Functional API

This is just another layout for coding the model graph. You can choose the following layout if you are more comfortable with Python style code writing:

from keras.models import Model
from keras.layers import Dense, Input

#defining input placeholder with input shape
inp = Input(shape = 100)

# layers
x = Dense(units = 128, activation = 'relu')
x = Dense(units = 64, activation = 'relu')

# taking output
predict = Dense(units = 4, activation = 'softmax')(x)

# defining model
model = Model(inputs = inp, outputs = predict)