[hts-users:00015] Re: FATAL error running demo

["Elliot Moore" <em80@xxxxxxxxxxxxxxx>]

Hi,

You are correct about the architecture and yes I did change that the
BYTESWAP value to 1 in that Makefile.  This is a very strange error.  It
would seem that even if the byteswap was the error then the files would
still process but just have nothing but garbage output.

----- Original Message -----
From: "Heiga ZEN" <zen@xxxxxxxxxxxxxxxx>
To: <hts-users@xxxxxxxxxxxxxxxxxxxxxxxxx>
Sent: Thursday, March 04, 2004 9:29 PM
Subject: [hts-users:00014] Re: FATAL error running demo


> Hi Elliot,
>
> I think you were running this demonstration on SPARC/Soraris architecture.
> Did you run HTS-demo_CMU-ARCTIC-AWB/Makefile with BYTESWAP=1 ?
>
> On Mon, 1 Mar 2004 12:41:41 -0500
> "Elliot Moore" <em80@xxxxxxxxxxxxxxx> wrote:
>
> > I have tried unsuccessfully to run the HTS demo on a unix-based machine.
> > The process gets to the same point in the script before stopping.
> > Here is what is output to the screen before it ceases processing (where
<dir> is just the directory location on my account):
> >
> > Iteration 2 of Embedded Re-estimation
> > <dir> /bin.sun4_solaris/HERest -A -B -C <dir>
/HTS-demo/configs/default.conf -D -T 1 -m 1 -u tmvw -t 1.0e+20 -w 3 -H <dir>
/HTS-demo/hmms/cmu_us_arctic/awb/cmp/fullcontext_cmp.mmf -I
<dir>/HTS-demo_CMU-ARCTIC-AWB/labels/full_awb.mlf -M <dir>
/htk/HTS-demo/hmms/cmu_us_arctic/awb/cmp -S
<dir>/HTS-demo_CMU-ARCTIC-AWB/lists/data_cmu_us_arctic_awb.scp
<dir>/HTS-demo_CMU-ARCTIC-AWB/lists/full_awb.list
> >
> > HTK Configuration Parameters[5]
> >   Module/Tool     Parameter                  Value
> > #                 NATURALWRITEORDER              TRUE
> > #                 NATURALREADORDER              TRUE
> > #                 DVFLOORSCALE            0.010000
> > #                 APPLYVFLOOR                 TRUE
> > #                 TREEMERGE                  FALSE
> >
> > HERest  Updating: Transitions Means Variances MixWeights
> >
> >  System is PLAIN
> > 40539 Logical/40539 Physical Models Loaded, VecSize=78
> > 1 MMF input files
> > Pruning-Off
> >  Processing Data: cmu_us_arctic_awb_a0001.cmp; Label
cmu_us_arctic_awb_a0001.lab
> >  Utterance prob per frame = 2.532082e+11
> >   ERROR [+7391]  StepAlpha: Forward/Backward Disagree
202060105613284.500000/202060105613283.531250
> >  FATAL ERROR - Terminating program <dir>/bin.sun4_solaris/HERest
> >
> > The HTK book says that this error is related to a "Sanity Check" that is
not correctable by the user.
> > Any suggestions on what I should do now?
>
> --
>  ------------------------------------------------
>   Heiga Zen
>   Department of Computer Science and Engineering
>   Graduate School of Engineering
>   Nagoya Institute of Technology
>   Japan
>
>   e-mail: zen@xxxxxxxxxxxxxxxx
>      web: http://kt-lab.ics.nitech.ac.jp/~zen
>  ------------------------------------------------
>