ScalarQuantizationCompression interface

Package:

@azure/search-documents

Contains configuration options specific to the scalar quantization compression method used during indexing and querying.

Extends

BaseVectorSearchCompression

Properties

kind	Polymorphic discriminator, which specifies the different types this object can be
parameters	Contains the parameters specific to Scalar Quantization.

Inherited Properties

compressionName	The name to associate with this particular configuration.
rescoringOptions	Contains the options for rescoring.
truncationDimension	The number of dimensions to truncate the vectors to. Truncating the vectors reduces the size of the vectors and the amount of data that needs to be transferred during search. This can save storage cost and improve search performance at the expense of recall. It should be only used for embeddings trained with Matryoshka Representation Learning (MRL) such as OpenAI text-embedding-3-large (small). The default value is null, which means no truncation.

Property Details

kind

Polymorphic discriminator, which specifies the different types this object can be

kind: "scalarQuantization"

Property Value

"scalarQuantization"

parameters

Contains the parameters specific to Scalar Quantization.

parameters?: ScalarQuantizationParameters

Property Value

ScalarQuantizationParameters

Inherited Property Details

compressionName

The name to associate with this particular configuration.

compressionName: string

Property Value

string

Inherited From VectorSearchCompression.compressionName

rescoringOptions

Contains the options for rescoring.

rescoringOptions?: RescoringOptions

Property Value

RescoringOptions

Inherited From VectorSearchCompression.rescoringOptions

truncationDimension

The number of dimensions to truncate the vectors to. Truncating the vectors reduces the size of the vectors and the amount of data that needs to be transferred during search. This can save storage cost and improve search performance at the expense of recall. It should be only used for embeddings trained with Matryoshka Representation Learning (MRL) such as OpenAI text-embedding-3-large (small). The default value is null, which means no truncation.

truncationDimension?: number

Property Value

number

Inherited From VectorSearchCompression.truncationDimension