SVD gives different results from Python with MWE

mkitti · March 9, 2021, 2:32pm

Having exhausted other options, you can try to reverse engineer Numpy to see if there are preprocessing the matrix somehow:

github.com

numpy/numpy/blob/v1.20.0/numpy/linalg/linalg.py#L1482-L1674


      
          @array_function_dispatch(_svd_dispatcher)
          def svd(a, full_matrices=True, compute_uv=True, hermitian=False):
              """
              Singular Value Decomposition.
          
          
    When `a` is a 2D array, it is factorized as ``u @ np.diag(s) @ vh
              = (u * s) @ vh``, where `u` and `vh` are 2D unitary arrays and `s` is a 1D
              array of `a`'s singular values. When `a` is higher-dimensional, SVD is
              applied in stacked mode as explained below.
          
          
    Parameters
              ----------
              a : (..., M, N) array_like
                  A real or complex array with ``a.ndim >= 2``.
              full_matrices : bool, optional
                  If True (default), `u` and `vh` have the shapes ``(..., M, M)`` and
                  ``(..., N, N)``, respectively.  Otherwise, the shapes are
                  ``(..., M, K)`` and ``(..., K, N)``, respectively, where
                  ``K = min(M, N)``.
              compute_uv : bool, optional

This file has been truncated. show original

It looks like they use a subset of LAPACK processed through f2c in the lapack_lite package.

Topic		Replies	Views
Julia SVD results differing from MATLAB Numerics linearalgebra	5	1048	February 2, 2021
Higher order SVD in Julia like Numpy? General Usage linearalgebra	2	526	July 30, 2022
SVD 2x slower than in Matlab and how to get best performance on Windows10 Performance	20	3440	October 9, 2024
`svd` function: difference between the show instance and the output New to Julia linearalgebra	1	304	October 31, 2020
LapackException(1) while svd but not svdvals General Usage linearalgebra	5	3111	May 2, 2019

SVD gives different results from Python with MWE

Related topics