xuhuihuang/graphvampnets

GraphVAMPnets combines graph neural networks with variational approach for Markovian process (VAMP) theory to identify the slow collective variables of the self-assembly processes.

GraphVAMPnets for Uncovering Slow Collective Variables of Self-Assembly Dynamics

Installation

The source code can be installed with a local clone:

git clone https://github.com/xuhuihuang/graphvampnets.git

(note that the dependencies should be installed first by two steps)

pip install -r ./requirements1.txt
pip install -r ./requirements2.txt

python ./setup.py install

Tutorial

A tutorial can be found here. Note that the data may be too large for colab to process, and you can load processed data for training or load pretrained models to play.

Reference

Our codebase builds heavily on

• https://github.com/msmbuilder/msmbuilder
• https://github.com/deeptime-ml/deeptime
• https://github.com/ghorbanimahdi73/GraphVampNet

Thanks for open-sourcing!