ArmNN
 25.11
Loading...
Searching...
No Matches
PadDescriptor Struct Reference

A PadDescriptor for the PadLayer. More...

#include <Descriptors.hpp>

Inheritance diagram for PadDescriptor:
[legend]
Collaboration diagram for PadDescriptor:
[legend]

Public Member Functions

 PadDescriptor ()
 PadDescriptor (const std::vector< std::pair< unsigned int, unsigned int > > &padList, const float &padValue=0, const PaddingMode &paddingMode=PaddingMode::Constant)
bool operator== (const PadDescriptor &rhs) const
Public Member Functions inherited from BaseDescriptor
virtual bool IsNull () const
virtual ~BaseDescriptor ()=default

Public Attributes

std::vector< std::pair< unsigned int, unsigned int > > m_PadList
 Specifies the padding for input dimension.
float m_PadValue
 Optional value to use for padding, defaults to 0.
PaddingMode m_PaddingMode
 Specifies the Padding mode (Constant, Reflect or Symmetric)

Detailed Description

A PadDescriptor for the PadLayer.

Definition at line 1196 of file Descriptors.hpp.

Constructor & Destructor Documentation

◆ PadDescriptor() [1/2]

PadDescriptor ( )
inline

Definition at line 1198 of file Descriptors.hpp.

1198 : m_PadValue(0), m_PaddingMode(PaddingMode::Constant)
1199 {}

References armnn::Constant, m_PaddingMode, and m_PadValue.

Referenced by operator==().

◆ PadDescriptor() [2/2]

PadDescriptor ( const std::vector< std::pair< unsigned int, unsigned int > > & padList,
const float & padValue = 0,
const PaddingMode & paddingMode = PaddingMode::Constant )
inline

Definition at line 1201 of file Descriptors.hpp.

1204 : m_PadList(padList)
1205 , m_PadValue(padValue)
1206 , m_PaddingMode(paddingMode)
1207 {}

References armnn::Constant, m_PaddingMode, m_PadList, and m_PadValue.

Member Function Documentation

◆ operator==()

bool operator== ( const PadDescriptor & rhs) const
inline

Definition at line 1209 of file Descriptors.hpp.

1210 {
1211 return m_PadList == rhs.m_PadList && m_PadValue == rhs.m_PadValue && m_PaddingMode == rhs.m_PaddingMode;
1212 }

References m_PaddingMode, m_PadList, m_PadValue, and PadDescriptor().

Member Data Documentation

◆ m_PaddingMode

◆ m_PadList

std::vector<std::pair<unsigned int, unsigned int> > m_PadList

Specifies the padding for input dimension.

First is the number of values to add before the tensor in the dimension. Second is the number of values to add after the tensor in the dimension. The number of pairs should match the number of dimensions in the input tensor.

Definition at line 1218 of file Descriptors.hpp.

Referenced by armnn::ClPadValidate(), ClPadWorkload::ClPadWorkload(), ConvertPadToTosaOperator(), PadLayer::CreateWorkload(), armnn::MirrorPad(), NeonPadWorkload::NeonPadWorkload(), armnn::NeonPadWorkloadValidate(), operator==(), armnn::Pad(), PadDescriptor(), StringifyLayerParameters< PadDescriptor >::Serialize(), and armnn::optimizations::pad_fold::TryFoldPadIntoLayer2d().

◆ m_PadValue


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