mindspore.ops.ormqr

View Source On AtomGit
mindspore.ops.ormqr(input, tau, other, left=True, transpose=False)[source]

Calculates the product of a matrix other and a matrix Q (represented by Householder vectors input and Householder reflection coefficients tau).

Warning

After version 2.9.0, the parameters tau and other will be renamed to input2 and input3, and an out parameter will be added.

If left is True, computes Q * other , otherwise, compute other * Q.

Parameters

  • input (Tensor) – The input tensor, shape \((*, mn, k)\), when left is True, mn is equal to m, otherwise, mn is equal to n.

  • tau (Tensor) – The input tensor, shape \((*, min(mn, k))\) .

  • other (Tensor) – The input tensor, shape \((*, m, n)\) .

  • left (bool, optional) – The order of computation. Default True .

  • transpose (bool, optional) – Whether the matrix Q is conjugate transposed or not. Default False .

Returns

Tensor

Supported Platforms:

GPU

Examples

>>> import mindspore
>>> input = mindspore.tensor(([[-114.6, 10.9, 1.1], [-0.304, 38.07, 69.38], [-0.45, -0.17, 62]]),
...                             mindspore.float32)
>>> tau = mindspore.tensor(([1.55, 1.94, 3.0]), mindspore.float32)
>>> other = mindspore.tensor(([[-114.6, 10.9, 1.1],
...                            [-0.304, 38.07, 69.38],
...                            [-0.45, -0.17, 62]]), mindspore.float32)
>>> output = mindspore.ops.ormqr(input, tau, other)
>>> print(output)
[[  63.82713   -13.823125 -116.28614 ]
 [ -53.659264  -28.157839  -70.42702 ]
 [ -79.54292    24.00183   -41.34253 ]]