diag

paddle. diag ( x, offset=0, padding_value=0, name=None ) [source]

If x is a vector (1-D tensor), a 2-D square tensor with the elements of x as the diagonal is returned.

If x is a matrix (2-D tensor), a 1-D tensor with the diagonal elements of x is returned.

The argument offset controls the diagonal offset:

If offset = 0, it is the main diagonal.

If offset > 0, it is superdiagonal.

If offset < 0, it is subdiagonal.

Parameters

  • x (Tensor) – The input tensor. Its shape is either 1-D or 2-D. Its data type should be float16, float32, float64, int32, int64, complex64, complex128.

  • offset (int, optional) – The diagonal offset. A positive value represents superdiagonal, 0 represents the main diagonal, and a negative value represents subdiagonal.

  • padding_value (int|float, optional) – Use this value to fill the area outside the specified diagonal band. Only takes effect when the input is a 1-D Tensor. The default value is 0.

  • name (str, optional) – For details, please refer to Name. Generally, no setting is required. Default: None.

Returns

Tensor, a square matrix or a vector. The output data type is the same as input data type.

Examples

>>> import paddle

>>> paddle.disable_static()
>>> x = paddle.to_tensor([1, 2, 3])
>>> y = paddle.diag(x)
>>> print(y)
Tensor(shape=[3, 3], dtype=int64, place=Place(cpu), stop_gradient=True,
[[1, 0, 0],
 [0, 2, 0],
 [0, 0, 3]])

>>> y = paddle.diag(x, offset=1)
>>> print(y)
Tensor(shape=[4, 4], dtype=int64, place=Place(cpu), stop_gradient=True,
[[0, 1, 0, 0],
 [0, 0, 2, 0],
 [0, 0, 0, 3],
 [0, 0, 0, 0]])

>>> y = paddle.diag(x, padding_value=6)
>>> print(y)
Tensor(shape=[3, 3], dtype=int64, place=Place(cpu), stop_gradient=True,
[[1, 6, 6],
 [6, 2, 6],
 [6, 6, 3]])

>>> import paddle

>>> paddle.disable_static()
>>> x = paddle.to_tensor([[1, 2, 3], [4, 5, 6]])
>>> y = paddle.diag(x)
>>> print(y)
Tensor(shape=[2], dtype=int64, place=Place(cpu), stop_gradient=True,
[1, 5])

>>> y = paddle.diag(x, offset=1)
>>> print(y)
Tensor(shape=[2], dtype=int64, place=Place(cpu), stop_gradient=True,
[2, 6])

>>> y = paddle.diag(x, offset=-1)
>>> print(y)
Tensor(shape=[1], dtype=int64, place=Place(cpu), stop_gradient=True,
[4])