Images keep moving up to higher energies and forces

[ooter]

Dear Prof. Henkelman and Community,

I manually construct the initial path by linear interpolation. The first 30 or so iterations appear to move the path towards the MEP judging by the energies and forces of the images. But after some iterations, the fifth image starts moving up to higher energies and eventually drags all the other images away from a path that appears closer to the MEP. The fifth image is next to one of the endpoints of the path. The shift to higher energies does not seem to stop but appears to make no sense as both the energies and forces are going up for all the images. Does this type of instability ever happen? Should I keep running in the hope that the system will eventually self-correct? Or am I doing something wrong? My input files are attached. I am new to NEB calculations.

Thanks for your help!