mindspore.ops.Select

class mindspore.ops.Select(*args, **kwargs)[source]

Returns the selected elements, either from input \(x\) or input \(y\), depending on the condition.

Given a tensor as input, this operation inserts a dimension of 1 at the dimension, it was invalid when both math: ‘x’ and math: ‘y’ are none. Keep in mind that the shape of the output tensor can vary depending on how many true values are in the input. Indexes are output in row-first order.

The conditional tensor acts as an optional compensation (mask), which determines whether the corresponding element / row in the output must be selected from \(x\) (if true) or \(y\) (if false) based on the value of each element.

It can be defined as:

\[\begin{split}out_i = \begin{cases} x_i, & \text{if } condition_i \\ y_i, & \text{otherwise} \end{cases}\end{split}\]

If condition is a vector, then \(x\) and \(y\) are higher-dimensional matrices, then it chooses to copy that row (external dimensions) from \(x\) and \(y\). If condition has the same shape as \(x\) and \(y\), you can choose to copy these elements from \(x\) and \(y\).

Inputs:

  • input_cond (Tensor[bool]) - The shape is \((x_1, x_2, ..., x_N, ..., x_R)\). The condition tensor, decides which element is chosen.

  • input_x (Tensor) - The shape is \((x_1, x_2, ..., x_N, ..., x_R)\). The first input tensor.

  • input_y (Tensor) - The shape is \((x_1, x_2, ..., x_N, ..., x_R)\). The second input tensor.

Outputs:

Tensor, has the same shape as input_x. The shape is \((x_1, x_2, ..., x_N, ..., x_R)\).

Raises

  • TypeError – If input_x or input_y is not a Tensor.

  • ValueError – If shape of input_x is not equal to shape of input_y or shape of input_cond.

Supported Platforms:

Ascend GPU CPU

Examples

>>> select = ops.Select()
>>> input_cond = Tensor([True, False])
>>> input_x = Tensor([2,3], mindspore.float32)
>>> input_y = Tensor([1,2], mindspore.float32)
>>> output = select(input_cond, input_x, input_y)
>>> print(output)
[2. 2.]