ArmNN
 24.08
Float32ToFloat16 Struct Reference

#include <ConvertConstants.hpp>

Static Public Member Functions

static void Func (std::shared_ptr< ConstTensorHandle > &handle)
 

Detailed Description

Definition at line 42 of file ConvertConstants.hpp.

Member Function Documentation

◆ Func()

static void Func ( std::shared_ptr< ConstTensorHandle > &  handle)
inlinestatic

Definition at line 44 of file ConvertConstants.hpp.

45  {
46  const TensorInfo& info = handle->GetTensorInfo();
47 
48  if (info.GetDataType() == DataType::Float32)
49  {
50  std::vector<Half> newValues(info.GetNumElements());
51 
52  armnnUtils::FloatingPointConverter::ConvertFloat32To16(handle->GetConstTensor<float>(),
53  info.GetNumElements(),
54  newValues.data());
55 
56  TensorInfo newInfo(info.GetShape(), DataType::Float16, 0.0f, 0, true);
57  ConstTensor newInput(newInfo, newValues);
58  handle.reset(new ScopedTensorHandle(newInput));
59  }
60  }

References FloatingPointConverter::ConvertFloat32To16(), armnn::Float16, armnn::Float32, and armnn::info.


The documentation for this struct was generated from the following file:
armnn::DataType::Float32
@ Float32
armnn::DataType::Float16
@ Float16
armnn::BoostLogSeverityMapping::info
@ info
armnnUtils::FloatingPointConverter::ConvertFloat32To16
static void ConvertFloat32To16(const float *srcFloat32Buffer, size_t numElements, void *dstFloat16Buffer)
Converts a buffer of FP32 values to FP16, and stores in the given dstFloat16Buffer.
Definition: FloatingPointConverter.cpp:17