mindspore.ops.diag

mindspore.ops.diag(input)[source]

Return a tensor with the input as its diagonal elements and 0 elsewhere.

Warning

This is an experimental API that is subject to change or deletion.

Parameters: input (Tensor) – The input tensor.
Returns: Tensor

Supported Platforms:: Ascend GPU CPU

Examples

>>> import mindspore
>>> # case 1: When input is a 1-D tensor:
>>> input = mindspore.ops.randn(3)
>>> output = mindspore.ops.diag(input)
>>> print(output)
[[ 1.7477764  0.         0.       ]
 [ 0.        -1.2616369  0.       ]
 [ 0.         0.         2.3283238]]
>>>
>>> # case 2: When input is a multi-dimensional tensor:
>>> input = mindspore.ops.randn(2, 3)
>>> print(input)
[[ 0.21546374 -0.0120403  -0.7330481 ]
 [-2.5405762   0.44775972 -1.4063131 ]]
>>> output = mindspore.ops.diag(input)
>>> print(output)
[[[[ 0.21546374  0.          0.        ]
   [ 0.          0.          0.        ]]
  [[ 0.         -0.0120403   0.        ]
   [ 0.          0.          0.        ]]
  [[ 0.          0.         -0.7330481 ]
   [ 0.          0.          0.        ]]]
 [[[ 0.          0.          0.        ]
   [-2.5405762   0.          0.        ]]
  [[ 0.          0.          0.        ]
   [ 0.          0.44775972  0.        ]]
  [[ 0.          0.          0.        ]
   [ 0.          0.         -1.4063131 ]]]]
>>> # Assume input has dimensions (D_1,... D_k), the output is a tensor of rank 2k with dimensions
    (D_1,..., D_k, D_1,..., D_k).
>>> print(output.shape)
(2, 3, 2, 3)