svdbpi

Contents

svdbpi#

Perform a singular value decomposition (SVD) using the block power iteration method. This method is usually better than svd where the matrices are small and batches are large

Added in version 0.6.0.

template<typename AType>
__MATX_INLINE__ auto matx::svdbpi(const AType &A, int max_iters = 10, float tol = 0.0f)#

Perform a SVD decomposition using the block power iteration. This version of SVD works well on small n/m with large batch.

Template Parameters:

AType – Tensor or operator type for output of A input tensors.

Parameters:

  • A – Input tensor or operator for tensor A input with size batches x m x n

  • max_iters – The approximate maximum number of QR iterations to perform.

  • tol – The termination tolerance for the QR iteration. Setting this to 0 will skip the tolerance check.

Examples#

auto A = make_tensor<AType>(Ashape);
auto U = make_tensor<AType>(Ushape);
auto VT = make_tensor<AType>(VTshape);
auto S = make_tensor<SType>(Sshape);

int iterations = 100;

(A = random<AType>(std::move(Ashape), NORMAL)).run(exec);


(U = 0).run(exec);
(S = 0).run(exec);
(VT = 0).run(exec);

(mtie(U, S, VT) = svdbpi(A, iterations)).run(exec);
Contents