How can I give a neural network multidimensional input data?

How can I give a neural network multidimensional input data in the following form:

dimension of input 1 = 10x600

dimension of input 2 = 10x700

and so on...

The dimensions of the input matrices cannot be further reduced!

How can you get a 10x600 matrix as input for example to relate to a target which has 1x1 dimensions?

Answer 1

[Srivardhan Gadila]

From the above information, I think the dimension of input 2 is also 10x 600. Then you can create a network with imageInputLayer as input layer with inputSize = [10 600]. Prepare/resize the data such that the dimension of the trainingData matrix would be 10x600x1xN. Add a fullyConnectedLayer with outputSize as 1 followed by output layer.

If your network takes two or more inputs then refer to Multiple-Input and Multiple-Output Networks.