Skip to content

An application of graph convolution neural network for Chemistry problem.

Notifications You must be signed in to change notification settings

thhung/Champs_GraphConV

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Champs_GraphConV

This project is my implementation for using the graph convolution neural network for the Champs competition on Kaggle. This is showcase from data preprocessing to the final result. It is developed base on the torch-geometric library.

To preprocessing data, you would need to install two library such as dscribe and rdkit to be able to extract the feature about atom and molecules.

Installation for dependencies:

pip install networkx dscribe
conda install -y -c rdkit rdkit

In this experiment, serveral configurations and different neural networks are experimented. It seems that adding the attention mechanism into network has a significant impact on the performance rather than pure simple graph conv.

About

An application of graph convolution neural network for Chemistry problem.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages