infolabai / custom-dida Goto Github PK

pip install <scatter NAME>.whl # In this case, from https://data.pyg.org/whl/torch-2.0.1%2Bcu117.html

pip install <geometric NAME>.whl ...

https://drive.google.com/file/d/19SOqzYEKvkna6DKd74gcJ50Wd4phOHr3/view?usp=share_link

python ./dataset_loader/preprocess_dgl_brw.py # Download dataset, produce DGL graphs and save DGL graphs as pt files to fix node features
python ./dataset_loader/preprocess_dict_from_dgl.py # Produce dictionaries from DGL graphs, and save dictionaries as pt files becuase we need to fix negative edges for test

# Our method with edge propagation with alpha_std
python main.py --model ours --seed 123 --device_id 0 --propagate dyaug --alpha_std 1 --dataset collab --ex_name "Dynamic aug"

# DIDA
python main.py --model dida --seed 123 --device_id 0 --dataset collab --ex_name "Dynamic aug"

bash script.sh
# Then, check the log folder <args.log_dir>/<args.ex_name>/* for this run

# To check logs, move the log folder to <TENSORBOARD FOLDER> and run tensorboard
tensorboard --logdir <TENSORBOARD FOLDER> --port <PORT>

# Check the logs in tensorboard