byun-jinyoung / graphvampnet

GraphVampNet

This repo contains the code for GraphVAMPNet

GraphVAMPNet code

Usage

training

python train.py --epochs 100 --batch-size 1000 --lr 0.0005 --hidden 16  
    --num-atoms 20 --num-classes 5 --num_neighbors 7 --conv_type SchNet --dmin 0 
    --dmax 8. --step 0.5 --dist-data dists.dat --nbr-data nbrs.dat --residual  --train

testing

python train.py --epochs 100 --batch-size 1000 --lr 0.0005 --hidden 16 
    --num-atoms 20 --num-classes 5 --num-neighbors 7 --conv_type SchNet --dmin 0
    --dmax 8. --step 0.5 --dist-data dists.dat --nbr-data nbrs.dat --residual --trained-model
    logs/logs_99.pt

Requirements

• pytorch
• deeptime
• torch_scatter

Sources:

• VAMPNet code is based on deeptime package deeptime
• SchNet code is based on the cgnet

Cite

If you use this code please cite the following paper: