Hessian error

[VishnuEriyat]

I got this error (below given) when I was trying to model rifampicin and doing the NPAG run. I specified 10000 cycles, but the run stopped at around 6000 with this error. I restarted the run using this as the prior and it again got stopped at around 2500 cycles.
Im not able to understand this error, can anybody please help.

Hessian matrix in interior point EM algorithm
is singular. Possibly number of grid points is too small,
or assay coefficients are too large.

Try again with a new assay polynomial or larger grid.

Suggested quick fix: rerun and select error model 2)
in response to the initial question; then enter a
initial value gamma = 10.0 in response to the prompt for
that value.

THE PROGRAM WILL CREATE OUTPUT FILES BEFORE STOPPING.

NOTE THAT IN SUBROUTINE emint, THE VALUE OF INFO
IS 110

IF THIS VALUE IS POSTIVE, IT IS LIKELY THE NO. OF THE SUBJECT
(OR AT LEAST THE FIRST SUBJECT) WHICH CAUSED THE HESSIAN
ERROR. SO IN THIS CASE, YOU MIGHT ALSO WANT TO EXAMINE THE
DATA IN THIS SUBJECT TO VERIFY THEY ARE CORRECT.