Input

Presentation

Input Layer must be the first layer of any neural network, it is not really a layer (has no neurons) but it provides to define the shape of input. For example for 32 x 32 color images you can declare input like Input(32, 32, 3) to have 2D input but you can also use Input(1024, 3) or Input(3072) to have 1D input. The shape of input (1D or 2D) determines the shape of the next layer. The shape does not change anything for simple Layers like FullyConnected if the number of input stay the same but for filter layers like Convolution or LoccalyConnected the input layer will have the same number of dimensions as the input.

Declaration

This is the function used to declare the size input of neural network.

template <typename ... TInt>
LayerModel Input(TInt... sizeOfInput)

Here is an example of neural networks with 2D input.

StraightforwardNeuralNetwork neuralNetwork({
        Input(28, 28, 1),
        Convolution(1, 5),
        FullyConnected(70),
        FullyConnected(10)
    });

See an example of GRU layer on dataset
3D inputs or more are not supported.