probly.representation.distribution.array_categorical.ArrayCategoricalDistribution

class probly.representation.distribution.array_categorical.ArrayCategoricalDistribution(unnormalized_probabilities: ndarray = <property object>)

Bases: ArrayAxisProtected[ndarray], CategoricalDistribution

A categorical distribution stored as a numpy array.

Shape: (…, num_classes) The last axis represents the category dimension.

property T: Self

The transposed version of the underlying array.

astype(dtype: DTypeLike, order: Order = 'K', casting: Literal['no', 'equiv', 'safe', 'same_kind', 'unsafe'] = 'unsafe', subok: bool = True, copy: bool = True) → Self

Copy of the array, cast to a specified type.

copy(order: Order = 'C') → Self

Return a copy of the array.

property device: str

Device of the array.

property dtype: dtype

Data type of the array.

entropy() → ArrayLike

Compute entropy.

property flags: ArrayFlagsLike

The flags of the array.

property mT: Self

The transposed version of the underlying array.

property ndim: int

Number of dimensions.

property num_classes: int

Get the number of classes.

permitted_functions: ClassVar[set[Callable]] = {<function average>, <function mean>}

permitted_ufuncs = {}

classmethod primary_protected_name() → str

Return the first protected field (dict order).

property probabilities: ndarray

Get the normalized probabilities of the categorical distribution.

protected_axes: ClassVar[dict[str, int]] = {'unnormalized_probabilities': 1}

property protected_shape: tuple[int, ...]

Protected trailing shape of the primary field.

protected_value() → ArrayProtectedValue

Return the primary protected value.

protected_values(func: Callable | None = None, method: str | None = None) → dict[str, ArrayProtectedValue] | None

Return all protected field values.

The values are preserved as-is and are not coerced to np.ndarray. Optionally takes the function that triggered the call for context. This can be used to conditionally modify the returned values or prevent them from being accessed.

reshape(*shape: int | tuple[int, ...], order: str = 'C', copy: bool | None = None) → Self

Return a copy with reshaped protected values.

sample(num_samples: int = 1, rng: Generator | None = None) → ArraySample[ndarray]

Sample from the categorical distribution (NumPy backend).

property shape: tuple[int, ...]

Shape of the array.

property size: int

Number of elements in the array.

to_device(device: Any, /, *, stream: Any = None) → Self

Move the array to a device.

transpose(*axes: int | None) → Self

Return a transposed version of the ArraySample.

This method implicitly also provides full axis tracking support for - np.moveaxis - np.rollaxis Those functions call out to transpose methods for custom array types.

Parameters:

axes – The axes to transpose.

Returns:

A transposed version of the ArraySample.

type = 'categorical'

unnormalized_probabilities: np.ndarray

with_protected_value(value: ArrayProtectedValue) → Self

Return a copy with a replaced primary protected value.

with_protected_values(values: dict[str, ArrayProtectedValue]) → Self

Return a copy with updated protected field values.