DeepGraphLearning / NBFNet

Official implementation of Neural Bellman-Ford Networks (NeurIPS 2021)
MIT License
197 stars 29 forks source link

Issues of the experiment settings of inductive link prediction #14

Closed smart-lty closed 2 years ago

smart-lty commented 2 years ago

First of all, thanks for the awesome code!

The authors claim that they follow the experiment settings of GraIL, which draws 50 negative triplets for each positive triplet and use the filtered ranking. However, I do not find the corresponding process of drawing 50 negative samples in the code. Can the authors please answer my question?