seongjunyun / Graph_Transformer_Networks

Graph Transformer Networks (Authors' PyTorch implementation for the NeurIPS 19 paper)
960 stars 179 forks source link

Applicable network type #7

Closed Fujiaoji closed 4 years ago

Fujiaoji commented 4 years ago

Hi, if this model can be used in this type of heterogeneous network? type of edge > 1 and type of nodes > 1

Fujiaoji commented 4 years ago

Hi, adjacency matrix has different dimensions, like, APC, APA, if AP belongs to Rn, PC belong to Rm, APA belongs to Rt, how do u deal with them?

seongjunyun commented 4 years ago

Hi, The dimension of adjacecny matrices is NxN and N is total number of nodes (Author, Paper and Conference). So they have same dimensions.

Thank you.