Skip to contents

Broadcast Tensors to a Common Shape

Source: R/api.R
nv_broadcast_tensors.Rd

Broadcast tensors to a common shape.

Usage

nv_broadcast_tensors(...)

Arguments

...

(nv_tensor)
Tensors to broadcast.

Value

(list() of nv_tensor)

Broadcasting Rules

We follow the standard NumPy broadcasting rules:

  1. If the tensors have different numbers of dimensions, prepend 1s to the shape of the smaller tensor.

  2. For each dimension, if:

    • the sizes are the same, do nothing.

    • one of the tensors has size 1, expand it to the corresponding size of the other tensor.

    • the sizes are different and neither is 1, raise an error.