Mila::Dnn::Loss< TDeviceType, TPrecision > Class Template Reference

Abstract base class for neural network loss functions. More...

Inheritance diagram for Mila::Dnn::Loss< TDeviceType, TPrecision >:

Public Types
using	ExecutionContextType = ExecutionContext<TDeviceType>
using	MR = typename DeviceTypeTraits<TDeviceType>::memory_resource
using	TensorType = Tensor<TPrecision, MR>

Additional Inherited Members
Public Member Functions inherited from Mila::Dnn::Component< TDeviceType, TPrecision >
	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).
virtual DeviceId	getDeviceId () const =0
	Get the compute device id associated with this component.
virtual std::vector< ITensor * >	getGradients () const =0
	Return non-owning pointers to parameter gradient tensors.
virtual MemoryStats	getMemoryStats () const =0
	Return the current memory allocation breakdown for this component.
const std::string	getName () const
	Get the component's name identifier.
virtual std::vector< std::string >	getParameterNames () const
	List all available parameter names for this component.
virtual std::vector< ITensor * >	getParameters () const =0
	Return non-owning pointers to parameter tensors.
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 const ComponentType	getType () const =0
	Get the component type identifier.
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	loadParameter (const std::string &, const Serialization::ITensorBlob &)
	Load a parameter from serialized tensor data.
virtual dim_t	parameterCount () const =0
	Return number of trainable parameters.
virtual void	requireSerializableParameters () const
	Verify this component can serialize whatever parameters it owns.
virtual void	save_ (ModelArchive &archive, SerializationMode mode) const =0
virtual void	saveFlatTensors (Serialization::SafeTensorsWriter &writer, const std::string &prefix, Serialization::TensorSavePass pass) const
	Drive this component's tensors through one pass of a flat safetensors save.
void	setTrainingMode (TrainingMode mode)
	Set the runtime behavioral mode for this Component.
virtual void	synchronize ()=0
virtual std::string	toString () const =0
	Produce a short, human-readable description of the component.
virtual void	zeroGradients ()
	Clear all model-owned gradients for this component.
Static Public Member Functions inherited from Mila::Dnn::Component< TDeviceType, TPrecision >
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, TPrecision >
using	HostStagingMemoryResource
	Host memory a device-resident parameter stages through.
Protected Member Functions inherited from Mila::Dnn::Component< TDeviceType, TPrecision >
IExecutionContext *	getExecutionContext () const
	Get the shared execution context.
bool	hasExecutionContext () const noexcept
	Check if execution context has been set.
template<TensorDataType TParameterPrecision, typename TMemoryResource>
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.
virtual void	onBuilding (const BuildContext &)
	Hook invoked by build() to allocate component buffers.
virtual void	onExecutionContextSet ()
	Lifecycle hook: Called immediately after ExecutionContext is set.
virtual void	onTrainingModeChanging (TrainingMode)
	Hook called before TrainingMode transitions.
template<TensorDataType TParameterPrecision, typename TMemoryResource>
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>".
template<TensorDataType TParameterPrecision, typename TMemoryResource>
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.
Protected Attributes inherited from Mila::Dnn::Component< TDeviceType, TPrecision >
BuildContext	build_context_ { shape_t{ 1 }, RuntimeMode::Training }
	The BuildContext stored at build time.

Detailed Description

template<Compute::DeviceType TDeviceType, TensorDataType TPrecision>
requires PrecisionSupportedOnDevice<TPrecision, TDeviceType>
class Mila::Dnn::Loss< TDeviceType, TPrecision >

Abstract base class for neural network loss functions.

Template Parameters

TDeviceType	Compile-time device identifier for this loss.
TPrecision	Data type used for computations.

Loss functions compute a scalar loss value given model predictions and target values. They may also provide hooks for optimizing the network graph and configuring reduction modes.

---

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

• Mila/Src/Dnn/Core/Loss.ixx