Is this the correct way to prepare the input and target vectors to feed them to neural network?

I am having a dataset of 50 images, i extracted the features of the 50 images and stored them as vectors in a matrix using the CSV so it was stored with .dat format, then in excel i created the target vectors where every column represents an image and every row represents a class. the target excel sheet is saved in .dat format

my questions are:

• Is it okey to use .dat format with neural network?
• how to make a dataset of the inputs and target?ex: i want to name the dataset : "pupiles"so [inputvectors,targets]=pupilesand so that i can use:

load pupiles

thank you in advance