$pip install -r requirement.txt
- $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
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.