Consult for the use of the BandUP program

[Lamed-git]

I'm new to BandUP and I'm trying to learn and apply it to my own flaw system but I find that my calculations are always not satisfactory. The results are roughly as follows:
I found that the delta_N in the figure is always small and there are some line segments that should not be in the energy band of the primitive cell. My calculation steps are roughly as follows:

1. Use ATAT or supercell to build a crystal model
2. Use VASP fixed cell parameters for optimization (ISIF=2)
3. The optimized structure is statically calculated to obtain CHGCAR
4. Use seekpath to find the high symmetry point of the primitive cell and use BandUP to generate the irreducible K points required for unfolding
5. Read the statically calculated CHGCAR and then calculate the WAVECAR
6. BandUP post-processing to get the result of unfolding
   Also my defect concentration is 2/27, not too big. I'm not sure if this is a problem with my crystal models or a problem with calculation processing, I hope you can help me, thank you very much！