add

→Tensor

add(input: Tensor, other: Tensor | Scalar)

source edit

Implementing kernel

Element-wise addition with broadcasting and dtype promotion.

Computes input + other element-wise. Scalars are promoted to a tensor of matching dtype; tensors with different shapes are broadcast following the standard rules. Mixed-dtype operands are promoted to their common type prior to the operation.

Parameters

inputTensor

Left operand.

otherTensor or scalar

Right operand. Broadcasts against input following the standard broadcasting rules; Python scalars are promoted to a tensor of matching dtype.

Returns

Tensor

Element-wise result with shape broadcast(input.shape, other.shape) and dtype determined by the usual type-promotion rules.

Notes

Mathematical definition:

\text{out}_i = \text{input}_i + \text{other}_i

Autograd: gradient flows back to both operands as the identity ( $\partial \text{out}/\partial \text{input} = 1$ , $\partial \text{out}/\partial \text{other} = 1$ ), with broadcasted dimensions summed.

Examples

>>> import lucid
>>> a = lucid.tensor([1.0, 2.0, 3.0])
>>> b = lucid.tensor([4.0, 5.0, 6.0])
>>> lucid.add(a, b)
Tensor([...])

add