make_clusters

# Import statement

from hottbox.datasets import make_clusters

make_clusters(dims, centers=5, n_samples=1000, center_bounds=(-10.0, 10.0), std=0.5, return_centers=False)

Generates a tensor of any dimension with isotropic gaussian blobs as clusters

Parameters

dimsint

specifies the order of the tensor

n_samplesint or list(int)

Specifies the size of each clusters

centersint or list(tuples)

The number of clusters in the dataset and their size (can be a list)

center_boundstuple(float, float)

Specifies the bound (min, max) for generating the centers

stdfloat

The standard deviation of each of the generated clusters

return_centersbool

Returns the positions of the centroids

Returns

tensor: Tensor

Generated tensor according to the parameters specified