mindspore.Tensor.nansum

Tensor.nansum(dim, keepdim=False, *, dtype=None) → Tensor

Computes sum of input Tensor over a given dimension, treating NaNs as zero.

Warning

  • It is only supported on Atlas A2 Training Series Products.

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

Parameters

  • dim (Union[int, tuple(int)], optional) – The dimensions to sum. Dim must be in the range [-rank(self), rank(self)). Default: None, which indicates the sum of all elements in a tensor.

  • keepdim (bool, optional) – Whether the output Tensor keeps dimensions or not. Default: False, indicating the output Tensor doesn't keep dimensions.

Keyword Arguments

dtype (mindspore.dtype, optional) – The dtype of output Tensor. Default: None.

Returns

Tensor, the sum of input Tensor in the given dimension dim, treating NaNs as zero.

  • If dim is None, keepdim is False, the output is a 0-D Tensor representing the sum of all elements in the self Tensor.

  • If dim is int, set as 2, and keepdim is False, the shape of output is \((self_1, self_3, ..., self_R)\).

  • If dim is tuple(int) or list(int), set as (2, 3), and keepdim is False, the shape of output is \((self_1, self_4, ..., self_R)\).

Raises

  • TypeError – If keepdim is not a bool.

  • TypeError – If the dtype of self or dtype is complex type.

  • ValueError – If dim not in [-rank(self), rank(self)).

Supported Platforms:

Ascend

Examples

>>> import mindspore
>>> import numpy as np
>>> from mindspore import Tensor
>>> x = Tensor(np.array([[float("nan"), 2, 3], [1, 2, float("nan")]]), mindspore.float32)
>>> output1 = x.nansum(dim=0, keepdim=False, dtype=mindspore.float32)
>>> output2 = x.nansum(dim=0, keepdim=True, dtype=mindspore.float32)
>>> print(output1)
[1. 4. 3.]
>>> print(output2)
[[1. 4. 3.]]
Tensor.nansum(axis=None, keepdims=False, *, dtype=None) → Tensor

Computes sum of input over a given dimension, treating NaNs as zero.

Parameters

  • axis (Union[int, tuple(int)], optional) – The dimensions to reduce. Supposed the rank of self is r, axis must be in the range [-r,r). Default: None, all dimensions are reduced.

  • keepdims (bool, optional) – Whether the output Tensor keeps dimensions or not. Default: False.

Keyword Arguments

dtype (mindspore.dtype, optional) – The dtype of output Tensor. Default: None.

Returns

Tensor, the sum of input Tensor in the given dimension dim, treating NaNs as zero.

  • If axis is None, keepdims is False, the output is a 0-D Tensor representing the sum of all elements in the input Tensor.

  • If axis is int, set as 2, and keepdims is False, the shape of output is \((self_1, self_3, ..., self_R)\).

  • If axis is tuple(int) or list(int), set as (2, 3), and keepdims is False, the shape of output is \((self_1, self_4, ..., self_R)\).

Raises

  • TypeError – If keepdims is not a bool.

  • TypeError – If the dtype of self or dtype is complex type.

  • ValueError – If axis not in [-rank(self), rank(self)).

Supported Platforms:

Ascend GPU CPU

Examples

>>> import mindspore
>>> import numpy as np
>>> from mindspore import Tensor, ops
>>> x = Tensor(np.array([[float("nan"), 2, 3], [1, 2, float("nan")]]), mindspore.float32)
>>> output1 = x.nansum(axis=0, keepdims=False, dtype=mindspore.float32)
>>> output2 = x.nansum(axis=0, keepdims=True, dtype=mindspore.float32)
>>> print(output1)
[1. 4. 3.]
>>> print(output2)
[[1. 4. 3.]]