CKKGEm: Context Knowledge for Knowledge Graph Embedding models

Usage

Step 1: Prepare datasets

• Command:

bash codes/preprocess.sh [task] [No.workers]

• Options:
  • No.workers: number of workers for parallel processing
  • task: on which dataset the script works on (possible values: wn18rr, fb15k237)

Step 2: Training and testing

• Command:

bash run.sh [Options]

• Options: please check run.py for more information

Step 3: Results

• Command:

python codes/results.py

Notes

According to the original project, best_config.sh contains all the commands with the best arguments for each models. Therefore, we use those commands to obtain the results for our models which are based on the original models.