Box explosion in NPT simulation #421
Unanswered
yangqifanshihaoren
asked this question in
Q&A
Replies: 2 comments 1 reply
-
I didn't use stress as a loss function but only energy and force since I haven't figure out the unit of stress I should use clearly. But I used the "StressForceOutput" so the stress is derived from energy and force. |
Beta Was this translation helpful? Give feedback.
1 reply
-
@Linux-cpp-lisp Could you please answer my quetion? Thank you very much! |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
When I use the trained model in lammps for npt simulation to calculate shear stress-strain curve, I found the box always explode to an extremely large volume(about 1000 times the previous amount). The training dataset is obtained by relaxing progress in vasp with "PSTRESS = 10" and "PSTRESS = -10" to get structures of different volumes. Is that because the training dataset is too little? Or because when using "erate" in lammps, the nequip trained model can't perform well? Thank you!
The following picture is the input file of the npt simulation.
Beta Was this translation helpful? Give feedback.
All reactions