Paper:InGram: Inductive Knowledge Graph Embedding via Relation Graphs
Clone the Openhgnn-DGL
python main.py -m Ingram -d NL-100 -t Ingram -g 0
If you do not have gpu, set -gpu -1.
Candidate dataset: NL-100
InGram[OpenHGNN] | MR | MRR | H@10 | H@1 |
---|---|---|---|---|
NL-100 | Paper:92.6 openhgnn:97.0 | Paper:0.309 openhgnn:0.295 | Paper:0.506 openhgnn:0.494 | Paper: 0.212 openhgnn: 0.193 |
Ingram
NL-100
Zikai Zhou[GAMMA LAB]
Submit an issue or email to [email protected].