[hts-users:02255] Re: result for mget [last:10 MIME/multipart] (2/2) (hts-users ML)

[谢旭 <xiexu2008@xxxxxxxxx>]

Hi all,

I attempted to install the HTS, but some problems appeared. Please help me. The details are following:

After the "make all" step, I did the "make install" step. When the "make install" finished, I inputed "HInit' just to check whether the installation was ok. The result was "bash: HInit: command not found".

The following is the scripts:

xiexu@xiexu-laptop:~/Desktop/htk$ sudo make install

(cd HTKTools && make all) \
      || case "" in *k*) fail=yes;; *) exit 1;; esac;
make[1]: Entering directory `/home/xiexu/Desktop/htk/HTKTools'
make[1]: Nothing to be done for `all'.
make[1]: Leaving directory `/home/xiexu/Desktop/htk/HTKTools'
(cd HTKTools && make install) \
    || case "" in *k*) fail=yes;; *) exit 1;; esac;
make[1]: Entering directory `/home/xiexu/Desktop/htk/HTKTools'
if [ ! -d /usr/local/HTS-2.1/bin ] ; then mkdir -p /usr/local/HTS-2.1/bin ; fi
for program in HSLab HBuild HCompV HCopy HDMan HERest HHEd HInit HLEd     HList HLRescore HLStats HMMIRest HParse HQuant HRest HResults HSGen HSmooth HVite HMGenS ; do /usr/bin/install -c -m 755 ${program} /usr/local/HTS-2.1/bin ; done
make[1]: Leaving directory `/home/xiexu/Desktop/htk/HTKTools'

xiexu@xiexu-laptop:~/Desktop/htk$ HInit
bash: HInit: command not found

Thanks 
Xu

2009/10/28 <hts-users-admin@xxxxxxxxxxxxxxx>

---------- 已转发邮件 ----------
From: "Lee Siu Wa" <yswlee@xxxxxxxxxxxxxx>
To: <hts-users@xxxxxxxxxxxxxxx>
Date: Mon, 26 Oct 2009 00:03:52 +0800
Subject: [hts-users:02251] duration unit

Hi,

 

    Does anyone know the unit of the duration models generated from HTS? i.e. if

 

~h "sil-E_ow+E_m"
<BEGINHMM>
<NUMSTATES> 3
<STATE> 2
<STREAM> 1
<MEAN> 1
 2.947347e+000
<VARIANCE> 1
 5.012488e-002
<GCONST> -1.155361e+000
<STREAM> 2
<MEAN> 1
 1.016725e+000
<VARIANCE> 1
 1.645324e-002
<GCONST> -2.269356e+000
<STREAM> 3
<MEAN> 1
 1.998965e+001
<VARIANCE> 1
 2.305229e-002
<GCONST> -1.932113e+000
<STREAM> 4
<MEAN> 1
 4.399481e+001
<VARIANCE> 1
 5.263355e-003
<GCONST> -3.409109e+000
<STREAM> 5
<MEAN> 1
 5.527111e+000
<VARIANCE> 1
 2.557698e-001
<GCONST> 4.743998e-001

 

    Does the mean of stream 1 refer to 2.947 msec or in other unit?  Thanks in advance.

 

Regards,

Lee Siu Wa

 

---------- 已转发邮件 ----------
From: "Nickolay V. Shmyrev" <nshmyrev@xxxxxxxxx>
To: hts-users@xxxxxxxxxxxxxxx
Date: Sun, 25 Oct 2009 22:55:40 +0300
Subject: [hts-users:02252] Re: duration unit
В Пнд, 26/10/2009 в 00:03 +0800, Lee Siu Wa пишет:
> Hi,
>
>     Does anyone know the unit of the duration models generated from
> HTS? i.e. if
>
> ~h "sil-E_ow+E_m"
> <BEGINHMM>
> <NUMSTATES> 3
> <STATE> 2
> <STREAM> 1
> <MEAN> 1
>  2.947347e+000
> <VARIANCE> 1
>  5.012488e-002
> <GCONST> -1.155361e+000
> <STREAM> 2
> <MEAN> 1
>  1.016725e+000
> <VARIANCE> 1
>  1.645324e-002
> <GCONST> -2.269356e+000
> <STREAM> 3
> <MEAN> 1
>  1.998965e+001
> <VARIANCE> 1
>  2.305229e-002
> <GCONST> -1.932113e+000
> <STREAM> 4
> <MEAN> 1
>  4.399481e+001
> <VARIANCE> 1
>  5.263355e-003
> <GCONST> -3.409109e+000
> <STREAM> 5
> <MEAN> 1
>  5.527111e+000
> <VARIANCE> 1
>  2.557698e-001
> <GCONST> 4.743998e-001
>
>     Does the mean of stream 1 refer to 2.947 msec or in other unit?
> Thanks in advance.

Most likely it's 2.947 frames, each frame traditionally 5 ms. So it's
14.6 ms.

See in HTS patch:

+static void GetLabStateDurations (GenInfo *genInfo)
...
+      genInfo->durations[j][k] = (int)((label->end -
label->start)/genInfo->frameRate+diff+0.5);
+



---------- 已转发邮件 ----------
From: Keiichiro Oura <uratec@xxxxxxxxxxxxxxx>
To: hts-users@xxxxxxxxxxxxxxx
Date: Mon, 26 Oct 2009 10:14:20 +0900
Subject: [hts-users:02253] Re: Error in train transitions matrix
Hi,

Technically, state alignment using HSMM is not implemented.
Your .hed file converted HSMMs to HMMs.
Therefore, '-R dur.mmf', '-N durdir', and 'durlist' are not required for HERest.

Regards,
Keiichiro Oura

Xingchi Xian wrote:

Hi all,

I want use the HSMM I've trained to do state alignment, thus I use Simon's recommendation to deal with this issue. I do this as follows.

1.Untie the shared transition stuctrue and retie it in terms of each phone, my .hed file is:
TR 2
UT {*.transP}
TI T_a {(*-a+*).transP}
TI T_ai {(*-ai+*).transP}
TI T_an {(*-an+*).transP}
TI T_ang {(*-ang+*).transP}
TI T_ao {(*-ao+*).transP}
...
AT 2 2 0.9 {*.transP}
AT 2 3 0.1 {*.transP}
AT 3 3 0.9 {*.transP}
AT 3 4 0.1 {*.transP}
AT 4 4 0.9 {*.transP}
AT 4 5 0.1 {*.transP}
AT 5 5 0.9 {*.transP}
AT 5 6 0.1 {*.transP}
AT 6 6 0.9 {*.transP}
AT 6 7 0.1 {*.transP}

There is no problem at this stage. The model file is stored in untiedt.mmf
2. Retrain the transition matrix using the commond below.

/usr/local/HTS-2.1/bin/HERest -A -B -C xxx/configs/trn.cnf -D -T 1 -S xxx/data/scp/train.scp -I xxx/data/labels/full.mlf -m 1 -u t -w 3 -t 1500 100 5000 -H xxx/models/qst001/ver1/cmp/untiedt.mmf -N xxx/models/qst001/ver1/dur/untied.mmf -M xxx/models/qst001/ver1/cmp -R xxx/models/qst001/ver1/dur -C xxx/configs/nvf.cnf -s xxx/stats/qst001/ver1/cmp.stats.untiedt -w 0.0 xxx/data/lists/full.list xxx/data/lists/full.list
However, it shows that  Processing Data: xxx_0001.cmp; Label xxx_0001.lab
Retrying Beta pass at 1600.0
Retrying Beta pass at 1700.0
Retrying Beta pass at 1800.0
Retrying Beta pass at 1900.0
Retrying Beta pass at 2000.0
Retrying Beta pass at 2100.0
Retrying Beta pass at 2200.0
Retrying Beta pass at 2300.0
...
 WARNING [-7324]  StepBack: Bad data or over pruning
ERROR [+7390]  StepAlpha: Alpha prune failed sq(43) > qHi(42) at time 1109

What's wrong with my manipulation? (Machine: 64bit Linux, 16GB mem)

Thanks for help!

Xian Xingchi

---------- 已转发邮件 ----------
From: Xingchi Xian <yeahxxc@xxxxxxxxx>
To: hts-users@xxxxxxxxxxxxxxx
Date: Wed, 28 Oct 2009 00:38:51 +0800
Subject: [hts-users:02254] Re: Error in train transitions matrix
Hi Keiichiro,

Thanks for your reply. However, the same error still occurs when I use command below,

/usr/local/HTS-2.1/bin/HERest -A -B -C xxx/configs/trn.cnf -D -T 1 -S xxx/data/scp/train.scp -I xxx/data/labels/full.mlf -m 1 -u t -w 3 -t 1500 100 5000 -H xxx/models/qst001/ver1/cmp/untied.mmf -M xxx/models/qst001/ver1/cmp -C xxx/configs/nvf.cnf -s xxx/stats/qst001/ver1/cmp.stats.untied -w 0.0 xxx/data/lists/full.list

Remove the -t command doesn't help either. Is there any other thing I have to do to turn off the HSMM BW training process or should I use a old version of HTS?

Many thanks.

2009/10/26 Keiichiro Oura <uratec@xxxxxxxxxxxxxxx>

Hi,

Technically, state alignment using HSMM is not implemented.
Your .hed file converted HSMMs to HMMs.
Therefore, '-R dur.mmf', '-N durdir', and 'durlist' are not required for HERest.

Regards,
Keiichiro Oura

Xingchi Xian wrote:

Hi all,

I want use the HSMM I've trained to do state alignment, thus I use Simon's recommendation to deal with this issue. I do this as follows.

1.Untie the shared transition stuctrue and retie it in terms of each phone, my .hed file is:
TR 2
UT {*.transP}
TI T_a {(*-a+*).transP}
TI T_ai {(*-ai+*).transP}
TI T_an {(*-an+*).transP}
TI T_ang {(*-ang+*).transP}
TI T_ao {(*-ao+*).transP}
...
AT 2 2 0.9 {*.transP}
AT 2 3 0.1 {*.transP}
AT 3 3 0.9 {*.transP}
AT 3 4 0.1 {*.transP}
AT 4 4 0.9 {*.transP}
AT 4 5 0.1 {*.transP}
AT 5 5 0.9 {*.transP}
AT 5 6 0.1 {*.transP}
AT 6 6 0.9 {*.transP}
AT 6 7 0.1 {*.transP}

There is no problem at this stage. The model file is stored in untiedt.mmf
2. Retrain the transition matrix using the commond below.

/usr/local/HTS-2.1/bin/HERest -A -B -C xxx/configs/trn.cnf -D -T 1 -S xxx/data/scp/train.scp -I xxx/data/labels/full.mlf -m 1 -u t -w 3 -t 1500 100 5000 -H xxx/models/qst001/ver1/cmp/untiedt.mmf -N xxx/models/qst001/ver1/dur/untied.mmf -M xxx/models/qst001/ver1/cmp -R xxx/models/qst001/ver1/dur -C xxx/configs/nvf.cnf -s xxx/stats/qst001/ver1/cmp.stats.untiedt -w 0.0 xxx/data/lists/full.list xxx/data/lists/full.list
However, it shows that  Processing Data: xxx_0001.cmp; Label xxx_0001.lab
Retrying Beta pass at 1600.0
Retrying Beta pass at 1700.0
Retrying Beta pass at 1800.0
Retrying Beta pass at 1900.0
Retrying Beta pass at 2000.0
Retrying Beta pass at 2100.0
Retrying Beta pass at 2200.0
Retrying Beta pass at 2300.0
...
 WARNING [-7324]  StepBack: Bad data or over pruning
ERROR [+7390]  StepAlpha: Alpha prune failed sq(43) > qHi(42) at time 1109

What's wrong with my manipulation? (Machine: 64bit Linux, 16GB mem)

Thanks for help!

Xian Xingchi