binomial

deeptrack.backend.array_api_compat_ext.torch.random.binomial(n: int | Tensor, p: float | Tensor, size: tuple[int, ...] | None = None) → Tensor

Sample from a Binomial distribution.

Mirrors numpy.random.binomial, including support for tensor parameters and broadcasting.

Parameters

n: int | torch.Tensor

Number of trials.

p: float | torch.Tensor

Probability of success.

size: tuple[int, …] | None, optional

Sample shape. If None, returns samples with the broadcasted parameter shape.

Returns

torch.Tensor

Samples drawn from Binomial(n, p). Output shape is size + batch_shape. The returned dtype is torch.int64 to match NumPy parity.

Examples

>>> import deeptrack.backend.array_api_compat_ext.torch.random as rnd

>>> rnd.binomial(10, 0.5)
tensor(6.)

>>> rnd.binomial(10, 0.5).dtype
torch.int64

>>> rnd.binomial(10, 0.5, (2, 3)).shape
torch.Size([2, 3])