Mila::Dnn::LayerNormConfig Class Reference

Configuration class for Layer Normalization module. More...

Inheritance diagram for Mila::Dnn::LayerNormConfig:

Collaboration diagram for Mila::Dnn::LayerNormConfig:

Public Member Functions
	LayerNormConfig ()=default
	Default constructor.
	LayerNormConfig (size_t normalized_dim)
	Constructor with normalized dimension size.
int64_t	getAxis () const
	Get the configured normalization axis.
float	getEpsilon () const
	Get the configured epsilon value.
const std::vector< size_t > &	getInputShape () const
	Get the configured input shape.
bool	hasBias () const
	Check if bias is enabled.
void	validate () const
	Validate configuration parameters.
LayerNormConfig &	withAxis (int64_t axis)
	Set the normalization axis.
LayerNormConfig &	withBias (bool has_bias)
	Set whether the layer should use bias.
LayerNormConfig &	withEpsilon (float epsilon)
	Set the epsilon value for numerical stability.
LayerNormConfig &	withInputShape (const std::vector< size_t > &input_shape)
	Set the input shape for the layer normalization.
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
int64_t	axis_ { -1 }
	The axis along which to normalize (default: -1 for last dimension)
float	epsilon_ { 1e-5f }
	Small constant added to variance for numerical stability.
bool	has_bias_ { true }
	Whether to include a learnable bias term.
std::vector< size_t >	input_shape_ {}
	Shape of the input tensor [batch_size, sequence_length, channels].

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 Layer Normalization module.

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

Constructor & Destructor Documentation

LayerNormConfig() [1/2]

Mila::Dnn::LayerNormConfig::LayerNormConfig ( )

default

Default constructor.

LayerNormConfig() [2/2]

Mila::Dnn::LayerNormConfig::LayerNormConfig ( size_t  normalized_dim )

inlineexplicit

Constructor with normalized dimension size.

Parameters

normalized_dim

The dimension size to normalize

Member Function Documentation

getAxis()

int64_t Mila::Dnn::LayerNormConfig::getAxis ( ) const

inline

Get the configured normalization axis.

Returns

int64_t The axis along which to normalize

Here is the caller graph for this function:

getEpsilon()

float Mila::Dnn::LayerNormConfig::getEpsilon ( ) const

inline

Get the configured epsilon value.

Returns

float The epsilon value for numerical stability

Here is the caller graph for this function:

getInputShape()

const std::vector< size_t > & Mila::Dnn::LayerNormConfig::getInputShape ( ) const

inline

Get the configured input shape.

Returns

const std::vector<size_t>& The input tensor shape

Here is the caller graph for this function:

hasBias()

bool Mila::Dnn::LayerNormConfig::hasBias ( ) const

inline

Check if bias is enabled.

Returns

bool Whether the layer has bias enabled

Here is the caller graph for this function:

validate()

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

inlinevirtual

Validate configuration parameters.

Exceptions

std::invalid_argument

If validation fails

Reimplemented from Mila::Dnn::ComponentConfig.

Here is the call graph for this function:

Here is the caller graph for this function:

withAxis()

LayerNormConfig & Mila::Dnn::LayerNormConfig::withAxis ( int64_t  axis )

inline

Set the normalization axis.

Parameters

axis	The axis along which to normalize (default is -1, the last dimension)

Returns

LayerNormConfig& Reference to this for method chaining

withBias()

LayerNormConfig & Mila::Dnn::LayerNormConfig::withBias ( bool  has_bias )

inline

Set whether the layer should use bias.

Parameters

has_bias

Whether to include a learnable bias term

Returns

LayerNormConfig& Reference to this for method chaining

withEpsilon()

LayerNormConfig & Mila::Dnn::LayerNormConfig::withEpsilon ( float  epsilon )

inline

Set the epsilon value for numerical stability.

Parameters

epsilon

Small constant added to variance for numerical stability

Returns

LayerNormConfig& Reference to this for method chaining

withInputShape()

LayerNormConfig & Mila::Dnn::LayerNormConfig::withInputShape ( const std::vector< size_t > &  input_shape )

inline

Set the input shape for the layer normalization.

Parameters

input_shape

The input tensor shape [batch_size, sequence_length, channels]

Returns

LayerNormConfig& Reference to this for method chaining

Member Data Documentation

axis_

int64_t Mila::Dnn::LayerNormConfig::axis_ { -1 }

private

The axis along which to normalize (default: -1 for last dimension)

epsilon_

float Mila::Dnn::LayerNormConfig::epsilon_ { 1e-5f }

private

Small constant added to variance for numerical stability.

has_bias_

bool Mila::Dnn::LayerNormConfig::has_bias_ { true }

private

Whether to include a learnable bias term.

input_shape_

std::vector<size_t> Mila::Dnn::LayerNormConfig::input_shape_ {}

private

Shape of the input tensor [batch_size, sequence_length, channels].

---

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

• /home/runner/work/Mila/Mila/Mila/Src/Dnn/Modules/Normalization/LayerNormConfig.ixx