Energy increase during minimization with indenters

Dear all,

I am running minimization of a MgO nanoparticle which is strained by using two indenters. I am using lammps 16Feb16.
I got the issue that during the minimization the energy might increase when using fire or quickmin minimizers. The energy typically first decreases, then after going through a minimum it can increase by more than one eV.
I do not get this issue when using conjugate gradient but I need to use fire of quickmin since I want to run NEB calculations.

I apparently do not get this issue without the indenter.

Are the indenters and the fire minimizer incompatible?
Am I doing something wrong?
I put my input script at the end of this email.

Thank you in advance for your help,

Emiile Maras