Mila::Dnn::GatedMLPConfig Class Reference

Configuration class for the gated feed-forward (GatedMLP) block. More...

Inheritance diagram for Mila::Dnn::GatedMLPConfig:

Public Member Functions
	GatedMLPConfig (dim_t input_features, dim_t hidden_size)
	Construct GatedMLP configuration.
void	fromMetadata (const SerializationMetadata &meta) override
	Populate configuration from provided metadata.
ActivationType	getGateActivation () const noexcept
dim_t	getHiddenSize () const noexcept
dim_t	getInputFeatures () const noexcept
bool	hasBias () const noexcept
SerializationMetadata	toMetadata () const override
	Convert configuration into a SerializationMetadata object.
std::string	toString () const override
	Produce a short, human-readable summary of the configuration.
void	validate () const override
	Validate configuration parameters.
template<typename Self>
Self &&	withBias (this Self &&self, bool has_bias)
template<typename Self>
decltype(auto)	withGateActivation (this Self &&self, ActivationType gate_activation)
	Set the gate activation function (serializable metadata).
Public Member Functions inherited from Mila::Dnn::ComponentConfig
virtual	~ComponentConfig ()=default
	Virtual destructor for polymorphic base.

Detailed Description

Configuration class for the gated feed-forward (GatedMLP) block.

Block structure: Input -> fc_gate_up Linear(in -> 2H, fused) -> Swiglu gate (2H -> H) -> fc_down Linear(H -> in) -> Output

Gated FFNs are typically bias-free; has_bias defaults to false. gate_activation is serializable metadata; the model factory bridges it to the compile-time gate.

Constructor & Destructor Documentation

GatedMLPConfig()

Mila::Dnn::GatedMLPConfig::GatedMLPConfig ( dim_t input_features, dim_t hidden_size )

inline

Construct GatedMLP configuration.

Parameters

input_features	Number of input (and output) features (must be > 0).
hidden_size	Size of the gated intermediate dimension H (must be > 0).

Member Function Documentation

fromMetadata()

void Mila::Dnn::GatedMLPConfig::fromMetadata ( const SerializationMetadata & meta )

inlineoverridevirtual

Populate configuration from provided metadata.

Implementations should read available keys and leave missing keys at their current/default values to preserve forward/backward compatibility.

Parameters

meta	Metadata to read configuration values from.

Implements Mila::Dnn::ComponentConfig.

toMetadata()

SerializationMetadata Mila::Dnn::GatedMLPConfig::toMetadata ( ) const

inlineoverridevirtual

Convert configuration into a SerializationMetadata object.

Implementations should include any fields required to fully reconstruct the configuration via fromMetadata.

Returns

SerializationMetadata Metadata representation of the config.

Implements Mila::Dnn::ComponentConfig.

toString()

std::string Mila::Dnn::GatedMLPConfig::toString ( ) const

inlineoverridevirtual

Produce a short, human-readable summary of the configuration.

Implementations should return a compact, single-line description suitable for logging and debugging.

Returns

std::string Human-readable summary of the configuration.

Implements Mila::Dnn::ComponentConfig.

validate()

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

inlineoverridevirtual

Validate configuration parameters.

Called by callers to ensure the configuration represents a valid, constructible component. Implementations must throw std::invalid_argument (or a derived exception) when validation fails.

Exceptions

std::invalid_argument

If the configuration is invalid.

Implements Mila::Dnn::ComponentConfig.

---

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

• Mila/Src/Dnn/Components/FFN/GatedMLP/GatedMLP.Config.ixx