probly.representation.distribution.mixture_dirichlet¶

Mixture distribution for Dirichlet components.

Classes

ArrayDirichletMixture(components, weights)

Dirichlet mixture distribution.

class probly.representation.distribution.mixture_dirichlet.ArrayDirichletMixture(components, weights)[source]¶

Bases: object

Dirichlet mixture distribution.

Parameters:

components (Sequence[ArrayDirichlet])
weights (np.ndarray)

sample(num_samples, rng=None)[source]¶

Sample from the mixture.

Returns:¶

ArraySample: Samples with shape: (num_samples, *alphas.shape).

Parameters:

num_samples (int)
rng (Generator | None)

Return type:

property T: Self¶: The transposed version of the distribution.

components: Sequence[ArrayDirichlet]¶

property device: str¶: The device of the underlying array.

property dtype: DTypeLike¶: The data type of the underlying array.

property ndim: int¶: Number of batch dimensions (excluding category axis).

property shape: tuple[int, ...]¶: Batch shape (excluding category axis).

property size: int¶: The total number of distributions.

weights: np.ndarray¶