Preprocessing Quantize Binary

Python module: cuvs.preprocessing.quantize.binary

transform

@auto_sync_resources @auto_convert_output

1 def transform(dataset, output=None, resources=None)

Applies binary quantization transform to given dataset

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	Type	Description
`dataset`	`row major host or device dataset to transform`
`output`	`optional preallocated output memory, on host or device memory`
`resources`	`cuvs.common.Resources, optional`

Returns

Name	Type	Description
`output`	`transformed dataset quantized into a uint8`

Examples

1 >>> import cupy as cp
2 >>> from cuvs.preprocessing.quantize import binary
3 >>> from cuvs.neighbors import cagra
4 >>> n_samples = 50000
5 >>> n_features = 50
6 >>> dataset = cp.random.standard_normal((n_samples, n_features),
7 ...                                   dtype=cp.float32)
8 >>> transformed = binary.transform(dataset)
9 >>>
10 >>> # build a cagra index on the binarized data
11 >>> params = cagra.IndexParams(metric="bitwise_hamming",
12 ...                            build_algo="iterative_cagra_search")
13 >>> idx = cagra.build(params, transformed)