best way to handle large input for neural network

The training data is a collection of 84x300 arrays for inputs paired with integers as outputs. Do I need to flatten the input to a 25200-wide vector? I also noticed that the output has to be the same data type as the input, so do I need one input for each element of the vector? Is there a more efficient way to handle this?

Answer 1

[Greg Heath]

When each of N I-dimensional "I"nput vectors is paired with one of corresponding N O-dimensional "O"utput target vectors,

 [ I N ] = size(input)

 [ O N ] = size(target)

Hope this helps.

Thank you for formally accepting my answer

Greg