Methodology:
adjusting hyperparameter such as
1、 the num of dense-layers, conv2D-layers,
2、the num of filters of each conv2D-layers.
learned about:
1、use dense_layers=[0,1,2]+ for i in dense_layers automatically create model names and avoid complex computing on how much models should I train
2、further, + for num in range(i), to determine how much layers in the model which is under training.
