s-tdep vasp calculation of forces

[Vahid999]

Dear Users,

The online documentation on s-tdep uses machine learning potential for the examples. I am using VASP.

Once the configurations are generated, do these configurations need to be relaxed in anyway? Such as atom-only relaxation or volume relaxation? Or should we do a self-consistent calculation directly on the unrelaxed configurations to get the forces?

Thanks,
Vahid