torch.geqrf — PyTorch 2.7 documentation (original) (raw)

torch.geqrf(input, *, out=None)¶

This is a low-level function for calling LAPACK’s geqrf directly. This function returns a namedtuple (a, tau) as defined in LAPACK documentation for geqrf .

Computes a QR decomposition of input. Both Q and R matrices are stored in the same output tensor a. The elements of R are stored on and above the diagonal. Elementary reflectors (or Householder vectors) implicitly defining matrix Qare stored below the diagonal. The results of this function can be used together with torch.linalg.householder_product()to obtain the Q matrix or with torch.ormqr(), which uses an implicit representation of the Q matrix, for an efficient matrix-matrix multiplication.

See LAPACK documentation for geqrf for further details.

Note

See also torch.linalg.qr(), which computes Q and R matrices, and torch.linalg.lstsq()with the driver="gels" option for a function that can solve matrix equations using a QR decomposition.

Parameters

input (Tensor) – the input matrix

Keyword Arguments

out (tuple, optional) – the output tuple of (Tensor, Tensor). Ignored if None. Default: None.