mindspore.ops.BesselK1

class mindspore.ops.BesselK1[source]

Computes modified Bessel function of the second kind, order 1 element-wise.

The formula is defined as:

\[\begin{split}\begin{array}{ll} \\ K_{1}(x)=\lim_{\nu \to 1} \left(\frac{\pi}{2}\right) \frac{I_{-\nu}(x)- I_{\nu}(x)}{\sin (\nu \pi)} = \int_{0}^{\infty} e^{-x \cosh t} \cosh (t) d t \end{array}\end{split}\]

where \(I_{1}\) is modified Bessel function of the first kind, order 1.

Warning

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

Inputs:

x (Tensor) - The input tensor. Data type must be float16, float32, float64.

Outputs:

Tensor, has the same shape as x.

Raises: TypeError – If x is not a Tensor of float16, float32, float64.

Supported Platforms:: GPU CPU

Examples

>>> import mindspore
>>> import numpy as np
>>> from mindspore import Tensor, ops
>>> bessel_k1 = ops.BesselK1()
>>> x = Tensor(np.array([0.24, 0.83, 0.31, 0.09]), mindspore.float32)
>>> output = bessel_k1(x)
>>> print(output)
[3.9190812  0.8143549  2.9440577 10.974864]