I hope you’re doing well. DeepMD-trained neural network potentials can be used in various computational software, but I haven’t found a way to implement them in GULP yet. Do you know of any methods or workarounds for calling DeepMD potentials in GULP?
I would greatly appreciate your assistance on this matter!
Hi Li Ma,
At the moment DeePMD isn’t interfaced to GULP & so it would be a matter to adding this. In principle it’s possible - it’s just a matter of interfacing Fortran & C++ together. I can add it to the wish list for future work. An indirect work around is that OpenKIM is interfaced to GULP already and this provides access to neural network models. You could always train the neural net in OpenKIM to reproduce the results from DeePMD.
Best regards,
Julian