Preprocessing Quantize Binary

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.

1 enum cuvsBinaryQuantizerThreshold { ... };

Values

Name	Value
`ZERO`	`0`
`MEAN`	`1`
`SAMPLING_MEDIAN`	`2`

cuvsBinaryQuantizerParams

Binary quantizer parameters.

1 struct cuvsBinaryQuantizerParams { ... };

Fields

Name	Type	Description
`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

1 CUVS_EXPORT cuvsError_t cuvsBinaryQuantizerParamsCreate(cuvsBinaryQuantizerParams_t* params);

Parameters

Name	Direction	Type	Description
`params`	in	`cuvsBinaryQuantizerParams_t*`	cuvsBinaryQuantizerParams_t to allocate

Returns

CUVS_EXPORT cuvsError_t

cuvsBinaryQuantizerParamsDestroy

De-allocate Binary Quantizer params

1 CUVS_EXPORT cuvsError_t cuvsBinaryQuantizerParamsDestroy(cuvsBinaryQuantizerParams_t params);

Parameters

Name	Direction	Type	Description
`params`	in	`cuvsBinaryQuantizerParams_t`

Returns

CUVS_EXPORT 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.

1 typedef struct { ... } cuvsBinaryQuantizer;

Fields

Name	Type	Description
`addr`	`uintptr_t`
`dtype`	`DLDataType`

cuvsBinaryQuantizerCreate

Allocate Binary Quantizer and populate with default values

1 CUVS_EXPORT cuvsError_t cuvsBinaryQuantizerCreate(cuvsBinaryQuantizer_t* quantizer);

Parameters

Name	Direction	Type	Description
`quantizer`	in	`cuvsBinaryQuantizer_t*`	cuvsBinaryQuantizer_t to allocate

Returns

CUVS_EXPORT cuvsError_t

cuvsBinaryQuantizerDestroy

De-allocate Binary Quantizer

1 CUVS_EXPORT cuvsError_t cuvsBinaryQuantizerDestroy(cuvsBinaryQuantizer_t quantizer);

Parameters

Name	Direction	Type	Description
`quantizer`	in	`cuvsBinaryQuantizer_t`

Returns

CUVS_EXPORT cuvsError_t

cuvsBinaryQuantizerTrain

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

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

Parameters

Name	Direction	Type	Description
`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

CUVS_EXPORT cuvsError_t

cuvsBinaryQuantizerTransform

Applies binary quantization transform to the given dataset

1 CUVS_EXPORT cuvsError_t cuvsBinaryQuantizerTransform(cuvsResources_t res,
2 DLManagedTensor* dataset,
3 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

Name	Direction	Type	Description
`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

CUVS_EXPORT cuvsError_t

cuvsBinaryQuantizerTransformWithParams

Applies binary quantization transform to the given dataset

1 CUVS_EXPORT cuvsError_t cuvsBinaryQuantizerTransformWithParams(cuvsResources_t res,
2 cuvsBinaryQuantizer_t quantizer,
3 DLManagedTensor* dataset,
4 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

Name	Direction	Type	Description
`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

CUVS_EXPORT cuvsError_t