uniform_op

→TensorImplPtr

int uniform_op(const int & shape, double low, double high, Dtype dt, Device device, Generator * gen)

source edit

Python wrappers

pylucid.nn.init.uniformfunction

Allocates a tensor filled with uniform samples on $[\mathrm{low}, \mathrm{high})$ .

Math

$x_i \sim \mathcal{U}(\mathrm{low}, \mathrm{high})$

See Also

rand_op — convenience shortcut for $[0, 1)$ .

Parameters

shapeconst Shape&

Shape of the output tensor.

lowdouble

Inclusive lower bound of the support.

highdouble

Exclusive upper bound of the support. Must be strictly greater than low.

dtDtype

Element dtype. Floating-point only.

deviceDevice

Memory domain (CPU or GPU).

genGenerator*= None

Bit-stream state; falls back to default_generator on nullptr.

Returns

TensorImplPtr

Newly allocated leaf tensor (requires_grad = false).

Raises

LucidError

If high <= low — surfaced as ValueError on the Python side.