Is it possible to train feed forward network using keras with different shapes of x_train and y_train batches? If so, how?
my goal is x_train to be
