mindspore.Tensor.hardshrink

Tensor.hardshrink(lambd=0.5) → Tensor

Hard Shrink activation function. Calculates the output according to the input elements.

The formula is defined as follows:

$$\begin{array}{r}\text{HardShrink}(x)=\{\begin{array}{ll}x,& \text{ if }x>\lambda \\ x,& \text{ if }x<-\lambda \\ 0,& \text{ otherwise }\end{array}\end{array}$$

HardShrink Activation Function Graph:

Note

The input Tensor. $x$ in the above formula. Supported dtypes:

• Ascend: float16, float32, bfloat16.

• CPU/GPU: float16, float32.

Parameters

lambd (number, optional) – The threshold $\lambda$ defined by the Hard Shrink formula. Default: 0.5 .

Returns

Tensor, has the same data type and shape as the input self.

Raises

• TypeError – If lambd is not a float, int or bool.

• TypeError – If self is not a tensor.

• TypeError – If dtype of self is not float16, float32 or bfloat16.

Supported Platforms:

Ascend GPU CPU

Examples

>>> import mindspore
>>> import numpy as np
>>> from mindspore import Tensor
>>> input = Tensor(np.array([[0.5, 1, 2.0], [0.0533, 0.0776, -2.1233]]), mindspore.float32)
>>> output = input.hardshrink()
>>> print(output)
[[ 0.      1.      2.    ]
 [ 0.      0.     -2.1233]]