divelab / DIG

A library for graph deep learning research
https://diveintographs.readthedocs.io/
GNU General Public License v3.0
1.84k stars 281 forks source link

Global features in graph generation #128

Closed gooaah closed 2 years ago

gooaah commented 2 years ago

Dear developers,

I am trying to produce crystal structures by graph generative models (like GraphDF) provided in DIG. Can these models generate global features of graph (e.g. the cell of crystal)? If so, how can I construct a new dataset of crystals to train generative models?

Thanks!

lyzustc commented 2 years ago

Hi gooaah,

Thank you for your interest in DIG! The graph generative models in DIG can only generate molecular graph structures, including atom types (e.g., carbon, oxygen) and chemical bond types. The global features like crystal cell cannot be generated in our methods.

Best regards, Youzhi Luo

gooaah commented 2 years ago

Thanks!