Skip to content

Generative deep learning model for inorganic materials

Notifications You must be signed in to change notification settings

sgbaird/MatGAN

 
 

Repository files navigation

MatGAN

Video demo

Installation of running environment

$pip install -r requirement.txt

How to run:

  • $python generate_data.py OQMD gan-materials-oqmd.csv
  • $python generate_data.py ICSD gan-materials-icsd.csv
  • $python generate_data.py MP gan-materials-mp.csv

This will generate 1024 hypothetical materials to corresponding csv files such as gan-materials-oqmd.csv

To generate more materials, just continue to re-run this same command, the results will be appended to the assigned output file

Generated hypothetical materials by GAN-ICSD

Citing

Dan, Y.; Zhao, Y.; Li, X.; Li, S.; Hu, M.; Hu, J. Generative Adversarial Networks (GAN) Based Efficient Sampling of Chemical Composition Space for Inverse Design of Inorganic Materials. npj Computational Materials 2020, 6 (1), 1–7. https://doi.org/10.1038/s41524-020-00352-0.

About

Generative deep learning model for inorganic materials

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 95.4%
  • Makefile 2.3%
  • Batchfile 2.3%