Dnn.Components.TokenEmbedding Module Reference

Exported Modules
module	Compute.UnaryOperation
module	Dnn.TensorDataType
module	Compute.MemoryResource
module	Dnn.TensorTypes
module	Dnn.Components.TokenEmbeddingConfig
module	Compute.DeviceTypeTraits
module	Dnn.TensorHelpers
module	Logging.Logger
module	Compute.Device
module	Compute.DeviceId
module	Dnn.Tensor
module	Serialization.Tensor
module	Dnn.TensorDataTypeTraits
module	Serialization.ModelArchive
module	Compute.ExecutionContextFactory
module	Dnn.Component
module	Compute.CpuMemoryResource
module	Dnn.ComponentType
module	Dnn.TensorOps
module	Dnn.TensorOps
module	Compute.OperationTraits
module	Compute.DeviceType
module	Dnn.ITensor
module	Serialization.Mode
module	Compute.ExecutionContext

Classes
class	Mila::Dnn::TokenEmbedding< TDeviceType, TIndex, TPrecision >
	Pure token embedding component (device-templated). More...

Typedefs
using	ComponentBase = Component<TDeviceType, TPrecision>
using	EmbeddingTensorType = Tensor<TPrecision, MR>
using	MR = typename DeviceTypeTraits<TDeviceType>::memory_resource
using	OpType = typename OperationTraits<OperationType::TokenEmbeddingOp, TDeviceType, TPrecision>::type
using	TokenIndexType = Tensor<TIndex, MR>

Functions
	TokenEmbedding (const std::string &name, const TokenEmbeddingConfig &config, std::optional< DeviceId > device_id=std::nullopt)
	Construct a TokenEmbedding component.
	~TokenEmbedding () override=default
TokenIndexType &	backward (const TokenIndexType &input, const EmbeddingTensorType &output_grad)
	Backward pass — accumulates gradients into wte.
void	createOperation ()
EmbeddingTensorType &	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.
MemoryStats	getMemoryStats () const override
	Return the current memory allocation breakdown 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	getVocabSize () const noexcept
EmbeddingTensorType *	getWteGrad () const noexcept
void	initializeParameterGradients ()
void	initializeParameters ()
void	loadParameter (const std::string &name, const ITensorBlob &blob) override
	Load a parameter from serialized tensor data.
void	onBuilding (const BuildContext &build_context) override
	Hook invoked by build() to allocate component buffers.
void	onExecutionContextSet () override
	Lifecycle hook: Called immediately after ExecutionContext is set.
void	onTrainingModeChanging (TrainingMode training_mode) override
	Hook called before TrainingMode transitions.
size_t	parameterCount () const override
	Return number of trainable parameters.
void	save_ (ModelArchive &archive, SerializationMode mode) const override
void	synchronize () override
	Wait for outstanding device work submitted by this component.
std::string	toString () const override
	Produce a short, human-readable description of the component.
void	validateBuildContext (const BuildContext &context) const
void	zeroGradients () override
	Clear all model-owned gradients for this component.

Variables
TokenEmbeddingConfig	config_
std::unique_ptr< EmbeddingTensorType >	current_output_view_ { nullptr }
std::unique_ptr< TokenIndexType >	input_grad_ { nullptr }
int64_t	max_batch_size_ { 0 }
int64_t	max_seq_len_ { 0 }
std::shared_ptr< OpType >	operation_ { nullptr }
std::unique_ptr< EmbeddingTensorType >	output_ { nullptr }
std::unique_ptr< IExecutionContext >	owned_exec_context_ { nullptr }
std::unique_ptr< EmbeddingTensorType >	wte_ { nullptr }
std::unique_ptr< EmbeddingTensorType >	wte_grad_ { nullptr }

Files
file	/__w/Mila/Mila/Mila/Src/Dnn/Components/Embeddings/TokenEmbedding.ixx
	Device-templated TokenEmbedding component.