Standard deviation of a tensor, alongside the specified axis.
tf.keras.backend.std(
x, axis=None, keepdims=False
)
It is an alias to tf.math.reduce_std
.
Arguments |
x
|
A tensor or variable. It should have numerical dtypes. Boolean type
inputs will be converted to float.
|
axis
|
An integer, the axis to compute the standard deviation. If None
(the default), reduces all dimensions. Must be in the range
[-rank(x), rank(x)) .
|
keepdims
|
A boolean, whether to keep the dimensions or not.
If keepdims is False , the rank of the tensor is reduced
by 1. If keepdims is True , the reduced dimension is retained with
length 1.
|
Returns |
A tensor with the standard deviation of elements of x with same dtype.
Boolean type input will be converted to float.
|