mindspore.ops.aminmax

mindspore.ops.aminmax(input, *, axis=0, keepdims=False)[源代码]

返回tensor在指定轴上的最小值和最大值。

警告

从2.9.0(不含)之后版本开始,该接口将发生非兼容性变更:参数 axis 将重命名为 dimaxis 的默认值将从 0 调整为 None

参数:
  • input (Tensor) - 输入tensor。

关键字参数:
  • axis (int,可选) - 指定计算的轴。如果为 None ,计算 input 中的所有元素。默认 0

  • keepdims (bool,可选) - 输出tensor是否保留维度。默认 False

返回:

由两个tensor组成的tuple(min, max)。

支持平台:

Ascend GPU CPU

样例:

>>> import mindspore
>>> input = mindspore.tensor([[9, 3, 4, 5],
...                           [5, 2, 7, 4],
...                           [8, 1, 3, 6]])
>>>
>>> # case 1: By default, compute along axis 0.
>>> mindspore.ops.aminmax(input)
(Tensor(shape=[4], dtype=Int64, value= [5, 1, 3, 4]),
 Tensor(shape=[4], dtype=Int64, value= [9, 3, 7, 6]))
>>>
>>> # case 2: Disregard NaN (Not a Number) values present in the input during computation.
>>> input = mindspore.tensor([[9, 3, 4, 5],
...                           [5, 2, 7, 4],
...                           [8, 1, 3, float('nan')]])
>>> mindspore.ops.aminmax(input, axis=None)
(Tensor(shape=[], dtype=Float32, value= 1),
 Tensor(shape=[], dtype=Float32, value= 9))
>>>
>>> # case 3: If keepdims=True, the output shape will be the same as that of the input.
>>> mindspore.ops.aminmax(input, axis=None, keepdims=True)
(Tensor(shape=[1, 1], dtype=Float32, value=
 [[ 1.00000000e+00]]),
 Tensor(shape=[1, 1], dtype=Float32, value=
 [[ 9.00000000e+00]]))