Download data from https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/data and put 'train' folder in 'Data'
dims_out=(13,32) shape = (13,32,1)
#train model: from generator_mfcc import DataGenerator
#test model on our own made test set: from generator_mfcc_test import DataGenerator
#test model for final submission: from generator_mfcc_finaltest import DataGenerator
dims_out=(99,161) shape = (99,161,1)
#train model: from generator_spectrum import DataGenerator
#test model on our own made test set: from generator_spectrum_test import DataGenerator
#test model for final submission: from generator_spectrum_finaltest import DataGenerator