tensor.nonzero
fn nonzero(self: @Tensor<T>) -> Tensor<usize>;
Produces indices of the elements that are non-zero (in row-major order - by dimension).
Args
self
(@Tensor<T>
) - Tensor of data to calculate non-zero indices.
Returns
A new Tensor<usize>
indices of the elements that are non-zero (in row-major order - by dimension).
Example
use core::array::{ArrayTrait, SpanTrait};
use orion::operators::tensor::{TensorTrait, Tensor, U32Tensor};
fn nonzero_example() -> Tensor<u32> {
let tensor = TensorTrait::<u32>::new(
shape: array![2, 4].span(),
data: array![0, 1, 2, 3, 4, 5, 6, 7].span(),
);
return tensor.nonzero();
}
>>> [[0 0 0 1 1 1 1]
[1 2 3 0 1 2 3]]
Last updated
Was this helpful?