This is a code for text matching,all the Deep model is run on the platform of pytorch
It is a competition about CIKM spanish text matching
python 3.6 or python 3.5
pytorch 0.4.0 for cpu or gpu (This code main runs on cpu)
tqdm
scikit-learn
data is used to save train,test,word-embeding vector or temporary file
model is used to save all kinds of Deep models
stacking is used to save predicted result on validation set of all kinds of models
data_propresse is used to preprocess data
submit is used to save submit file
run the file runme.py in the dir data_propresse
run the file train.py
run the file ensemble2.py