st: question on gllamm with ip(f), the discrete latent variable

[Hey Sky <heyskywalker@yahoo.com>]

hey, All

I am new to gllamm and have read the manual and some other documents for it,
such as A review of random effects modelling using gllamm in Stata by Grilli, L. 
and Rampichini on gllamm.org website. 

I am modeling people training behavior base on their formal education history, ie 
ed9, acedu, etc, and accumulative on/off job training history, ie acojt, acoft, etc
 with panel data. I assume there is one latent variable, epsilon, with two types. 
thus every individual  has a probability belong to each type. and I have code like 
below, which different choice has different explain variables controlled by constr. 

now my question is, everytime I try to run the gllamm, it will give me the 
following warning:

"specify at least one file or folder to include in the zip file
(error occurred in ML computation)
(use trace option and check correctness of initial model)"

I have tried Grilli, L's code and data set for discrete latent variable, which are given
 on gllamm.org, but the same warning appears and the program cannot continue. 
the trace does not help or maybe I skiped the message.

I know the question should be a very basic but why is that and how to solve it? 

any hints are appreciated and thanks a lot in advance



code:
local contr234 "ed9 ed10 ed11 ed12 ed13 ed14 ed15 ed16 ed17"
constr def 1 [c1]: acedu acexp acojt acoft
constr def 2 [c2]: `contr234'
constr def 3 [c3]: `contr234'
constr def 4 [c4]: `contr234'
 
gen cons=1
eq mu1: cons

gllamm choice ed9-ed18 xvar2-xvar7 xvar56 acedu acexp acojt acoft, ///
 i(id) base(5) link(mlogit) family(binom) constr(1 2 3 4) nip(2) ip(f) nrf(1) ///
 eq(mu1) dots trace




*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/