mindspore.Tensor.var

Tensor.var(axis=None, ddof=0, keepdims=False) → Tensor

Compute the variance along the specified axis.

The variance is the average of the squared deviations from the mean, i.e., \(var = mean(abs(x - x.mean())**2)\).

Return the variance, which is computed for the flattened array by default, otherwise over the specified axis.

Note

Numpy arguments dtype, out and where are not supported.

Parameters

  • axis (Union[None, int, tuple(int)], optional) – Axis or axes along which the variance is computed. The default is to compute the variance of the flattened array. Default: None .

  • ddof (int, optional) – Means Delta Degrees of Freedom. Default: 0 . The divisor used in calculations is \(N - ddof\), where \(N\) represents the number of elements.

  • keepdims (bool, optional) – Whether the output Tensor has dim retained or not. If True , keep these reduced dimensions and the length is 1. If False , don't keep these dimensions. Default: False .

Returns

Variance tensor.

Raises

  • TypeError – If axis is not one of the followings: None, int, tuple.

  • TypeError – If ddof is not an int.

  • TypeError – If keepdims is not a bool.

  • ValueError – If axis is out of range \([-self.ndim, self.ndim)\).

See also

Supported Platforms:

Ascend GPU CPU

Examples

>>> import numpy as np
>>> from mindspore import Tensor
>>> input_x = Tensor(np.array([1., 2., 3., 4.], np.float32))
>>> output = input_x.var()
>>> print(output)
1.25
Tensor.var(dim=None, *, correction=1, keepdim=False) → Tensor

Calculates the variance over the dimensions specified by dim. dim can be a single dimension, list of dimensions, or None to reduce over all dimensions.

The variance (\(\delta ^2\)) is calculated as:

\[\delta ^2 = \frac{1}{\max(0, N - \delta N)}\sum^{N - 1}_{i = 0}(x_i - \bar{x})^2\]

where \(x\) is the sample set of elements, \(\bar{x}\) is the sample mean, \(N\) is the number of samples and \(\delta N\) is the correction.

Parameters

dim (None, int, tuple(int), optional) – The dimension or dimensions to reduce. Defaults to None. If None, all dimensions are reduced.

Keyword Arguments

  • correction (int, optional) – The difference between the sample size and sample degrees of freedom. Defaults to Bessel's correction. Defaults to 1.

  • keepdim (bool, optional) – Whether the output tensor has dim retained or not. If True , keep these reduced dimensions and the length is 1. If False, don't keep these dimensions. Defaults to False.

Returns

Tensor, the variance. Suppose the shape of self is \((x_0, x_1, ..., x_R)\):

  • If dim is () and keepdim is set to False , returns a 0-D Tensor, indicating the variance of all elements in self.

  • If dim is int, e.g. 1 and keepdim is set to False , then the returned Tensor has shape \((x_0, x_2, ..., x_R)\).

  • If dim is tuple(int) or list(int), e.g. (1, 2) and keepdim is set to False , then the returned Tensor has shape \((x_0, x_3, ..., x_R)\).

Raises

  • TypeError – If dim is not one of the followings: None, int, list, tuple.

  • TypeError – If correction is not an int.

  • TypeError – If keepdim is not a bool.

  • ValueError – If dim is out of range \([-self.ndim, self.ndim)\).

Supported Platforms:

Ascend

Examples

>>> import mindspore
>>> from mindspore import Tensor
>>> input_x = Tensor([[8, 2, 1], [5, 9, 3], [4, 6, 7]], mindspore.float32)
>>> output = input_x.var(dim=0, correction=1, keepdim=True)
>>> print(output)
[[ 4.333333, 12.333333, 9.333333]]