Sonata165 / GraphTransformerDemo

NUS CS6208 assignment code repository

Code for CS6208 assignment

Reproduce Graph Transformer

Paper: https://arxiv.org/abs/2012.09699

Ou Longshen's assignment submission.

Environment

# On linux (CUDA 11.7)
pip install torch==1.13.1+cu117 torchvision==0.14.1+cu117 torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu117
pip install  dgl -f https://data.dgl.ai/wheels/cu117/repo.html
pip install  dglgo -f https://data.dgl.ai/wheels-test/repo.html

Files

data            Dataset used in the experiment.
models_gt.py    Code to define the Graph Transformer model.
train.py        Code for model training and evaluation.
train.ipynb     Training log of one experiment.

Run the code

Check the train.ipynb for previous training logs.

Or use below command:

python train.py

to run the training script.