ONE - On-device Neural Engine
Loading...
Searching...
No Matches
onert::backend::train::ops::TrainingKernelRegistry Class Referenceabstract

#include <PoolLayer.h>

Public Member Functions

virtual void forward (const IPortableTensor *in, IPortableTensor *out)=0
 
virtual void backward (const IPortableTensor *back_prop_out, IPortableTensor *back_prop_in)=0
 
 TrainingKernelRegistry ()=default
 
virtual ~TrainingKernelRegistry ()=default
 

Detailed Description

This is to register the pair of (forward, backward) training kernel.

Definition at line 36 of file PoolLayer.h.

Constructor & Destructor Documentation

◆ TrainingKernelRegistry()

onert::backend::train::ops::TrainingKernelRegistry::TrainingKernelRegistry ( )
default

◆ ~TrainingKernelRegistry()

virtual onert::backend::train::ops::TrainingKernelRegistry::~TrainingKernelRegistry ( )
virtualdefault

Member Function Documentation

◆ backward()

virtual void onert::backend::train::ops::TrainingKernelRegistry::backward ( const IPortableTensor back_prop_out,
IPortableTensor back_prop_in 
)
pure virtual

◆ forward()

virtual void onert::backend::train::ops::TrainingKernelRegistry::forward ( const IPortableTensor in,
IPortableTensor out 
)
pure virtual

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