We will not perform KGC using different link prediction models.
cd pykeen
We first generate data that will be used for comparing the KGC results.
./generate_data.sh
We need to optimize the hyperparameters for multiple models and multiple datasets. The following script runs the HPO in a sequential manner, so you may wish to change the script based on your own needs.
./run_hpo.sh
Using the best hyperparameter found above for each dataset and model pair, we get the statistics on the test set.
./run_test_using_best_models.sh