You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Congratulations on your newly published KDD 2022 paper: GraphMAE: Self-Supervised Masked Graph Autoencoders. We found its idea interesting and inspiring, and would like to do further research over it.
Recently we run this public source code on NCI1 dataset for graph classification and the reproduced result is merely 76.9%, which exists a large gap compared to the reported result (80.4%) in the main paper. Would you please share the complete parameter setting of NCI1 with us for academic research?
Best regards
The text was updated successfully, but these errors were encountered:
@XinPeng97 Thanks for your attention to our work. Important hyperparameters are listed in this file. And all hyperparameters will be printed when running the experiment. The reported results can be reproduced by running scripts/run_graph.sh NCI1 <gpu_id>. And the experiments are conducted on NVIDIA 2080Ti. Could you provide more details about your experiment?
Dear Authors,
Congratulations on your newly published KDD 2022 paper: GraphMAE: Self-Supervised Masked Graph Autoencoders. We found its idea interesting and inspiring, and would like to do further research over it.
Recently we run this public source code on NCI1 dataset for graph classification and the reproduced result is merely 76.9%, which exists a large gap compared to the reported result (80.4%) in the main paper. Would you please share the complete parameter setting of NCI1 with us for academic research?
Best regards
The text was updated successfully, but these errors were encountered: