Skip to content

Latest commit

 

History

History
47 lines (23 loc) · 973 Bytes

README.md

File metadata and controls

47 lines (23 loc) · 973 Bytes

ROD: Reception-aware Online Distillation for Sparse Graphs .

This repository is the official implementation of ROD.

Requirements

Environment: 16GB(RAM), Ubuntu 16.04 (OS), TITAN RTX (GPU), Xeon Gold 5120 (CPU)

To install requirements:

pip install -r requirements.txt

Training

To train the model(s) in the paper:

cd the “examples” data

run the ipynb notebook file "test_classification.ipynb", "test_link_prediction.ipynb", "test_clustering.ipynb"

Results

  1. Node Classification:

    ​ (1) transductive setting:

​ (2) inductive setting:

  1. Link Prediction:

  1. Node Clustering: