Skip to content

Frank-LIU-520/ape-MPNN

Repository files navigation

ape-MPNN(DeepMoleNet)

Transferable Multilevel Attention Neural Network for Accurate Prediction of Quantum Chemistry Properties via Multitask Learning https://pubs.acs.org/doi/10.1021/acs.jcim.0c01224#

Champion solution of Codalab Alchemy Contest
https://alchemy.tencent.com/

DeepMoleNet is a deep learning package for molecular properties prediction. This code was developed by Ziteng Liu@Nanjing Univerisity and Liqiang Lin@Nanjing University during Codalab Alchemy Contest https://alchemy.tencent.com/. We formed one team called "NJU_Chem" and won the champion in the competition. We also released document to disclose our thoughts during different stages in the competition. Former DeepMoleNet was called ape-MPNN in the contest. Please see https://alchemy.tencent.com/data/2019/1st_solution-ape-MPNN_NJU_Chem.pdf for detail.

image

To cite this algorithm, please reference: Liu, Ziteng, et al. "Transferable multi-level attention neural network for accurate prediction of quantum chemistry properties via multi-task learning." ChemRxiv 12588170 (2020): v1.

Please go to the DeepMoleNet homepage, register, and download the code before usage in the following steps.

1 unzip the zip file and make sure all packagement in the requirement.txt file are installed;

2 put molecule sdf files in the .\data-bin\raw\dev, with its sdf file name and all 12 properties saved in .\data-bin\raw\dev_target.csv in templet.

3 run the code, python DeepMoleNet.py

Contact [email protected] if you have any questions.

About

Champion solution of Coalab Alchemy Contest

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages