You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am using ALIGNN for a DFT dataset and need to include stress term in the training. I am using ase to read VASP OUTCAR and converting the stress into GPa using the formula below. int_stress = traj[0].get_stress(voigt=False)*(-160.21766208)
I am hoping this should convert the DFT stress values to be aligned with as expected by ALIGN model. Can you please confirm if that's the case or I should use different formula to pull the stress values from the VASP.
As a reference I am also attaching code snippet from AlignAtomwiseCalculator.
Hi,
I am using ALIGNN for a DFT dataset and need to include stress term in the training. I am using ase to read VASP OUTCAR and converting the stress into GPa using the formula below.
int_stress = traj[0].get_stress(voigt=False)*(-160.21766208)
I am hoping this should convert the DFT stress values to be aligned with as expected by ALIGN model. Can you please confirm if that's the case or I should use different formula to pull the stress values from the VASP.
As a reference I am also attaching code snippet from AlignAtomwiseCalculator.
stress = ( -1 * 160.21766208 * ( torch.matmul(r.T, pair_forces) # / (2 * g.edata["V"]) / (2 * g.ndata["V"][0]) ) )
The text was updated successfully, but these errors were encountered: