That was it! I was using the old version from the voxforge quickstart. I downloaded and installed Julius 4.1.2, ran yum install libesd.so.0, and your command ran fine for me. Thanks again!

--- (Edited on 5/11/2009 2:54 pm [GMT-0500] by Visitor) ---

I have a similar need and have followed the instructions in this thread.
I created a bin file by using  

         mkbingram -nlr lm_giga_64k_nvp_3gram.arpa julius.bin

But when I run julius with -d with the above bin and -v with the dic file from the same LM dir.

Julius wants an hmmdef_file.

Can anyone tell me how to use the giga LM including the recipe for all the files?

Thankx

--- (Edited on 5/27/2009 2:11 pm [GMT-0500] by kburden000) ---

hmmdef file you can dowload here as a part of Voxforge acoustic model. It's not related to a language model at all.

--- (Edited on 5/28/2009 4:55 pm [GMT-0500] by nsh) ---

Hello I'm getting sil word not exist error:

mertyildiran@Corsair:~/Downloads/lm_giga_5k_nvp_3gram$ julius -d julius.bin -v lm_giga_5k_nvp.sphinx.dic -h hmmdefs -hlist tiedlist 
STAT: jconf successfully finalized
STAT: *** loading AM00 _default
Stat: init_phmm: Reading in HMM definition
Stat: rdhmmdef: ascii format HMM definition
Stat: rdhmmdef: limit check passed
Stat: check_hmm_restriction: an HMM with several arcs from initial state found: "sp"
Stat: rdhmmdef: this HMM requires multipath handling at decoding
Stat: init_phmm: defined HMMs: 16032
Stat: init_phmm: loading ascii hmmlist
Stat: init_phmm: logical names: 24402 in HMMList
Stat: init_phmm: base phones:    41 used in logical
Stat: init_phmm: finished reading HMM definitions
STAT: making pseudo bi/mono-phone for IW-triphone
Stat: hmm_lookup: 799 pseudo phones are added to logical HMM list
STAT: *** AM00 _default loaded
STAT: *** loading LM00 _default
Stat: init_voca: read 5882 words
ERROR: m_fusion: head sil word "<s>" not exist in voca
ERROR: m_fusion: failed to initialize dictionary
ERROR: Error in loading model
mertyildiran@Corsair:~/Downloads/lm_giga_5k_nvp_3gram$

--- (Edited on 4/20/2016 9:47 am [GMT-0500] by Visitor) ---

Cross posting. :(

--- (Edited on 2016-04-21 5:51 am [GMT-0400] by colbec) ---

I met the same trouble as you described here. I am trying to train an Acoustic Model in other language which is out of the lists given in this web. and I want to test my model,as i know, it must give HVite  a wdnet (grammar) ,or word network. i still can not figure it out.

so , i wonder if you got your problems solved? and how?

besides, i want to know how does n-gram,trained by HTK, used in my task?

many thanks,and waiting for your reply.

best regards.

Eric.

--- (Edited on 4/19/2012 8:09 pm [GMT-0500] by ericsiat) ---

--- (Edited on 4/19/2012 8:11 pm [GMT-0500] by ericsiat) ---

>so , i wonder if you got your problems solved? and how?

The solution would be to use decoder with n-gram model

> besides, i want to know how does n-gram,trained by HTK, used in my task?

Julius can decode with arpa models, see the parameter list of julius

HTK also provides large vocabulary decoder called HDecode, you can find details on it in HTK Book.

--- (Edited on 4/22/2012 14:45 [GMT+0400] by nsh) ---