Inheritance diagram for Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >:

[legend]

Collaboration diagram for Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >:

[legend]

Public Types
using	ConfigType = TokenEmbeddingConfig
using	CudaExecutionContext = ExecutionContext<DeviceType::Cuda>
using	MR = CudaDeviceMemoryResource
using	NativeType = typename Mila::Dnn::Compute::Cuda::TensorDataTypeMap<TPrecision>::device_type
using	TensorType = Tensor<TPrecision, MR>
using	UnaryOperationBase = UnaryOperation<DeviceType::Cuda, TInput, TPrecision>
Public Types inherited from Mila::Dnn::Compute::UnaryOperation< DeviceType::Cuda, TInput, TInput >
using	MR
using	TensorInputType
using	TensorOutputType
Public Types inherited from Mila::Dnn::Compute::Operation< TDeviceType, TPrecision >
using	DataTypeTraits

Public Member Functions
	CudaTokenEmbeddingOp (IExecutionContext *context, const TokenEmbeddingConfig &config)
void	backward (const ITensor &input, const ITensor &output_grad, ITensor &input_grad) const override
	Backward pass accumulating gradients into wte (hot path).
void	build (const BuildContext &config) override
	Prepare the operation for a concrete input shape (cold path).
void	decode (const ITensor &input, ITensor &output) const
	Single-token decode pass (hot path).
void	forward (const ITensor &input, ITensor &output) const override
	Full-sequence forward pass (hot path).
std::string	getName () const override
	Human-readable operation name.
OperationType	getOperationType () const override
	Operation type identifier.
void	setGradients (ITensor wte_grad, ITensor ) override
	Bind the wte gradient tensor for training (module retains ownership).
void	setParameters (ITensor wte, ITensor ) override
	Bind the wte parameter tensor (module retains ownership).
Public Member Functions inherited from Mila::Dnn::Compute::UnaryOperation< DeviceType::Cuda, TInput, TInput >
virtual	~UnaryOperation ()=default
Public Member Functions inherited from Mila::Dnn::Compute::Operation< TDeviceType, TPrecision >
virtual	~Operation ()=default
virtual void	clearGradients () noexcept
	Clear any cached gradient pointers held by the operation.
virtual TensorDataType	getDataType () const
	Tensor data type for this operation.
virtual DeviceType	getDeviceType () const
	Device type for this operation.
virtual std::size_t	getStateMemorySize () const
	Returns the number of bytes of state memory allocated by this operation.
virtual bool	isBuilt () const
	Whether build() completed successfully for a concrete input shape.
virtual bool	isEvalMode () const
	Query whether operation is configured for training.
virtual void	setTrainingMode (TrainingMode training_mode)
	Configure operation training-mode behavior.

Private Member Functions
void	validateInputShape (const shape_t &shape) const
void	validateRuntimeShape (int B, int T) const

Private Attributes
int	batch_size_ { 0 }
TokenEmbeddingConfig	config_
CudaExecutionContext *	context_
int	embedding_dim_ { 0 }
int	seq_length_ { 0 }
int	vocab_size_ { 0 }
NativeType *	wte_ { nullptr }
NativeType *	wte_grad_ { nullptr }

Additional Inherited Members
Static Public Attributes inherited from Mila::Dnn::Compute::Operation< TDeviceType, TPrecision >
static constexpr TensorDataType	data_type
static constexpr DeviceType	device_type
Static Protected Member Functions inherited from Mila::Dnn::Compute::UnaryOperation< DeviceType::Cuda, TInput, TInput >
static const TensorInputType &	asInputTensor (const ITensor &t)
static TensorOutputType &	asOutputTensor (ITensor &t)
Protected Attributes inherited from Mila::Dnn::Compute::Operation< TDeviceType, TPrecision >
bool	is_built_
TrainingMode	training_mode_

Member Typedef Documentation

◆ ConfigType

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

using Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::ConfigType = TokenEmbeddingConfig

◆ CudaExecutionContext

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

using Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::CudaExecutionContext = ExecutionContext<DeviceType::Cuda>

◆ MR

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

using Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::MR = CudaDeviceMemoryResource

◆ NativeType

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

using Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::NativeType = typename Mila::Dnn::Compute::Cuda::TensorDataTypeMap<TPrecision>::device_type

◆ TensorType

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

using Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::TensorType = Tensor<TPrecision, MR>

◆ UnaryOperationBase

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

using Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::UnaryOperationBase = UnaryOperation<DeviceType::Cuda, TInput, TPrecision>

Constructor & Destructor Documentation

◆ CudaTokenEmbeddingOp()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::CudaTokenEmbeddingOp	(	IExecutionContext *	context,
		const TokenEmbeddingConfig &	config )

inline

Member Function Documentation

◆ backward()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

void Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::backward	(	const ITensor &	input,
		const ITensor &	output_grad,
		ITensor &	input_grad ) const

inlineoverridevirtual

Backward pass accumulating gradients into wte (hot path).

Token indices are non-differentiable; input_grad is unused.

Parameters

input	Token indices used in forward [B, T] (INT32).
output_grad	Upstream embedding gradient [B, T, C].
input_grad	Unused (non-differentiable input).

Implements Mila::Dnn::Compute::UnaryOperation< DeviceType::Cuda, TInput, TInput >.

◆ build()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

void Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::build ( const BuildContext & config )

inlineoverridevirtual

Prepare the operation for a concrete input shape (cold path).

Parameters

input_shape Token index input shape [B, T].

Exceptions

std::runtime_error	if wte is not bound.
std::invalid_argument	if input shape is invalid.

Reimplemented from Mila::Dnn::Compute::Operation< TDeviceType, TPrecision >.

◆ decode()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

void Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::decode	(	const ITensor &	input,
		ITensor &	output ) const

inline

Single-token decode pass (hot path).

Computes output[b,:] = wte[X[b,0],:] for each batch element. No position argument — positional encoding is handled downstream.

Parameters

input	Single-token indices [B, 1] (INT32).
output	Pre-allocated output buffer [B, C].

◆ forward()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

void Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::forward	(	const ITensor &	input,
		ITensor &	output ) const

inlineoverridevirtual

Full-sequence forward pass (hot path).

For each (b, t): output[b,t,:] = wte[X[b,t],:].

Parameters

input	Token indices [B, T] (INT32).
output	Pre-allocated embeddings [B, T, C].

Implements Mila::Dnn::Compute::UnaryOperation< DeviceType::Cuda, TInput, TInput >.

◆ getName()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

std::string Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::getName ( ) const

inlineoverridevirtual

Human-readable operation name.

Implements Mila::Dnn::Compute::Operation< TDeviceType, TPrecision >.

◆ getOperationType()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

OperationType Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::getOperationType ( ) const

inlineoverridevirtual

Operation type identifier.

Implements Mila::Dnn::Compute::Operation< TDeviceType, TPrecision >.

◆ setGradients()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

void Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::setGradients	(	ITensor *	wte_grad,
		ITensor *	)

inlineoverridevirtual

Bind the wte gradient tensor for training (module retains ownership).

Parameters

wte_grad Gradient buffer for wte — CUDA tensor of shape [vocab_size, C].

Exceptions

std::invalid_argument on null or non-CUDA tensor.

Reimplemented from Mila::Dnn::Compute::Operation< TDeviceType, TPrecision >.

◆ setParameters()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

void Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::setParameters	(	ITensor *	wte,
		ITensor *	)

inlineoverridevirtual

Bind the wte parameter tensor (module retains ownership).

Parameters

wte	Token embedding table — CUDA tensor of shape [vocab_size, C].

Exceptions

std::invalid_argument on null, non-CUDA, or shape-mismatched tensor.

Reimplemented from Mila::Dnn::Compute::Operation< TDeviceType, TPrecision >.

◆ validateInputShape()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

void Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::validateInputShape ( const shape_t & shape ) const

inlineprivate

Here is the caller graph for this function:

◆ validateRuntimeShape()

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

void Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::validateRuntimeShape	(	int	B,
		int	T ) const

inlineprivate

Here is the caller graph for this function:

Member Data Documentation

◆ batch_size_

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

int Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::batch_size_ { 0 }

private

◆ config_

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

TokenEmbeddingConfig Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::config_

private

◆ context_

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

CudaExecutionContext* Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::context_

private

◆ embedding_dim_

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

int Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::embedding_dim_ { 0 }

private

◆ seq_length_

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

int Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::seq_length_ { 0 }

private

◆ vocab_size_

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

int Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::vocab_size_ { 0 }

private

◆ wte_

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

NativeType* Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::wte_ { nullptr }

private

◆ wte_grad_

template<TensorDataType TInput, TensorDataType TPrecision = TInput>

NativeType* Mila::Dnn::Compute::Cuda::TokenEmbedding::CudaTokenEmbeddingOp< TInput, TPrecision >::wte_grad_ { nullptr }

private

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

/__w/Mila/Mila/Mila/Src/Dnn/Compute/Devices/Cuda/Operations/Embeddings/CudaTokenEmbeddingOp.ixx

Public Types

Public Member Functions

Private Member Functions

Private Attributes

Additional Inherited Members

Member Typedef Documentation

◆ ConfigType

◆ CudaExecutionContext

◆ MR

◆ NativeType

◆ TensorType

◆ UnaryOperationBase

Constructor & Destructor Documentation

◆ CudaTokenEmbeddingOp()

Member Function Documentation

◆ backward()

◆ build()

◆ decode()

◆ forward()

◆ getName()

◆ getOperationType()

◆ setGradients()

◆ setParameters()

◆ validateInputShape()

◆ validateRuntimeShape()

Member Data Documentation

◆ batch_size_

◆ config_

◆ context_

◆ embedding_dim_

◆ seq_length_

◆ vocab_size_

◆ wte_

◆ wte_grad_