Configuration class for GELU module. More...

Inheritance diagram for Mila::Dnn::GeluConfig:

Collaboration diagram for Mila::Dnn::GeluConfig:

Public Types
enum class	ApproximationMethod { Exact , Tanh , Sigmoid }
	Approximation methods for the GELU activation function. More...

Public Member Functions
	GeluConfig ()=default
	Default constructor.

ApproximationMethod	getApproximationMethod () const
	Get the configured approximation method.

void	validate () const
	Validate configuration parameters.

GeluConfig &	withApproximationMethod (ApproximationMethod method)
	Configure the approximation method for GELU computation.

Public Member Functions inherited from Mila::Dnn::ComponentConfig
virtual	~ComponentConfig ()=default
	Virtual destructor to support proper polymorphic destruction.

const std::string &	getName () const
	Gets the configured component name.

ComputePrecision::Policy	getPrecision () const
	Gets the configured precision policy.

bool	isTraining () const
	Gets the configured training mode.

template<typename Self >
auto &	withName (this Self &&self, std::string name)
	Sets the name of the component with fluent interface.

template<typename Self >
auto &	withPrecision (this Self &&self, ComputePrecision::Policy policy)
	Sets the compute precision policy with fluent interface.

template<typename Self >
auto &	withTraining (this Self &&self, bool is_training)
	Sets the training mode with fluent interface.

Private Attributes
ApproximationMethod	approximation_method_ = ApproximationMethod::Tanh

Additional Inherited Members
Protected Attributes inherited from Mila::Dnn::ComponentConfig
bool	is_training_ = false
	Training mode flag, defaults to false (inference mode)

std::string	name_ = "unnamed"
	Component name, defaults to "unnamed" if not explicitly set.

ComputePrecision::Policy	precision_ = ComputePrecision::Policy::Auto
	Precision policy for computation, defaults to Auto.

Detailed Description

Configuration class for GELU module.

Provides a type-safe fluent interface for configuring GELU modules.

Member Enumeration Documentation

◆ ApproximationMethod

enum class Mila::Dnn::GeluConfig::ApproximationMethod

strong

Approximation methods for the GELU activation function.

Enumerator
Exact	Exact implementation using error function.
Tanh	Fast approximation using tanh.
Sigmoid	Fast approximation using sigmoid.

Constructor & Destructor Documentation

◆ GeluConfig()

Mila::Dnn::GeluConfig::GeluConfig ( )

default

Default constructor.

Member Function Documentation

◆ getApproximationMethod()

ApproximationMethod Mila::Dnn::GeluConfig::getApproximationMethod ( ) const

inline

Get the configured approximation method.

Returns: ApproximationMethod The approximation method

Here is the caller graph for this function:

◆ validate()

void Mila::Dnn::GeluConfig::validate ( ) const

inlinevirtual

Validate configuration parameters.

Currently, only the Tanh approximation method is supported for GELU computation. Setting other approximation methods will cause validation to fail.

Exceptions

std::invalid_argument If validation fails or an unsupported approximation method is selected

Reimplemented from Mila::Dnn::ComponentConfig.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ withApproximationMethod()

GeluConfig & Mila::Dnn::GeluConfig::withApproximationMethod ( ApproximationMethod method )

inline

Configure the approximation method for GELU computation.

Note: Currently, only the Tanh approximation method is supported. Setting other methods will cause validation to fail when the configuration is used.

Parameters

method The approximation method to use (only ApproximationMethod::Tanh is currently supported)

Returns: GeluConfig& Reference to this for method chaining

Member Data Documentation

◆ approximation_method_

ApproximationMethod Mila::Dnn::GeluConfig::approximation_method_ = ApproximationMethod::Tanh

private

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

/home/runner/work/Mila/Mila/Mila/Src/Dnn/Modules/Activations/GeluConfig.ixx

Public Types

Public Member Functions

Private Attributes

Additional Inherited Members