[MLFF error] Inconsistent initialization of ion. and lat. DOFs

[exk301]

Hello!

I'm running MLFF calculation using vasp6.3.2.For some reason, My job ends with the error below:

"Inconsistent initialization of ion. and lat. DOFs |
| |
| ----> I REFUSE TO CONTINUE WITH THIS SICK JOB ... BYE!!! <---- "

I had my ML_FFN and ML_ABN files generated from the previous run (72Pd atoms, 1Cl). Then, I brought the force field and the configurations (ML_FFN, ML_ABN) to run the new POSCAR file (72 Pd, 2Cl this time) and goes through the same calculation. And My new POSCAR file has two bottoms layers fixed. However, I have the error above, and my job dies.

I wonder if this is because I don't have enough relaxed atoms to be used for the MLFF? When I relax all my slab layers with the same setting(when I do selective dynamiacs T for all atoms), this issue has been resolved.

Any advice would be appreciated!

Thank you for your time!

[ferenc_karsai]

This error is not related to MLFF but to molecular dynamics. You have most likely changed some parameters and now the predictor/corrector entries in the POSCAR file are incompatible. Just delete everything after the Wycoff positions in the POSCAR file.