mindspore.ops.tile

mindspore.ops.tile(input, dims)

Creates a new tensor by repeating the elements in the input tensor dims times.

The i'th dimension of output tensor has input.shape[i] * dims[i] elements, and the values of input are repeated dims[i] times along the i'th dimension.

Note

• On Ascend, the number of dims should not exceed 8, and currently does not support scenarios where more than 4 dimensions are repeated simultaneously.

• If input.dim = d, then the shape of their corresponding positions can be multiplied, and the shape of Outputs is $(x_1\ast y_1,x_2\ast y_2,...,x_S\ast y_S)$.

• If input.dim < d, prepend 1 to the shape of input until their lengths are consistent. Such as set the shape of input as $(1,...,x_1,x_2,...,x_S)$, then the shape of their corresponding positions can be multiplied, and the shape of Outputs is $(1\ast y_1,...,x_R\ast y_R,x_S\ast y_S)$.

• If input.dim > d, prepend 1 to dims until their lengths are consistent. Such as set the dims as $(1,...,y_1,y_2,...,y_S)$, then the shape of their corresponding positions can be multiplied, and the shape of Outputs is $(x_1\ast 1,...,x_R\ast y_R,x_S\ast y_S)$.

Parameters

• input (Tensor) – The input tensor.

• dims (tuple[int]) – The specified number of repetitions in each dimension.

Returns

Tensor

Supported Platforms:

Ascend GPU CPU

Examples

>>> import mindspore
>>> input = mindspore.tensor([[1, 2], [3, 4]])
>>> mindspore.ops.tile(input, (2, 3))
Tensor(shape=[4, 6], dtype=Int64, value=
[[1, 2, 1, 2, 1, 2],
 [3, 4, 3, 4, 3, 4],
 [1, 2, 1, 2, 1, 2],
 [3, 4, 3, 4, 3, 4]])
>>> mindspore.ops.tile(input, (2, 3, 2))
Tensor(shape=[2, 6, 4], dtype=Int64, value=
[[[1, 2, 1, 2],
  [3, 4, 3, 4],
  [1, 2, 1, 2],
  [3, 4, 3, 4],
  [1, 2, 1, 2],
  [3, 4, 3, 4]],
 [[1, 2, 1, 2],
  [3, 4, 3, 4],
  [1, 2, 1, 2],
  [3, 4, 3, 4],
  [1, 2, 1, 2],
  [3, 4, 3, 4]]])