Could you possibly share the inductive split dataset and also the negative samples used for evaluation? I believe it's quite important for correctly benchmarking the model.
I also noticed the fixed transductive split provided in SSI-DDI paper -- I would greatly appreciate it if the inductive split can be provided in that way too!
Hi! Thanks for sharing your great work!
Could you possibly share the inductive split dataset and also the negative samples used for evaluation? I believe it's quite important for correctly benchmarking the model.
I also noticed the fixed transductive split provided in SSI-DDI paper -- I would greatly appreciate it if the inductive split can be provided in that way too!