极速阅览

Source header: cuvs/preprocessing/quantize/binary.h

C API for Binary Quantizer

cuvsBinaryQuantizerThreshold

In the cuvsBinaryQuantizerTransform function, a bit is set if the corresponding element in the dataset vector is greater than the corresponding element in the threshold vector. The mean and sampling_median thresholds are calculated separately for each dimension.

enum cuvsBinaryQuantizerThreshold {
  ZERO = 0,
  MEAN = 1,
  SAMPLING_MEDIAN = 2
};

Values

NameValue

ZERO	0
MEAN	1
SAMPLING_MEDIAN	2

cuvsBinaryQuantizerParams

Binary quantizer parameters.

struct cuvsBinaryQuantizerParams {
  /* * specifies the threshold to set a bit in cuvsBinaryQuantizerTransform */ enum cuvsBinaryQuantizerThreshold threshold;
  /* * specifies the sampling ratio */ float sampling_ratio;
};

Fields

NameTypeDescription

threshold	/* * specifies the threshold to set a bit in cuvsBinaryQuantizerTransform */ enum cuvsBinaryQuantizerThreshold
sampling_ratio	/* * specifies the sampling ratio */ float

cuvsBinaryQuantizerParamsCreate

Allocate Binary Quantizer params, and populate with default values

cuvsError_t cuvsBinaryQuantizerParamsCreate(cuvsBinaryQuantizerParams_t* params);

Parameters

NameDirectionTypeDescription

params

cuvsBinaryQuantizerParams_t*

cuvsBinaryQuantizerParams_t to allocate

Returns

cuvsError_t

cuvsBinaryQuantizerParamsDestroy

De-allocate Binary Quantizer params

cuvsError_t cuvsBinaryQuantizerParamsDestroy(cuvsBinaryQuantizerParams_t params);

Parameters

NameDirectionTypeDescription

params

cuvsBinaryQuantizerParams_t

Returns

cuvsError_t

cuvsBinaryQuantizer

Defines and stores threshold for quantization upon training

The quantization is performed by a linear mapping of an interval in the float data type to the full range of the quantized int type.

typedef struct {
  uintptr_t addr;
  DLDataType dtype;
} cuvsBinaryQuantizer;

Fields

NameTypeDescription

addr	uintptr_t
dtype	DLDataType

cuvsBinaryQuantizerCreate

Allocate Binary Quantizer and populate with default values

1	cuvsError_t cuvsBinaryQuantizerCreate(cuvsBinaryQuantizer_t* quantizer);

Parameters

NameDirectionTypeDescription

quantizer

cuvsBinaryQuantizer_t*

cuvsBinaryQuantizer_t to allocate

Returns

cuvsError_t

cuvsBinaryQuantizerDestroy

De-allocate Binary Quantizer

1	cuvsError_t cuvsBinaryQuantizerDestroy(cuvsBinaryQuantizer_t quantizer);

Parameters

NameDirectionTypeDescription

quantizer

cuvsBinaryQuantizer_t

Returns

cuvsError_t

cuvsBinaryQuantizerTrain

Trains a binary quantizer to be used later for quantizing the dataset.

cuvsError_t cuvsBinaryQuantizerTrain(cuvsResources_t res,
cuvsBinaryQuantizerParams_t params,
DLManagedTensor* dataset,
cuvsBinaryQuantizer_t quantizer);

Parameters

NameDirectionTypeDescription

res	in	cuvsResources_t	raft resource
params	in	cuvsBinaryQuantizerParams_t	configure binary quantizer, e.g. threshold
dataset	in	DLManagedTensor*	a row-major host or device matrix
quantizer	out	cuvsBinaryQuantizer_t	trained binary quantizer

Returns

cuvsError_t

cuvsBinaryQuantizerTransform

Applies binary quantization transform to the given dataset

cuvsError_t cuvsBinaryQuantizerTransform(cuvsResources_t res,
DLManagedTensor* dataset,
DLManagedTensor* out);

This applies binary quantization to a dataset, changing any positive values to a bitwise 1. This is useful for searching with the BitwiseHamming distance type.

Parameters

NameDirectionTypeDescription

res	in	cuvsResources_t	raft resource
dataset	in	DLManagedTensor*	a row-major host or device matrix to transform
out	out	DLManagedTensor*	a row-major host or device matrix to store transformed data

Returns

cuvsError_t

cuvsBinaryQuantizerTransformWithParams

Applies binary quantization transform to the given dataset

cuvsError_t cuvsBinaryQuantizerTransformWithParams(cuvsResources_t res,
cuvsBinaryQuantizer_t quantizer,
DLManagedTensor* dataset,
DLManagedTensor* out);

This applies binary quantization to a dataset, changing any values that are larger than the threshold specified in the param to a bitwise 1. This is useful for searching with the BitwiseHamming distance type.

Parameters

NameDirectionTypeDescription

res	in	cuvsResources_t	raft resource
quantizer	in	cuvsBinaryQuantizer_t	binary quantizer
dataset	in	DLManagedTensor*	a row-major host or device matrix to transform
out	out	DLManagedTensor*	a row-major host or device matrix to store transformed data

Returns

cuvsError_t

Binary

C API for Binary Quantizer

cuvsBinaryQuantizerThreshold

cuvsBinaryQuantizerParams

cuvsBinaryQuantizerParamsCreate

cuvsBinaryQuantizerParamsDestroy

cuvsBinaryQuantizer

cuvsBinaryQuantizerCreate

cuvsBinaryQuantizerDestroy

cuvsBinaryQuantizerTrain

cuvsBinaryQuantizerTransform

cuvsBinaryQuantizerTransformWithParams

Neighbors Vamana

Preprocessing PCA

1	enum cuvsBinaryQuantizerThreshold {
2	ZERO = 0,
3	MEAN = 1,
4	SAMPLING_MEDIAN = 2
5	};

1	struct cuvsBinaryQuantizerParams {
2	/* * specifies the threshold to set a bit in cuvsBinaryQuantizerTransform */ enum cuvsBinaryQuantizerThreshold threshold;
3	/* * specifies the sampling ratio */ float sampling_ratio;
4	};

1	typedef struct {
2	uintptr_t addr;
3	DLDataType dtype;
4	} cuvsBinaryQuantizer;

1	cuvsError_t cuvsBinaryQuantizerTrain(cuvsResources_t res,
2	cuvsBinaryQuantizerParams_t params,
3	DLManagedTensor* dataset,
4	cuvsBinaryQuantizer_t quantizer);

1	cuvsError_t cuvsBinaryQuantizerTransform(cuvsResources_t res,
2	DLManagedTensor* dataset,
3	DLManagedTensor* out);

1	cuvsError_t cuvsBinaryQuantizerTransformWithParams(cuvsResources_t res,
2	cuvsBinaryQuantizer_t quantizer,
3	DLManagedTensor* dataset,
4	DLManagedTensor* out);