Method SVD

static member SVD: Tensor<'T> -> Tensor<'T> * Tensor<'T> * Tensor<'T>

Computes the (batched) singular value decomposition (SVD) of the specified matrix.

Declaration
static member SVD: a:Tensor<'T> -> Tensor<'T> * Tensor<'T> * Tensor<'T>
Parameters
Type Name Description
Tensor<'T> a

The input matrix or tensor to this operation.
Returns
Type Description
Tensor<'T> * Tensor<'T> * Tensor<'T>

A tuple consisting of (U, S, V) where S is a vector consisting of the singular values and U, V are the transformation matrices.
Remarks

The singular value decomposition of matrix a is computed. It is defined by the property that a = U .* Tensor.diagMat(S) .* V.T.

If a is a tensor, the operation is batched over the matrices consisting of the last two dimensions.

See Also