inputLayer

Input layer

Since R2023b

Description

An input layer inputs data into a neural network with a custom format.

• For 2-D image input, use imageInputLayer.

• For 3-D image input, use image3dInputLayer.

• For sequence and time series input, use sequenceInputLayer.

• For tabular and feature data input, use featureInputLayer.

Creation

Syntax

layer = inputLayer(inputSize,inputFormat)

layer = inputLayer(inputSize,format,Name=name)

Description

layer = inputLayer(inputSize,inputFormat) creates an input layer and sets the InputSize and InputFormat properties.

layer = inputLayer(inputSize,format,Name=name) also sets the Name property.

Properties

Input

InputSize — Size of input

row vector of positive integers or NaN

This property is read-only.

Size of the input, specified as a row vector of positive integers or NaN.

For networks that support variable sizes for the batch or time dimensions, specify the size of the corresponding dimension as NaN.

Example: [224 224 3]

InputFormat — Description of data dimensions

character vector | string scalar

Description of the data dimensions, specified as a character vector or string scalar.

A data format is a string of characters, where each character describes the type of the corresponding data dimension.

The characters are:

• "S" — Spatial

• "C" — Channel

• "B" — Batch

• "T" — Time

• "U" — Unspecified

For example, consider an array containing a batch of sequences where the first, second, and third dimensions correspond to channels, observations, and time steps, respectively. You can specify that this array has the format "CBT" (channel, batch, time).

You can specify multiple dimensions labeled "S" or "U". You can use the labels "C", "B", and "T" at most once. When the number of dimensions is greater than two, the input size corresponding to the rightmost "U" dimension in the input format must be greater than 1.

For more information, see Deep Learning Data Formats.

Example: "SSCB"

Example: "SCBT"

Example: "TCB"

Data Types: char | string

Layer

Name — Layer name

"" (default) | character vector | string scalar

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

The InputLayer object stores this property as a character vector.

Data Types: char | string

NumInputs — Number of inputs

0 (default)

This property is read-only.

Number of inputs of the layer. The layer has no inputs.

Data Types: double

InputNames — Input names

{} (default)

This property is read-only.

Input names of the layer. The layer has no inputs.

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 1-D Image Input Layer

Create an input layer that inputs 1-D image data (3-D data, with dimensions corresponding to space, channels, and observations). Specify that the data has three channels and a image height size of 64. Specify that the batch dimension can vary.

inputSize = [64 3 NaN];
inputFormat = "SCB";
layer = inputLayer(inputSize,inputFormat)

layer = 
  InputLayer with properties:

           Name: ''
      InputSize: [64 3 NaN]
    InputFormat: 'SCB'

Include the input layer in a network.

layers = [
    inputLayer([64 3 NaN],"SCB")
    convolution1dLayer(5,32)
    batchNormalizationLayer
    reluLayer
    fullyConnectedLayer(10)
    softmaxLayer];

Create Spatiotemporal Input Layer

Create an input layer that inputs spatiotemporal data (4-D data, with dimensions corresponding to space, channels, time, and observations). Specify that the data has three channels and a spatial size of 64. Specify that the batch and time dimensions can vary.

inputSize = [64 3 NaN NaN];
inputFormat = "SCBT";
layer = inputLayer(inputSize,inputFormat)

layer = 
  InputLayer with properties:

           Name: ''
      InputSize: [64 3 NaN NaN]
    InputFormat: 'SCBT'

Include the input layer in a network.

layers = [
    inputLayer([64 3 NaN NaN],"SCBT")
    convolution1dLayer(5,32)
    batchNormalizationLayer
    reluLayer
    globalAveragePooling1dLayer
    flattenLayer
    lstmLayer(100,OutputMode="last")
    fullyConnectedLayer(10)
    softmaxLayer];

Algorithms

Complex Numbers

InputLayer objects support passing complex-valued data to subsequent layers. (since R2024a)

For an example showing how to train a network with complex-valued data, see Train Network with Complex-Valued Data.

Extended Capabilities

C/C++ Code Generation

Generate C and C++ code using MATLAB® Coder™.

Usage notes and limitations:

You can generate C or C++ code that does not depend on any deep learning third-party libraries for formats with any number of spatial dimensions. For example, code generation supports the input format "SSSCBT" (spatial, spatial, spatial, channel, batch, time) for inputLayer.

Code generation for ARM^® Compute and Intel^® MKL-DNN only supports permutations of these input formats:

• "CB" (channel, batch)

• "SSCB" (spatial, spatial, channel, batch)

• "CBT" (channel, batch, time)

• "SSCBT" (spatial, spatial, channel, batch, time)

GPU Code Generation

Generate CUDA® code for NVIDIA® GPUs using GPU Coder™.

Usage notes and limitations:

• You can generate plain CUDA code that is independent of deep learning libraries for formats with any number of spatial dimensions. For example, code generation supports the input format "SSSCBT" (spatial, spatial, spatial, channel, batch, time) for inputLayer.

• You can generate code that takes advantage of the NVIDIA^® CUDA^® deep neural network library (cuDNN), or the NVIDIA TensorRT™ high performance inference library.

  The cuDNN library supports permutations of these input formats:

  • "CB" (channel, batch)

  • "SSCB" (spatial, spatial, channel, batch)

  • "CBT" (channel, batch, time)

  • "SSCBT" (spatial, spatial, channel, batch, time)

• The TensorRT library supports permutations of these input formats:

  • "CB" (channel, batch)

  • "SSCB" (spatial, spatial, channel, batch)

  • "CBT" (channel, batch, time)

Version History

Introduced in R2023b

R2024a: Complex-valued outputs

For complex-valued input to the neural network, the layer passes complex-valued data to subsequent layers.