HRGAT

The code for hyper-node relational graph attention network

requirement

torch == 1.7.0, dgl == 0.4.2, sentence-bert

Train Model

To start a simple training process:

python run.py --data FB15k-237 --text --img --attr

--model denotes the link prediction score score function
--gpu for specifying the GPU to use
--epoch for number of epochs
--batch for batch size
--text for text information
--img for batch information
--attr for batch information
Rest of the arguments can be listed using python run.py -h