The dataset in this example is delicious, which has been processed and split into training (70%), validation(10%) and testing(20%). The processed data is stored in data.pkl.
You need to place other supporting data files in the /Data/ directory to the same directory with model.py to train and test the model.
The pre-trained model is located in /checkpoint/ directory, which can be used to initialized the model.
run python3 model.py train to train the model and python3 model.py test to perform testing
The negative sampling rate is different for the training, validating and testing phase, and it can be modified in the source file
The recommendation performance is evaluating using HR@K AND NDCG@K, and they reside in the file res.dat in the directory /model/ after testing