WorkloadFactoryBase Class Reference

#include <WorkloadFactoryBase.hpp>

Inheritance diagram for WorkloadFactoryBase:

Collaboration diagram for WorkloadFactoryBase:

Public Member Functions
bool	SupportsSubTensors () const override
std::unique_ptr< ITensorHandle >	CreateSubTensorHandle (ITensorHandle &, TensorShape const &, unsigned int const *) const override
std::unique_ptr< ITensorHandle >	CreateTensorHandle (const TensorInfo &, const bool) const override
std::unique_ptr< ITensorHandle >	CreateTensorHandle (const TensorInfo &, DataLayout, const bool) const override
std::unique_ptr< IWorkload >	CreateWorkload (LayerType, const QueueDescriptor &, const WorkloadInfo &) const override
	Backends should implement their own CreateWorkload function with a switch statement. More...
Public Member Functions inherited from IWorkloadFactory
virtual	~IWorkloadFactory ()
virtual void	AfterWorkloadsCreated ()
virtual const BackendId &	GetBackendId () const =0

Additional Inherited Members
Static Public Member Functions inherited from IWorkloadFactory
static bool	IsLayerSupported (const BackendId &backendId, const IConnectableLayer &layer, Optional< DataType > dataType, std::string &outReasonIfUnsupported)
static bool	IsLayerSupported (const IConnectableLayer &layer, Optional< DataType > dataType, std::string &outReasonIfUnsupported)
static bool	IsLayerSupported (const IConnectableLayer &layer, Optional< DataType > dataType, std::string &outReasonIfUnsupported, const ModelOptions &modelOptions)
static bool	IsLayerSupported (const BackendId &backendId, const IConnectableLayer &layer, Optional< DataType > dataType, std::string &outReasonIfUnsupported, const ModelOptions &modelOptions)

Detailed Description

Definition at line 13 of file WorkloadFactoryBase.hpp.

Member Function Documentation

CreateSubTensorHandle()

std::unique_ptr<ITensorHandle> CreateSubTensorHandle ( ITensorHandle &  , TensorShape const &  , unsigned int const *    ) const

inlineoverridevirtual

Implements IWorkloadFactory.

Reimplemented in ClWorkloadFactory, and NeonWorkloadFactory.

Definition at line 19 of file WorkloadFactoryBase.hpp.

    { return nullptr; };

CreateTensorHandle() [1/2]

std::unique_ptr<ITensorHandle> CreateTensorHandle ( const TensorInfo &  , const bool    ) const

inlineoverridevirtual

Implements IWorkloadFactory.

Reimplemented in ClWorkloadFactory, and NeonWorkloadFactory.

Definition at line 25 of file WorkloadFactoryBase.hpp.

    { return nullptr; }

CreateTensorHandle() [2/2]

std::unique_ptr<ITensorHandle> CreateTensorHandle ( const TensorInfo &  , DataLayout  , const bool    ) const

inlineoverridevirtual

Implements IWorkloadFactory.

Reimplemented in ClWorkloadFactory, and NeonWorkloadFactory.

Definition at line 29 of file WorkloadFactoryBase.hpp.

    { return nullptr; }

CreateWorkload()

std::unique_ptr<IWorkload> CreateWorkload ( LayerType  type, const QueueDescriptor &  descriptor, const WorkloadInfo &  info  ) const

inlineoverridevirtual

Backends should implement their own CreateWorkload function with a switch statement.

The case for the switch should be the LayerType and based on that they will call their specific workload creation functionality.

Implements IWorkloadFactory.

Reimplemented in ClWorkloadFactory, and NeonWorkloadFactory.

Definition at line 34 of file WorkloadFactoryBase.hpp.

    { return nullptr; }

SupportsSubTensors()

bool SupportsSubTensors ( ) const

inlineoverridevirtual

Implements IWorkloadFactory.

Reimplemented in ClWorkloadFactory, and NeonWorkloadFactory.

Definition at line 16 of file WorkloadFactoryBase.hpp.

    { return false; };

---

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

• src/backends/backendsCommon/WorkloadFactoryBase.hpp