1, data cleansing (pre-process)
The data pre-processing code used in Section 3.1.
(1)Building conversations from original data and cleansing the typos and slangs. 
transformation.py, inlinepatterns.py

(2) Vocabulary truncation
    proc2.py, filter.py 

(3) Dialog truncation 
    filter.py 

2, train and test Neural Response Model (model)

train*model.sh
test*model.sh