Distance | cuVS

Python module: cuvs.distance

DISTANCE_NAMES

1 DISTANCE_NAMES = {v: k for k, v in DISTANCE_TYPES.items()}

DISTANCE_TYPES

1 DISTANCE_TYPES = {
2 "l2": cuvsDistanceType.L2SqrtExpanded,
3 "sqeuclidean": cuvsDistanceType.L2Expanded,
4 "euclidean": cuvsDistanceType.L2SqrtExpanded,
5 "l1": cuvsDistanceType.L1,
6 "cityblock": cuvsDistanceType.L1,
7 "inner_product": cuvsDistanceType.InnerProduct,
8 "chebyshev": cuvsDistanceType.Linf,
9 "canberra": cuvsDistanceType.Canberra,
10 "cosine": cuvsDistanceType.CosineExpanded,
11 "lp": cuvsDistanceType.LpUnexpanded,
12 "correlation": cuvsDistanceType.CorrelationExpanded,
13 "jaccard": cuvsDistanceType.JaccardExpanded,
14 "hellinger": cuvsDistanceType.HellingerExpanded,
15 "braycurtis": cuvsDistanceType.BrayCurtis,
16 "jensenshannon": cuvsDistanceType.JensenShannon,
17 "hamming": cuvsDistanceType.HammingUnexpanded,
18 "kl_divergence": cuvsDistanceType.KLDivergence,
19 "minkowski": cuvsDistanceType.LpUnexpanded,
20 "russellrao": cuvsDistanceType.RusselRaoExpanded,
21 "dice": cuvsDistanceType.DiceExpanded,
22 "bitwise_hamming": cuvsDistanceType.BitwiseHamming
23 }

pairwise_distance

@auto_sync_resources @auto_convert_output

1 def pairwise_distance(X, Y, out=None, metric="euclidean", p=2.0, resources=None)

Compute pairwise distances between X and Y

Valid values for metric: [“euclidean”, “l2”, “l1”, “cityblock”, “inner_product”, “chebyshev”, “canberra”, “lp”, “hellinger”, “jensenshannon”, “kl_divergence”, “russellrao”, “minkowski”, “correlation”, “cosine”]

Parameters

Name	Type	Description
`X`	`CUDA array interface compliant matrix shape (m, k)`
`Y`	`CUDA array interface compliant matrix shape (n, k)`
`out`	`Optional writable CUDA array interface matrix shape (m, n)`
`metric`	`string denoting the metric type (default="euclidean")`
`p`	`metric parameter (currently used only for "minkowski")`
`resources`	`cuvs.common.Resources, optional`

Examples

1 >>> import cupy as cp
2 >>> from cuvs.distance import pairwise_distance
3 >>> n_samples = 5000
4 >>> n_features = 50
5 >>> in1 = cp.random.random_sample((n_samples, n_features),
6 ...                               dtype=cp.float32)
7 >>> in2 = cp.random.random_sample((n_samples, n_features),
8 ...                               dtype=cp.float32)
9 >>> output = pairwise_distance(in1, in2, metric="euclidean")