train_CAML.py is used for train a CAML model.

test_CAML.py is used for predicting and generating explanations for a new user-item pair.

run_train.cmd and run_test.cmd are corresponding example instructions.

The descriptions and default values of parameters for training and testing are in parser_CAML.py.

Codes in folder tf_models are corresponding model files of CAML. Other necessary operations are in the tylib folder.

Evaluate folder contains the ROUGE-1.5.5 tool and a example script of ROUGE evaluation.

data folder shows the names and formats of each input file in the dataset.