Dnn.Components.Residual Module Reference

Exported Modules
module	Compute.MemoryResource
module	Dnn.TensorTypes
module	Compute.DeviceId
module	Compute.Device
module	Dnn.Tensor
module	Logging.Logger
module	Dnn.Components.Linear
module	Dnn.TensorDataTypeTraits
module	Serialization.ModelArchive
module	Compute.ExecutionContextFactory
module	Dnn.Component
module	Compute.CpuMemoryResource
module	Dnn.ComponentType
module	Compute.IExecutionContext
module	Compute.BinaryOperation
module	Compute.OperationTraits
module	Compute.UnaryOperation
module	Dnn.TensorDataType
module	Compute.DeviceType
module	Serialization.Mode
module	Dnn.ITensor
module	Dnn.Components.ResidualConfig
module	Compute.DeviceTypeTraits

Classes
class	Mila::Dnn::Residual< TDeviceType, TPrecision >
	Device-templated Residual connection component. More...

Typedefs
using	ComponentBase = Component<TDeviceType, TPrecision>
using	MR = typename DeviceTypeTraits<TDeviceType>::memory_resource
using	OpType = typename OperationTraits<OperationType::ResidualOp, TDeviceType, TPrecision>::type
using	TensorType = Tensor<TPrecision, MR>

Functions
	Residual (const std::string &name, const ResidualConfig &config, std::optional< DeviceId > device_id=std::nullopt)
	Construct Residual component with optional ExecutionContext ownership.
	~Residual () override=default
std::pair< TensorType &, TensorType & >	backward (const TensorType &input_a, const TensorType &input_b, const TensorType &output_grad)
	Execute the backward pass and return component-owned gradient for input_a.
void	createOperation ()
	Create backend BinaryOperation from OperationRegistry.
TensorType &	forward (const TensorType &input_a, const TensorType &input_b)
	Execute the forward pass and return component-owned output tensor.
DeviceId	getDeviceId () const override
	Get the compute device id associated with this component.
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.
void	onBuilding (const BuildContext &build_config) override
	Build the Residual component from the provided BuildContext.
void	onExecutionContextSet () override
	Hook invoked after ExecutionContext is set on the base Component.
void	onTrainingModeChanging (TrainingMode training_mode) override
	Hook invoked when training mode is about to change.
size_t	parameterCount () const override
	Number of trainable parameters.
void	save_ (ModelArchive &archive, SerializationMode mode) const override
	Serialize component parameters into the provided archive.
void	synchronize () override
	Block until all device operations submitted by this component complete.
std::string	toString () const override
	Return a human-readable description of the component.

Variables
ResidualConfig	config_
std::unique_ptr< TensorType >	input_a_grad_ { nullptr }
std::unique_ptr< TensorType >	input_b_grad_ { nullptr }
shape_t	leading_shape_
std::shared_ptr< OpType >	operation_ { nullptr }
std::unique_ptr< TensorType >	output_ { nullptr }
std::unique_ptr< TensorType >	output_view_ { nullptr }
std::unique_ptr< IExecutionContext >	owned_exec_context_ { nullptr }

Files
file	/__w/Mila/Mila/Mila/Src/Dnn/Components/Connections/Residual.ixx
	Device-templated Residual connection component.