Struct SvdResult 

pub struct SvdResult<T: Scalar> {
    pub u: Tensor<T>,
    pub s: Tensor<T>,
    pub vt: Tensor<T>,
}

SVD result: A = U * diag(S) * Vt.

For an input of shape (m, n, *) with k = min(m, n):

• u: shape (m, k, *)
• s: shape (k, *) (singular values, descending order)
• vt: shape (k, n, *)

Examples

use tenferro_linalg::svd;
use tenferro_tensor::{Tensor, MemoryOrder};
use tenferro_device::LogicalMemorySpace;

let a = Tensor::<f64>::zeros(&[3, 4],
    LogicalMemorySpace::MainMemory, MemoryOrder::ColumnMajor);
let result = svd(&a, None).unwrap();
assert_eq!(result.s.ndim(), 1);

Fields

u: Tensor<T>

Left singular vectors. Shape: (m, k, *).

s: Tensor<T>

Singular values (descending order). Shape: (k, *).

vt: Tensor<T>

Right singular vectors (conjugate-transposed). Shape: (k, n, *).