help create lm and dict for pocketsphix

[nguyen duy nam]

hi, guys
i'm new with pocketsphinx, i installed pocketsphinx 0.7 on my computer(os is
fedora 16 64bit) and i want to create a small dict for it follow steps:
create text file corpus.txt:

open browser
new e-mail
forward
backward
next window
last window
open music player

and excute command

text2wfreq.exe < etc\sent.txt > etc\project.wfreq
wfreq2vocab.exe -top 20000 < etc\project.wfreq > etc\project0.vocab
text2idngram.exe -vocab etc\project.vocab < etc\sent.txt > etc\project.idngram
idngram2lm.exe -idngram etc\project.idngram -vocab etc\project.vocab -binary
etc\project.binlm
binlm2arpa -binary etc\project.binlm -arpa etc\project.arpa
mkdir dmp
lm3g2dmp etc\project.arpa dmp
cp dmp/project.arpa.DMP etc/project.DMP

and create dict

perl scripts_pl/make_pronunciation.pl \
-tools "{your path}//trunk//logios//Tools" \
-words project.vocab \
-dict project.dic \
-handdict {your-words}.handdict \
-dictdir etc
cp etc/project.dic etc/NewVoiceProject.dic

and i run command:

pocketsphinx_continuous -lm /Data/dataModellanguale/dmp/corpus.arpa.DMP -dict
/Data/dataModellanguale/corpus.dic

but when i speak it not found any word, but when i run pocketsphinx_continuous
so it run ormal.
anybody can help me that my steps where worng.
thanks

[Nickolay V. Shmyrev]

and i want to create a small dict for it follow steps:

You need to write jsgf grammar instead of the language model. You need to use
existing dictionary, you do not need to generate a new one. You shouldn't use
lm3g2dmp, instead you should use sphinx_lm_convert. See the tutorial for
details

http://cmusphinx.sourceforge.net/wiki/tutoriallm

anybody can help me that my steps where worng.

In order to enable us to help you need to provide the files you created. You
can pack them to archive and upload to a public file sharing resource. Don't
forget to give a link after that.

[nguyen duy nam]

i need create new dict because i want to dic on other language not english.
can anybody help me?

[Nickolay V. Shmyrev]

i need create new dict because i want to dic on other language not english.
can anybody help me?

This issue is covered in tutorial

[nguyen duy nam]

hi nshmyrev,
i used sphinx_lm_convert to create file DMP, but not ok,
and i want to create a application speech to text in vietnammese not english
so i neet know steps create model language and acoutis and dict, and to
understand them i tried do small step but i meet dificute can you help me.

[Nickolay V. Shmyrev]

nd to understand them i tried do small step but i meet dificute can you help
me.

In order to let me help you need to ask more exact questions. You also need to
read the tutorial first.

[nguyen duy nam]

thank very much nshmyrev.
i create file DMP as your tutorial http://cmusphinx.sourceforge.net/wiki/tuto
riallm
when i run: pocketsphinx_continuous -lm /Data/weather/weather.lm.DMP
it ok.
but i want to create dict and only use the dict so i make:

../pocketsphinxdemo/cmusphinx/trunk/logios/Tools/MakeDict/make_pronunciation.p
l -tools "/Data/pocketsphinxdemo/cmusphinx/trunk/logios/Tools" -words
weather.vocab -dict weather.dic -handdict weather.handdic -dictdir
/Data/weather

then i run:

pocketsphinx_continuous -lm /Data/weather/weather.lm.DMP -dict
/Data/weather/weather.dic

result when i speak it not return any work.
i not right in anywhere, can you show me the steps to create dict and using it
in pocketsphinx 0.7.
thanks becase you get me your time.

[Nickolay V. Shmyrev]

i not right in anywhere, can you show me the steps to create dict and using
it in pocketsphinx 0.7.

Steps are described in tutorial

http://cmusphinx.sourceforge.net/wiki/tutorialdict

[nguyen duy nam]

hi nshmyrev.
i'm creating AM, and i want to know that if i had file .dic, do have anytool
create file .phone from the dic file?
if yes, can you show me. thanks.
and i sorry because i don't care about my post so you warring

[Nickolay V. Shmyrev]

To create a phone file you can use script make_phoneset.pl from the
distribution

perl make_phoneset.pl etc/your.dic etc/your.filler > etc/your.phone

[nguyen duy nam]

thanks you indeed

[nguyen duy nam]

hi, when i run perl scripts_pl/20.ci_hmm/slave_convg.pl
in terminal:
**

perl ./scripts_pl/20.ci_hmm/slave_convg.pl
MODULE: 20 Training Context Independent models
Phase 1: Cleaning up directories:
accumulator...logs...qmanager...models...
Phase 2: Flat initialize
Phase 3: Forward-Backward
Baum welch starting for 1 Gaussian(s), iteration: 1 (1 of 1)
0%
This step had 6 ERROR messages and 0 WARNING messages. Please check the log
file for details.
Only 0 parts of 1 of Baum Welch were successfully completed
Parts 1 failed to run!
Training failed in iteration 1

and veiwer in logdir:
file words.1.1-1.bw have error:

INFO: main.c(432): Reading filler lexicon:
/Application/Data/speechtotext/demo/datatest3/etc/words.filler
ERROR: "hash.c", line 263: Hash table lex: Duplicate definition:
FATAL_ERROR: "lexicon.c", line 70: hash add failed
Sat Feb 4 13:40:32 2012

during file filler:

SIL
SIL
<sil> SIL
**
i don't understand it, can you show me to solve it. pls help</sil>

[Nickolay V. Shmyrev]

You need to remove from the main dictionary.