We provide the code (in pytorch) and datasets for our paper "HGPROMPT: Bridging Homogeneous and Heterogeneous Graphs for Few-shot Prompt Learning", which is accepted by AAAI2024.
- cuda 11.3
- dgl0.9.0-cu113
- dgllife
The default dataset is ACM. You need to change the corresponding parameters in pre_train.py and run.py to train and evaluate on other datasets.
Pretrain:
- python pre_train.py
Prompting and test:
- python run.py