dlhdl.layer.mishLayer

Mish layer

Since R2024a

Description

Use mishLayer objects to apply the mish function to the layer inputs.

This equation describes the mish operation:

$$f(x)=x\tanh (\ln (1+e^x))$$

Creation

Syntax

layer = dlhdl.layer.mishLayer(Name)

Description

layer = dlhdl.layer.mishLayer(Name) creates a mish layer with the name specified by Name. For example, dlhdl.layer.mishLayer("mish1") creates a mish layer with the name "mish1".

Properties

Name — Layer name

"" (default) | character vector | string scalar

Layer name, specified as a character vector or string scalar. For Layer array input, the trainnet and dlnetwork functions automatically assign names to layers with the name "".

The mishLayer object stores this property as a character vector.

Data Types: char | string

NumInputs — Number of inputs

1 (default)

This property is read-only.

Number of inputs to the layer, returned as 1. This layer accepts a single input only.

Data Types: double

InputNames — Input names

{'in'} (default)

This property is read-only.

Input names, returned as {'in'}. This layer accepts a single input only.

Data Types: cell

NumOutputs — Number of outputs

1 (default)

This property is read-only.

Number of outputs from the layer, returned as 1. This layer has a single output only.

Data Types: double

OutputNames — Output names

{'out'} (default)

This property is read-only.

Output names, returned as {'out'}. This layer has a single output only.

Data Types: cell

Examples

Create Mish Layer

Create a mish layer with the name "mish1".

layer = dlhdl.layer.mishLayer("mish1")

layer = 
  mishLayer with properties:

    Name: 'mish1'
   Learnable Parameters
    No properties.
   State Parameters
    No properties.
  Show all properties

Include the mish layer in a Layer array.

 layers = [imageInputLayer([20,20,3],'Normalization',"none",'Name','input')
                convolution2dLayer([5 5],3,'Padding',[1 2 1 2],'Stride',[1 1],'Name', 'conv')
                batchNormalizationLayer('Name','batchnorm')
                dlhdl.layer.mishLayer("mish1")
                convolution2dLayer([5 5],3,'Padding',[1 2 1 2],'Stride',[2 2],'Name', 'conv')
                batchNormalizationLayer('Name','batchnorm')
                swishLayer('Name','swish')]

layers = 
  7×1 Layer array with layers:

     1   'input'       Image Input             20×20×3 images
     2   'conv'        2-D Convolution         3 5×5 convolutions with stride [1  1] and padding [1  2  1  2]
     3   'batchnorm'   Batch Normalization     Batch normalization
     4   'mish1'       dlhdl.layer.mishLayer   Custom mish Layer
     5   'conv'        2-D Convolution         3 5×5 convolutions with stride [2  2] and padding [1  2  1  2]
     6   'batchnorm'   Batch Normalization     Batch normalization
     7   'swish'       Swish                   Swish

Algorithms

Mish Layer

A mish activation layer applies the mish function to the layer inputs. The mish operation uses this equation $$f(x)=x\tanh (\ln (1+e^x))$$

The mish layer does not change the size of the input. Activation layers such as mish layers improve the training accuracy for some applications and usually follow convolution and normalization layers. Other nonlinear activation layers perform different operations. For a list of activation layers, see Activation Layers.

Version History

Introduced in R2024a