FP16 to FP32 precision utilities

Set of functions to convert from FP32 to FP16 and vice versa. More...

Typedefs
using	InferenceEngine::ie_fp16 = short
	A type difinition for FP16 data type. Defined as a singed short.

Functions
ie_fp16	InferenceEngine::PrecisionUtils::f32tof16 (float x)
	Converts a single-precision floating point value to a half-precision floating poit value. More...
float	InferenceEngine::PrecisionUtils::f16tof32 (ie_fp16 x)
	Convers a half-precision floating point value to a single-precision floating point value. More...
void	InferenceEngine::PrecisionUtils::f16tof32Arrays (float *dst, const ie_fp16 *src, size_t nelem, float scale=1.f, float bias=0.f)
	Converts a half-precision floating point array to single-precision floating point array and applies scale and bias is needed. More...
void	InferenceEngine::PrecisionUtils::f32tof16Arrays (ie_fp16 *dst, const float *src, size_t nelem, float scale=1.f, float bias=0.f)
	Converts a single-precision floating point array to a half-precision floating point array and applies scale and bias if needed. More...
template<class OutT , class InT , typename std::enable_if< std::is_integral< OutT >::value &&std::is_integral< InT >::value &&std::is_signed< InT >::value &&!std::is_same< OutT, InT >::value >::type * = nullptr>
OutT	InferenceEngine::PrecisionUtils::saturate_cast (const InT &value)
	Converts one integral type to another saturating the result if the source value doesn't fit into destination type range. More...
template<class InT >
InT	InferenceEngine::PrecisionUtils::saturate_cast (const InT &value)
	Converts one integral type to another saturating the result if the source value doesn't fit into destination type range. More...

Detailed Description

Set of functions to convert from FP32 to FP16 and vice versa.

Function Documentation

f16tof32()

float InferenceEngine::PrecisionUtils::f16tof32

(

ie_fp16

x

)

Convers a half-precision floating point value to a single-precision floating point value.

Parameters

[in]

x

A half-precision floating point value

Returns

A single-precision floating point value

f16tof32Arrays()

void InferenceEngine::PrecisionUtils::f16tof32Arrays	(	float *	dst,
		const ie_fp16 *	src,
		size_t	nelem,
		float	scale = 1.f,
		float	bias = 0.f
	)

Converts a half-precision floating point array to single-precision floating point array and applies scale and bias is needed.

Parameters

	dst	A destination array of single-precision floating point values
[in]	src	A source array of half-precision floating point values
[in]	nelem	A number of elements in arrays
[in]	scale	An optional scale parameter
[in]	bias	An optional bias parameter

f32tof16()

ie_fp16 InferenceEngine::PrecisionUtils::f32tof16

(

float

x

)

Converts a single-precision floating point value to a half-precision floating poit value.

Parameters

[in]

x

A single-precision floating point value

Returns

A half-precision floating point value

f32tof16Arrays()

void InferenceEngine::PrecisionUtils::f32tof16Arrays	(	ie_fp16 *	dst,
		const float *	src,
		size_t	nelem,
		float	scale = 1.f,
		float	bias = 0.f
	)

Converts a single-precision floating point array to a half-precision floating point array and applies scale and bias if needed.

Parameters

	dst	A destination array of half-precision floating point values
[in]	src	A sources array of single-precision floating point values
[in]	nelem	A number of elements in arrays
[in]	scale	An optional scale parameter
[in]	bias	An optional bias parameter

saturate_cast() [1/2]

template<class OutT , class InT , typename std::enable_if< std::is_integral< OutT >::value &&std::is_integral< InT >::value &&std::is_signed< InT >::value &&!std::is_same< OutT, InT >::value >::type * = nullptr>

OutT InferenceEngine::PrecisionUtils::saturate_cast ( const InT &  value )

inline

Converts one integral type to another saturating the result if the source value doesn't fit into destination type range.

Parameters

value

Value to be converted

Returns

A saturated value

saturate_cast() [2/2]

template<class InT >

InT InferenceEngine::PrecisionUtils::saturate_cast ( const InT &  value )

inline

Converts one integral type to another saturating the result if the source value doesn't fit into destination type range.

Parameters

value

Value to be converted

Returns

A saturated value