This repo contains code accompaning the paper, GPPT: Graph Pre-training and Prompt Tuning to Generalize Graph Neural Networks (Mingchen Sun et al., KDD 2022).
This code requires the following:
- python 2.* or python 3.*
- PyTorch v1.8.+
- DGL v0.7+
We evaluate our model on eight benchmark datasets, see the usage instructions in load_graph.py and utils.py respectively.
The hyperparameters settings see get_args.py.
To run the code, see the usage instructions at the top of GPPT.py or run_all.py.