sakura.utils.distributions.rand_ring2d

sakura.utils.distributions.rand_ring2d(batch_size, n_dim=2)

Generates 2D samples from a hollowed-circle (ring) distribution using sklearn.make_circles.

Samples are drawn with uniform radial distance from the origin.

Parameters:

• batch_size (int) – Number of samples to generate

• n_dim (Literal[2], optional) – Dimension of output samples, should be 2

Returns:

Tensor of shape (batch_size, 2) with ring samples

Return type:

torch.FloatTensor