Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Latest commit

 

History

History
 
 

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 

README.md

CogDL examples for ogbn-arxiv

CogDL implementation of VRGCN for ogbn-arxiv:

Jianfei Chen, Jun Zhu, Le Song. Stochastic Training of Graph Convolutional Networks with Variance Reduction. Paper in arXiv. In ICML'2018.

Requires CogDL 0.5-alpha0 or later versions.

Training & Evaluation

# Run with model with default config
python main.py

For more hyper-parameters, please find them in the main.py.

Results

Here are the results over 10 runs which are comparable with OGB official results reported in the leaderboard.

Method Test Accuracy Validation Accuracy #Parameters
VRGCN 0.7224 ± 0.0042 0.7260 ± 0.0030 44,328