Encoder module for token and positional embeddings (device-templated). More...

Inheritance diagram for Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >:

Public Types
using	ComponentBase = Component<TDeviceType, TPrecision>
using	EmbeddingsTensorType = Tensor<TPrecision, MR>
using	MR = typename DeviceTypeTraits<TDeviceType>::memory_resource
using	TokenIndexType = Tensor<TIndex, MR>

Public Member Functions
	Lpe (const std::string &name, const LpeConfig &config, std::optional< DeviceId > device_id=std::nullopt)
	Construct Encoder component.
TokenIndexType &	backward (const TokenIndexType &input, const EmbeddingsTensorType &output_grad)
	Backward pass - compute parameter gradients and return owned input-grad.
EmbeddingsTensorType &	decode (const TokenIndexType &input, dim_t position)
	Decode pass - single token embedding at a specific sequence position.
EmbeddingsTensorType &	forward (const TokenIndexType &input)
	Forward pass - returns component-owned embeddings tensor.
DeviceId	getDeviceId () const override
	Get the compute device id associated with this component.
int64_t	getEmbeddingDim () const noexcept
std::vector< ITensor * >	getGradients () const override
	Return non-owning pointers to parameter gradient tensors.
int64_t	getMaxSequenceLength () const noexcept
MemoryStats	getMemoryStats () const override
	Return the current memory allocation breakdown for this component.
std::vector< std::string >	getParameterNames () const override
	List all available parameter names for this component.
std::vector< ITensor * >	getParameters () const override
	Return non-owning pointers to parameter tensors.
const ComponentType	getType () const override
	Get the component type identifier.
int64_t	getVocabularyLength () const noexcept
EmbeddingsTensorType *	getWpeGrad () const noexcept
EmbeddingsTensorType *	getWteGrad () const noexcept
void	loadParameter (const std::string &name, const ITensorBlob &blob) override
	Load a parameter from serialized tensor data.
dim_t	parameterCount () const override
	Return number of trainable parameters.
void	save_ (ModelArchive &archive, SerializationMode mode) const override
void	saveFlatTensors (Serialization::SafeTensorsWriter &writer, const std::string &prefix, Serialization::TensorSavePass pass) const override
	Drive this component's tensors through one pass of a flat safetensors save.
void	synchronize () override
std::string	toString () const override
	Produce a short, human-readable description of the component.
void	zeroGradients () override
	Clear all model-owned gradients for this component.
Public Member Functions inherited from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >
	Component (const std::string &name)
	Construct component with required name identifier.
virtual void	build (const BuildContext &context) final
	Build the component with the provided BuildContext (canonical overload).
const std::string	getName () const
	Get the component's name identifier.
virtual MemoryStats	getRequiredMemory (const BuildContext &context) const
	Report what build( context ) would allocate, without allocating it.
TrainingMode	getTrainingMode () const noexcept
	The current runtime behavioral mode of this Component.
virtual bool	isBuilt () const final
	Returns true if build() has completed successfully.
virtual void	load_ (ModelArchive &archive, SerializationMode mode)
	Restore this component's parameters from its archive scope.
virtual void	requireSerializableParameters () const
	Verify this component can serialize whatever parameters it owns.
void	setTrainingMode (TrainingMode mode)
	Set the runtime behavioral mode for this Component.

Protected Member Functions
void	onBuilding (const BuildContext &build_config) override
	Hook invoked by build() to allocate component buffers.
void	onExecutionContextSet () override
	Called after ExecutionContext is set on the base Component.
void	onTrainingModeChanging (TrainingMode training_mode) override
	Hook called before TrainingMode transitions.
Protected Member Functions inherited from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >
IExecutionContext *	getExecutionContext () const
	Get the shared execution context.
bool	hasExecutionContext () const noexcept
	Check if execution context has been set.
void	loadParameterFromBlob (const std::string &param_name, const Serialization::ITensorBlob &blob, Tensor< TParameterPrecision, TMemoryResource > &target, const shape_t &expected_shape)
	Load a tensor blob into a parameter tensor with validation.
void	saveParameterToArchive (ModelArchive &archive, const std::string &parameter_name, const Tensor< TParameterPrecision, TMemoryResource > &parameter) const
	Write one parameter tensor into the archive under "tensors/<name>".
void	saveParameterToWriter (Serialization::SafeTensorsWriter &writer, const std::string &flat_name, const Tensor< TParameterPrecision, TMemoryResource > &parameter, Serialization::TensorSavePass pass) const
	Drive one parameter through one pass of a flat safetensors save.
void	setExecutionContext (IExecutionContext *context)
	Set the execution context for this component.

Additional Inherited Members
Static Public Member Functions inherited from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >
static constexpr DeviceType	getDeviceType ()
	Compile-time device type for this component instance.
static constexpr TensorDataType	getPrecision () noexcept
	Compile-time tensor precision for this component instance.
Protected Types inherited from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >
using	HostStagingMemoryResource
	Host memory a device-resident parameter stages through.
Protected Attributes inherited from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >
BuildContext	build_context_
	The BuildContext stored at build time.

Detailed Description

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>
requires PrecisionSupportedOnDevice<TPrecision, TDeviceType>
class Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >

Encoder module for token and positional embeddings (device-templated).

Delegates computation to a device-specific UnaryOperation implementation registered in the OperationRegistry.

The Encoder transforms input token IDs into continuous vector representations:

Looks up token embeddings from vocabulary table (wte)
Adds positional embeddings (wpe) based on sequence position

Module owns trainable parameters (wte, wpe) and exposes them via accessors. The operation implements embedding lookup and position encoding addition.

Construction modes:

Standalone: provide a DeviceId to create and own an ExecutionContext.
Deferred/shared: omit DeviceId and caller must call setExecutionContext() before build().

Template Parameters

TDeviceType	Device type (DeviceType::Cpu or DeviceType::Cuda)
TIndex	Data type for token indices (typically INT32)
TPrecision	Abstract tensor precision (TensorDataType) for embeddings

Constructor & Destructor Documentation

◆ Lpe()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::Lpe	(	const std::string &	name,
		const LpeConfig &	config,
		std::optional< DeviceId >	device_id = std::nullopt )

inlineexplicit

Construct Encoder component.

Two modes:

Standalone mode: provide DeviceId and component will create and own an ExecutionContext.
Child mode: omit DeviceId and parent must call setExecutionContext() before build().

Parameters

name	Component name identifier (mandatory)
config	Encoder configuration
device_id	Optional DeviceId to create owned ExecutionContext (standalone mode)

Member Function Documentation

◆ backward()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

TokenIndexType & Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::backward	(	const TokenIndexType &	input,
		const EmbeddingsTensorType &	output_grad )

inline

Backward pass - compute parameter gradients and return owned input-grad.

Token indices are discrete and not differentiable; the backend may still expect an input-gradient tensor. The component owns a token-index-typed input-gradient buffer that is passed to the backend and returned.

Parameters

input	Input token indices tensor used during forward.
output_grad	Gradient w.r.t. embeddings [B, T, C].

Returns: Reference to component-owned token-index-typed input-grad tensor.

Exceptions

std::runtime_error if component is not built, not in training mode, or backend/buffers are not initialized.

◆ decode()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

EmbeddingsTensorType & Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::decode	(	const TokenIndexType &	input,
		dim_t	position )

inline

Decode pass - single token embedding at a specific sequence position.

Unlike forward() which processes a full sequence [B, T] and uses positions 0..T-1, decode() processes a single token and uses the caller-supplied position for the positional embedding lookup. This is critical for correctness in KV cache autoregressive generation – without the correct position, wpe[0] would be used for every generated token, corrupting all subsequent attention computations.

Parameters

input	Single token index tensor [1, 1]
position	Actual sequence position (prefill_len + decode_step)

Returns: Reference to component-owned embedding tensor [1, 1, C]

Exceptions

std::runtime_error if component is not built.

◆ forward()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

EmbeddingsTensorType & Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::forward ( const TokenIndexType & input )

inline

Forward pass - returns component-owned embeddings tensor.

Parameters

input Input token indices tensor [B, T]

Returns: Reference to component-owned embeddings tensor [B, T, C]

Exceptions

std::runtime_error if component is not built or backend not initialized.

◆ getDeviceId()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

DeviceId Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::getDeviceId ( ) const

inlineoverridevirtual

Get the compute device id associated with this component.

Must return the device on which parameters and operations execute.

Implements Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ getGradients()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

std::vector< ITensor * > Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::getGradients ( ) const

inlineoverridevirtual

Return non-owning pointers to parameter gradient tensors.

Gradient buffers are allocated only when the component is built in training mode, so a component built for inference returns an empty vector. Stateless components return empty in either mode. This is the accessor counterpart to getParameters() and does not throw on mode.

Returns: Vector of gradient pointers; empty when built for inference.

Exceptions

std::runtime_error if called before the component has been built.

Implements Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ getMemoryStats()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

MemoryStats Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::getMemoryStats ( ) const

inlineoverridevirtual

Return the current memory allocation breakdown for this component.

Reflects allocations at the moment of the call. The returned stats naturally track the component lifecycle:

After construction – nothing; construction allocates none After build( Inference ) – parameters + T=1 state buffers After build( Training ) – parameters + T=full state buffers After setTrainingMode( Train ) – parameters + state + gradients

For CompositeComponent and Network, the returned stats are the recursive aggregate of all child components.

May be called at any time – no lifecycle preconditions.

Returns: MemoryStats reflecting current allocations.

Implements Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ getParameterNames()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

std::vector< std::string > Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::getParameterNames ( ) const

inlineoverridevirtual

List all available parameter names for this component.

Returns an empty vector by default. Leaf components with parameters should override to return their canonical parameter name list in the same stable order used by save_() and loadParameter().

Reimplemented from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ getParameters()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

std::vector< ITensor * > Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::getParameters ( ) const

inlineoverridevirtual

Return non-owning pointers to parameter tensors.

The returned tensor pointers remain valid for the lifetime of the component. Order should be canonical (weights before biases).

Implements Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ getType()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

const ComponentType Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::getType ( ) const

inlineoverridevirtual

Get the component type identifier.

Used for serialization and runtime type identification.

Returns: Component type enum value.

Implements Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ loadParameter()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

void Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::loadParameter	(	const std::string &	name,
		const ITensorBlob &	blob )

inlineoverridevirtual

Load a parameter from serialized tensor data.

Loads raw tensor bytes directly into an existing parameter tensor, handling precision conversion and device upload as needed.

The component validates that the blob's shape matches the parameter's expected shape, then delegates to the backend to perform:

Precision conversion (blob dtype -> parameter dtype)
Device upload (CPU bytes -> target device)

Takes the parameter name used to locate the target tensor, and a blob holding serialized tensor metadata and raw bytes.

Exceptions

std::runtime_error	if component has no parameters to load.
std::runtime_error	if blob shape doesn't match parameter shape.

Reimplemented from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ onBuilding()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

void Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::onBuilding ( const BuildContext & build_config )

inlineoverrideprotectedvirtual

Hook invoked by build() to allocate component buffers.

Receives the stored BuildContext. Implementations must use config.allocationSeqLen() when sizing output buffers – this is the single call that makes Inference and Training allocate the correct buffer sizes automatically without per-component logic.

// Example -- Linear component:
shape_t out_shape =
{
    config.batchSize(),
    config.allocationSeqLen(),   // 1 for Inference, T for Training
    config_.getOutputFeatures()
};
output_ = std::make_unique<TensorType>( device, out_shape,
    this->getName() + ".output" );

The default implementation forwards to the legacy onBuilding( const shape_t& ) overload for backwards compatibility. New components should override this overload directly.

Note: Do not call build() or onBuilding() from within this hook.; Implementations should either succeed fully or leave no partial state, as a failed build() may be retried.

Takes the build-time configuration; use its allocationSeqLen() to obtain the correct output buffer sequence dimension.

Reimplemented from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ onExecutionContextSet()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

void Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::onExecutionContextSet ( )

inlineoverrideprotectedvirtual

Called after ExecutionContext is set on the base Component.

Initialize device-bound parameters and create the backend operation.

Reimplemented from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ onTrainingModeChanging()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

void Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::onTrainingModeChanging ( TrainingMode training_mode )

inlineoverrideprotectedvirtual

Hook called before TrainingMode transitions.

Called by setTrainingMode() after validation and lock acquisition, before the internal state is updated. Derived classes override to respond to the transition – e.g. zeroing gradient buffers on transition to Eval, or re-enabling dropout on transition to Training.

The default implementation is a no-op.

Takes the incoming TrainingMode.

Reimplemented from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ parameterCount()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

dim_t Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::parameterCount ( ) const

inlineoverridevirtual

Return number of trainable parameters.

For leaf components this is the element count of owned parameter tensors. CompositeComponent and Network implementations should return the recursive aggregate across all children.

Implements Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ save_()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

void Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::save_	(	ModelArchive &	archive,
		SerializationMode	mode ) const

inlineoverridevirtual

Implements Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ saveFlatTensors()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

void Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::saveFlatTensors	(	Serialization::SafeTensorsWriter &	writer,
		const std::string &	prefix,
		Serialization::TensorSavePass	pass ) const

inlineoverridevirtual

Drive this component's tensors through one pass of a flat safetensors save.

This is the path save_() cannot serve. A quantized weight is packed storage plus a scale companion and ModelArchive has no representation for that pairing, so save_() refuses it; the flat container expresses it as sibling tensors and this writes them.

The vocabulary is the flat format's: a fully qualified dotted component path plus a parameter name, exactly what loadParameters() parses on the way back in. Composites contribute nothing themselves – they recurse and extend the prefix – so the emitted set matches what the reader will route back through loadParameter().

Concrete parameter types, not ITensor, on purpose: a type-erased walk would have to ask at runtime whether a tensor's memory is host-accessible, and MemoryResource::is_host_accessible is a compile-time constant. Keeping concrete types here is what avoids widening an exported core type for a serialization concern.

The default refuses rather than writing nothing. A component that owns parameters and does not implement this would otherwise contribute silently to an artifact that loads, runs, and produces garbage – the same failure mode Phase 0 removed from save_().

Parameters

writer	Writer being driven.
prefix	Fully qualified component path, e.g. "tf_layer_0.qkv_proj".
pass	Declare reserves byte ranges; Write streams bytes in declaration order.

Reimplemented from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ synchronize()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

void Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::synchronize ( )

inlineoverridevirtual

         @brief Convenience accessor -- true if currently in Eval mode.

         Equivalent to getTrainingMode() == TrainingMode::Eval.
         Valid for both RuntimeMode::Inference and RuntimeMode::Training
         built components.

         @return true if in Eval mode.
        &zwj;/

bool isEvalMode() const noexcept { return getTrainingMode() == TrainingMode::Eval; }

    RuntimeMode getRuntimeMode() const noexcept
    {
        return build_context_.getRuntimeMode();
    }

    bool isInferenceMode() const noexcept
    {
        return build_context_.isInferenceMode();
    }

    bool isTrainingMode() const noexcept
    {
        return build_context_.isTrainingMode();
    }

====================================================================

Synchronization

    /**
       @brief Wait for outstanding device work submitted by this component.

       On CPU this may be a no-op. Use to ensure results are visible to
       the host or to measure synchronous timings.

Implements Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ toString()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

std::string Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::toString ( ) const

inlineoverridevirtual

Produce a short, human-readable description of the component.

Implementations should keep output concise and avoid throwing.

Implements Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

◆ zeroGradients()

template<DeviceType TDeviceType, TensorDataType TIndex = dtype_t::INT32, TensorDataType TPrecision = dtype_t::FP32>

void Mila::Dnn::Lpe< TDeviceType, TIndex, TPrecision >::zeroGradients ( )

inlineoverridevirtual

Clear all model-owned gradients for this component.

Default implementation is a no-op. Composite components should override to recurse to children. Leaf components should override to zero their parameter and activation gradients using device-aware helpers.

Reimplemented from Mila::Dnn::Component< TDeviceType, dtype_t::FP32 >.

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

Mila/Src/Dnn/Components/Encodings/Lpe/Lpe.ixx

Public Types

Public Member Functions

Protected Member Functions

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ Lpe()

Member Function Documentation

◆ backward()

◆ decode()

◆ forward()

◆ getDeviceId()

◆ getGradients()

◆ getMemoryStats()

◆ getParameterNames()

◆ getParameters()

◆ getType()

◆ loadParameter()

◆ onBuilding()

◆ onExecutionContextSet()

◆ onTrainingModeChanging()

◆ parameterCount()

◆ save_()

◆ saveFlatTensors()

◆ synchronize()

Synchronization

◆ toString()

◆ zeroGradients()