getInput(size_t index) const -> std::shared_ptr< Tensor >: Get the input tensor at the specified index.
auto getOutput(size_t index) const -> std::shared_ptr< Tensor >: Get the output tensor at the specified index.
auto getInputs() -> std::vector< std::shared_ptr< Tensor >> &: Get the inputs of the model.
auto getOutputs() -> std::vector< std::shared_ptr< Tensor >> &: Get the outputs of the model.
auto getDelegate() const -> DELEGATE: Get the current delegate used for model execution.
auto applyDelegate(const DELEGATE& delegate) -> STATUS pure virtual: Apply a delegate for model execution.
auto execute() -> STATUS pure virtual: Execute the model.
auto name() const -> const std::string &: Get the name of the model.
auto getCreationStatus() const -> STATUS: Get the status of model creation.
auto getPrecision() const -> TensorType: Get the precision used for model execution.

Protected functions

void setDelegate(const DELEGATE& delegate): Set the delegate for model execution.
void setPrecision(const TensorType& precision): Set the precision for model execution.
void setCreationStatus(const STATUS& status): Set the status of model creation.

Function documentation

edge::Model::Model(const std::filesystem::path& modelPath) explicit

Constructor for the Model class.

Parameters
modelPath	The path to the model file.

This constructor initializes a Model object with the given model path.

auto edge::Model::loadModel(const std::filesystem::path& modelPath) -> STATUS pure virtual

Pure virtual function to load a model from a file path.

Parameters
modelPath	The path to the model file
Returns	STATUS The status of the model loading operation

This function is a pure virtual function that must be implemented by any derived classes. It is used to load a model from a file path.

auto edge::Model::loadModel(const nonstd::span<uint8_t>& modelBuffer) -> STATUS pure virtual

Pure virtual function to load a model from a file buffer.

Parameters
modelBuffer	The buffer containing the model
Returns	STATUS The status of the model loading operation

This function is a pure virtual function that must be implemented by any derived classes. It is used to load a model from a file buffer.

auto edge::Model::getNumInputs() const -> size_t

Get the number of input tensors in the model.

Returns	The number of input tensors

auto edge::Model::getNumOutputs() const -> size_t

Get the number of output tensors in the model.

Returns	The number of output tensors

auto edge::Model::getInput(size_t index) const -> std::shared_ptr< Tensor >

Get the input tensor at the specified index.

Parameters
index	The index of the input tensor
Returns	The input tensor at the specified index, or nullptr if index is out of bounds

auto edge::Model::getOutput(size_t index) const -> std::shared_ptr< Tensor >

Get the output tensor at the specified index.

Parameters
index	The index of the output tensor
Returns	The output tensor at the specified index, or nullptr if index is out of bounds

auto edge::Model::getInputs() -> std::vector< std::shared_ptr< Tensor >> &

Get the inputs of the model.

Returns	A reference to a vector of shared pointers to Tensor objects representing the inputs of the model.

This function returns a reference to a vector of shared pointers to Tensor objects, which represent the inputs of the model.

auto edge::Model::getOutputs() -> std::vector< std::shared_ptr< Tensor >> &

Get the outputs of the model.

Returns	A reference to a vector of shared pointers to Tensor objects representing the outputs of the model.

This function returns a reference to a vector of shared pointers to Tensor objects, which represent the outputs of the model.

auto edge::Model::getDelegate() const -> DELEGATE

Get the current delegate used for model execution.

Returns	The delegate currently set for model execution

auto edge::Model::applyDelegate(const DELEGATE& delegate) -> STATUS pure virtual

Apply a delegate for model execution.

Parameters
delegate	The delegate to apply
Returns	The status of the operation

auto edge::Model::execute() -> STATUS pure virtual

Execute the model.

Returns	The status of the operation

auto edge::Model::name() const -> const std::string &

Get the name of the model.

Returns	The name of the model

auto edge::Model::getCreationStatus() const -> STATUS

Get the status of model creation.

Returns	The status of model creation

Verify that the model was created successfully

auto edge::Model::getPrecision() const -> TensorType

Get the precision used for model execution.

Returns	The pricsion used model execution

void edge::Model::setDelegate(const DELEGATE& delegate) protected

Set the delegate for model execution.

Parameters
delegate	The delegate to set

This method is used by derivatives to allow users to query the currently set delegate

void edge::Model::setPrecision(const TensorType& precision) protected

Set the precision for model execution.

This method is used by derivatives to allow users to query the execution precision

void edge::Model::setCreationStatus(const STATUS& status) protected

Set the status of model creation.

Parameters
status	The status to set

This method is used by derivatives to allow querying of model creation status

edge::Model class #include <model.hpp>

Contents