lucid.random.permutation¶

lucid.random.permutation(n: int, dtype: type[bool | int | float | complex] | ~lucid.types.Numeric = <class 'int'>, requires_grad: bool = False, keep_grad: bool = False, device: ~typing.Literal['cpu', 'gpu'] = 'cpu') → Tensor¶

The permutation function returns a random permutation of integers from \(0\) to \(n - 1\), without repetition. It produces a 1-dimensional tensor containing all integers in a shuffled order.

Function Signature¶

def permutation(
    n: int,
    dtype: type = np.int64,
    requires_grad: bool = False,
    keep_grad: bool = False,
    device: _DeviceType = "cpu",
) -> Tensor

Parameters¶

n (int): The number of elements to permute. The output tensor will contain all integers from 0 to n - 1, randomly shuffled.
dtype (type, optional): The data type of the resulting tensor. Defaults to lucid.Int64.
requires_grad (bool, optional): Whether the resulting tensor should track gradients. Defaults to False.
keep_grad (bool, optional): Whether to preserve gradient history during further operations. Defaults to False.
device ({ “cpu”, “gpu” }, optional): The device on which to place the resulting tensor. Defaults to “cpu”.

Returns¶

Tensor: A 1-dimensional tensor of length n, containing a random permutation of integers from 0 to n - 1.

\[\operatorname{shape}(\text{out}) = (n,)\]

Example¶

>>> import lucid
>>> lucid.random.permutation(5)
Tensor([3, 0, 2, 4, 1], grad=None)

The permutation contains all integers from 0 to 4 in a random order.

To place the tensor on GPU:

>>> lucid.random.permutation(6, device="gpu")
Tensor([4, 0, 2, 1, 5, 3], device=gpu)

Note

The permutation is generated using NumPy and transferred to the specified device.
Values are unique and cover the full range from 0 to \(n-1\).
This operation is non-differentiable, and gradients will not flow through the result.
For reproducibility, use lucid.random.seed() before calling this function.