HSigmoid

Versioned name: HSigmoid-5

Category: Activation function

Short description: HSigmoid takes one input tensor and produces output tensor where the hard version of sigmoid function is applied to the tensor elementwise.

Detailed description: For each element from the input tensor calculates corresponding element in the output tensor with the following formula:

\[HSigmoid(x) = \frac{min(max(x + 3,\ 0),\ 6)}{6}\]

The HSigmoid operation is introduced in the following article.

Attributes: operations has no attributes.

Inputs:

  • 1: A tensor of type T. Required.

Outputs:

  • 1: The resulting tensor of the same shape and type as input tensor.

Types

  • T: any floating-point type.

Examples

<layer ... type="HSigmoid">
    <input>
        <port id="0">
            <dim>256</dim>
        </port>
    </input>
    <output>
        <port id="1">
            <dim>256</dim>
        </port>
    </output>
</layer>