Mila::Dnn::DropoutConfig Class Reference

Configuration class for Dropout module. More...

Inheritance diagram for Mila::Dnn::DropoutConfig:

Collaboration diagram for Mila::Dnn::DropoutConfig:

Public Member Functions
	DropoutConfig ()=default
	Default constructor.
	DropoutConfig (float probability)
	Constructor with dropout probability.
float	getProbability () const
	Get the configured dropout probability.
bool	scalesDuringInference () const
	Check if scaling during inference is enabled.
bool	usesSameMaskPerBatch () const
	Check if the same mask is used for all elements in a batch.
void	validate () const
	Validate configuration parameters.
DropoutConfig &	withProbability (float probability)
	Configure the dropout probability.
DropoutConfig &	withSameMaskPerBatch (bool use_same_mask_per_batch)
	Configure whether to use the same dropout mask for all elements in a batch.
DropoutConfig &	withScalingDuringInference (bool scale_during_inference)
	Configure whether to apply scaling during inference.
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
float	probability_ { 0.5f }
	The probability of zeroing elements.
bool	scale_during_inference_ { false }
	Whether to apply scaling during inference.
bool	use_same_mask_per_batch_ { false }
	Whether to use the same mask for entire batch.

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 Dropout module.

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

Constructor & Destructor Documentation

DropoutConfig() [1/2]

Mila::Dnn::DropoutConfig::DropoutConfig ( )

default

Default constructor.

DropoutConfig() [2/2]

Mila::Dnn::DropoutConfig::DropoutConfig ( float  probability )

inlineexplicit

Constructor with dropout probability.

Parameters

probability

The dropout probability (0.0 to 1.0)

Member Function Documentation

getProbability()

float Mila::Dnn::DropoutConfig::getProbability ( ) const

inline

Get the configured dropout probability.

Returns

float The dropout probability

Here is the caller graph for this function:

scalesDuringInference()

bool Mila::Dnn::DropoutConfig::scalesDuringInference ( ) const

inline

Check if scaling during inference is enabled.

Returns

bool Whether scaling during inference is enabled

Here is the caller graph for this function:

usesSameMaskPerBatch()

bool Mila::Dnn::DropoutConfig::usesSameMaskPerBatch ( ) const

inline

Check if the same mask is used for all elements in a batch.

Returns

bool Whether the same mask is used per batch

Here is the caller graph for this function:

validate()

void Mila::Dnn::DropoutConfig::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:

withProbability()

DropoutConfig & Mila::Dnn::DropoutConfig::withProbability ( float  probability )

inline

Configure the dropout probability.

Parameters

probability

The probability of zeroing elements (0.0 to 1.0)

Returns

DropoutConfig& Reference to this for method chaining

withSameMaskPerBatch()

DropoutConfig & Mila::Dnn::DropoutConfig::withSameMaskPerBatch ( bool  use_same_mask_per_batch )

inline

Configure whether to use the same dropout mask for all elements in a batch.

Parameters

use_same_mask_per_batch

Whether to use the same mask for entire batch

Returns

DropoutConfig& Reference to this for method chaining

withScalingDuringInference()

DropoutConfig & Mila::Dnn::DropoutConfig::withScalingDuringInference ( bool  scale_during_inference )

inline

Configure whether to apply scaling during inference.

When true, outputs during inference will be scaled by 1/(1-p) to maintain the same expected value between training and inference. When false, dropout is completely disabled during inference.

Parameters

scale_during_inference

Whether to apply scaling during inference

Returns

DropoutConfig& Reference to this for method chaining

Member Data Documentation

probability_

float Mila::Dnn::DropoutConfig::probability_ { 0.5f }

private

The probability of zeroing elements.

scale_during_inference_

bool Mila::Dnn::DropoutConfig::scale_during_inference_ { false }

private

Whether to apply scaling during inference.

use_same_mask_per_batch_

bool Mila::Dnn::DropoutConfig::use_same_mask_per_batch_ { false }

private

Whether to use the same mask for entire batch.

---

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

• /home/runner/work/Mila/Mila/Mila/Src/Dnn/Modules/Regularization/DropoutConfig.ixx