onert::backend::basic::train::TrainableTensor Class Reference

#include <TrainableTensor.h>

Collaboration diagram for onert::backend::basic::train::TrainableTensor:

Public Member Functions
	TrainableTensor ()=delete
virtual	~TrainableTensor ()=default
	TrainableTensor (const ir::OperandInfo &info)
void	setBuffer (uint8_t *buffer)
	Set the Buffer object. This method is called for static and non-const tensor.
uint8_t *	buffer () const override
std::vector< ITensor * >	optVars () override
	Get optimizer variables of this trainable tensor.
void	appendOptVar (std::unique_ptr< Tensor > opt_var)
void	setOptVarBuffer (uint8_t *buffer, size_t pos)
void	fillBuffer (const std::shared_ptr< ir::Data > &data)
Public Member Functions inherited from onert::backend::train::ITrainableTensor
virtual	~ITrainableTensor ()=default
	IPortableTensor (const ir::OperandInfo &info)
Public Member Functions inherited from onert::backend::IPortableTensor
	IPortableTensor (const ir::OperandInfo &info)
virtual	~IPortableTensor ()
const ir::OperandInfo &	get_info () const
const ir::Sparsity *	sparsity () const
size_t	total_size () const override final
size_t	calcOffset (const ir::Coordinates &coords) const override final
ir::DataType	data_type () const override final
float	data_scale () const override final
int32_t	data_zero_point () const override final
const std::vector< float > &	data_scales () const override final
const std::vector< int32_t > &	data_zero_points () const override
bool	is_constant () const override final
	Return true if the tensor is constant.
bool	is_dynamic () const override final
	Return true if the tensor needs dynamic allocation, meaning that during compile-time the outpus shape cannot be known and the output shape is calculated during kernel execution-time.
ir::Shape	getShape () const override final
	Get ir::Shape of tensor.
bool	has_padding () const final
void	access (const std::function< void(ITensor &tensor)> &fn) final
Public Member Functions inherited from onert::backend::ITensor
virtual	~ITensor ()
virtual void	deallocBuffer ()
	Dealloc the buffer (only for dynamic tensors)
virtual bool	is_subtensor () const
virtual bool	needMemoryMap () const
virtual void	enqueueWriteBuffer (const void *, bool)
virtual void	enqueueReadBuffer (void *, bool)

Protected Attributes
Tensor	_tensor
std::vector< std::unique_ptr< Tensor > >	_opt_vars
Protected Attributes inherited from onert::backend::IPortableTensor
ir::OperandInfo	_info

Detailed Description

Definition at line 27 of file TrainableTensor.h.

Constructor & Destructor Documentation

TrainableTensor() [1/2]

onert::backend::basic::train::TrainableTensor::TrainableTensor ( )

delete

~TrainableTensor()

virtual onert::backend::basic::train::TrainableTensor::~TrainableTensor ( )

virtualdefault

TrainableTensor() [2/2]

onert::backend::basic::train::TrainableTensor::TrainableTensor ( const ir::OperandInfo &  info )

inline

Definition at line 34 of file TrainableTensor.h.

    : ITrainableTensor{info}, _tensor{info, nullptr}, _opt_vars{}
  {
    // DO NOTHING
  }

Member Function Documentation

appendOptVar()

void onert::backend::basic::train::TrainableTensor::appendOptVar ( std::unique_ptr< Tensor >  opt_var )

inline

Definition at line 51 of file TrainableTensor.h.

{ _opt_vars.emplace_back(std::move(opt_var)); }

References _opt_vars.

buffer()

uint8_t * onert::backend::basic::train::TrainableTensor::buffer ( ) const

inlineoverridevirtual

Implements onert::backend::ITensor.

Definition at line 47 of file TrainableTensor.h.

{ return _tensor.buffer(); }

References _tensor, and onert::backend::basic::Tensor::buffer().

Referenced by fillBuffer(), setBuffer(), and setOptVarBuffer().

fillBuffer()

void onert::backend::basic::train::TrainableTensor::fillBuffer

(

const std::shared_ptr< ir::Data > &

data

)

Definition at line 32 of file TrainableTensor.cc.

{
  auto *buffer = _tensor.buffer();
  assert(buffer);
  assert(total_size() == data->size());
  std::memcpy(buffer, data->base(), data->size());
}

References _tensor, onert::backend::basic::Tensor::buffer(), buffer(), and onert::backend::IPortableTensor::total_size().

Referenced by onert::backend::train::BackendContext::gen().

optVars()

std::vector< ITensor * > onert::backend::basic::train::TrainableTensor::optVars ( )

overridevirtual

Get optimizer variables of this trainable tensor.

Returns

Optimizer variables

Implements onert::backend::train::ITrainableTensor.

Definition at line 22 of file TrainableTensor.cc.

{
  std::vector<ITensor *> ret;
  for (auto &&e : _opt_vars)
  {
    ret.emplace_back(e.get());
  }
  return ret;
}

References _opt_vars.

setBuffer()

void onert::backend::basic::train::TrainableTensor::setBuffer ( uint8_t *  buffer )

inline

Set the Buffer object. This method is called for static and non-const tensor.

Definition at line 44 of file TrainableTensor.h.

{ _tensor.setBuffer(buffer); }

References _tensor, buffer(), and onert::backend::basic::Tensor::setBuffer().

setOptVarBuffer()

void onert::backend::basic::train::TrainableTensor::setOptVarBuffer ( uint8_t *  buffer, size_t  pos  )

inline

Definition at line 52 of file TrainableTensor.h.

{ _opt_vars.at(pos)->setBuffer(buffer); }

References _opt_vars, and buffer().

Field Documentation

_opt_vars

std::vector<std::unique_ptr<Tensor> > onert::backend::basic::train::TrainableTensor::_opt_vars

protected

Definition at line 64 of file TrainableTensor.h.

Referenced by appendOptVar(), optVars(), and setOptVarBuffer().

_tensor

Tensor onert::backend::basic::train::TrainableTensor::_tensor

protected

Definition at line 63 of file TrainableTensor.h.

Referenced by buffer(), fillBuffer(), and setBuffer().

---

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

• runtime/onert/core/include/backend/basic/train/TrainableTensor.h
• runtime/onert/core/src/backend/basic/train/TrainableTensor.cc