UnifiedLauncher< CpuBackend > Class Reference

CPU launcher specialization. More...

#include <unified_launcher.hpp>

Public Member Functions
	UnifiedLauncher ()=default
template<int In, int Out, typename Activation >
void	addLayer ()
	Add a layer to the CPU network. More...
void	buildNetwork ()
	Finalize parameters and internal buffers. More...
void	setData (const UnifiedDataset &data)
	Attach the training/test dataset. More...
void	train (UnifiedOptimizer< CpuBackend > &optimizer, const UnifiedConfig &config)
	Run training for the selected optimizer. More...
void	test ()
	Evaluate on test data. More...
NetworkWrapper< CpuBackend > &	getWrapper ()
	Access the underlying wrapper. More...

Detailed Description

CPU launcher specialization.

Constructor & Destructor Documentation

UnifiedLauncher()

UnifiedLauncher< CpuBackend >::UnifiedLauncher ( )

default

Member Function Documentation

addLayer()

template<int In, int Out, typename Activation >

void UnifiedLauncher< CpuBackend >::addLayer ( )

inline

Add a layer to the CPU network.

Template Parameters

In	Input dimension.
Out	Output dimension.
Activation	Activation function type.

buildNetwork()

void UnifiedLauncher< CpuBackend >::buildNetwork ( )

inline

Finalize parameters and internal buffers.

Allocates memory for weights and gradients based on added layers.

getWrapper()

NetworkWrapper<CpuBackend>& UnifiedLauncher< CpuBackend >::getWrapper ( )

inline

Access the underlying wrapper.

Returns

Reference to the NetworkWrapper.

setData()

void UnifiedLauncher< CpuBackend >::setData ( const UnifiedDataset &  data )

inline

Attach the training/test dataset.

Parameters

data	The UnifiedDataset containing train/test splits.

test()

void UnifiedLauncher< CpuBackend >::test ( )

inline

Evaluate on test data.

Prints MSE and Accuracy metrics to stdout.

train()

void UnifiedLauncher< CpuBackend >::train ( UnifiedOptimizer< CpuBackend > &  optimizer, const UnifiedConfig &  config  )

inline

Run training for the selected optimizer.

Parameters

optimizer	The optimization strategy to use.
config	configuration parameters for the experiment.

Here is the call graph for this function:

---

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

• src/unified_launcher.hpp