This repository is the official PyTorch implementation of "Monotone Operator Theory-Inspired Message Passing for Learning Long-Range Interaction on Graphs."
Dockerfile: Recommended install using Dockerfile.drgnn.
Manual: Alternatively use pip3 install -r requirements.txt to manually install.
agg: The ./agg/ directory contains all the relevant code for DRGNN.
tasks: Individual tasks are located in the ./tasks/ directory with .py driver files and .yaml settings files.
python3 ./tasks/arxiv.py
sweeps: Sweep results are run using wandb and can be easily instantiated and run by performing the following:
- Move desired runs to
./sweeps/todo/.
mv ./sweeps/tables/* ./sweeps/todo/
- Generate
sweep_ids.log.
bash ./sweeps/generate.sh todo
- Run the sweep.
bash ./sweeps/run.sh <device>
analysis: All other results are found in the ./analysis/ directory.
baselines: All baseline codes can be found in the ./baselines/ directory.
Awaiting publication.