Menu

language models and sphinx2-demo

Help
2000-12-28
2012-09-22
  • Jeremy Impson

    Jeremy Impson - 2000-12-28

    Hi.

    I created a new language model with http://www.speech.cs.cmu.edu/tools/lmtool.html and unpacked the resultant tar.gz file into a directory (lm4795).  If I run "sphinx2-demo" alone it works, but if I run "sphin2-demo lm4795" is craps out after printinf "[initializing]".  I looked at the perl code, and set it to print debugging statements, and it says:

        kb.c(345): Reading phone file [/usr/local/share/sphinx2/model/hmm/6k/phone]
        Using 0 phonological mappings
        kb.c(355): Reading map file [/usr/local/share/sphinx2/model/hmm/6k/map]
        Read Map: 125716 phones map to 23355 unique senone sequences
        kb.c(360): Reading dict file [lm4795/4795.dic]
        hash_lookup: Bad Argument [(null)]
        dict_load: first line of lm4795/4795.dic was "6, expecting '!'
        dict_load: will assume first line contains a word

    ...and then it returns..

    Any thoughts??

    --Jeremy

    P.S. anybody here know John Moody??  I work with him...

     
    • Kevin A. Lenzo

      Kevin A. Lenzo - 2001-07-10

      Hi,

      Email me directly (lenzo@cs.cmu.edu) and I can help you on this a bit more.  I'll want to get the original corpus you were uploading to the lmtool.

      kevin

       
    • Ng Hong I

      Ng Hong I - 2002-08-02

      I wanted to create a new language model from the TRAINS corpus.

      To make sure that SimpleLM.pl and CMU-Cam_Toolkit_v2 are working, I have re-compiled turtle.corpus that comes together with sphinx2-0.4.tar.gz and tried using sphinx2-simple. No problem occurs.

      So, I created a new language model from one of the TRAINS corpus using SimpleLM.pl and CMU_Cam_Toolkit_v2 and tried using sphinx2-simple. The following error message occurs:

      READY....
      Listening...
      Stopped listening, please wait...
      /home/nghongi/bin/sphinx2-simple1: line 19: 14996 Segmentation fault      (core dumped) $S2CONTINUOUS -live TRUE -ctloffset 0 -ctlcount 100000000 -cepdir ${TASK}/ctl -datadir ${TASK}/ctl -agcemax TRUE -langwt 6.5 -fwdflatlw 8.5 -rescorelw 9.5 -ugwt 0.5 -fillpen 1e-10 -silpen 0.005 -inspen 0.65 -top 1 -topsenfrm 3 -topsenthresh -70000 -beam 2e-06 -npbeam 2e-06 -lpbeam 2e-05 -lponlybeam 0.0005 -nwbeam 0.0005 -fwdflat FALSE -fwdflatbeam 1e-08 -fwdflatnwbeam 0.0003 -bestpath TRUE -kbdumpdir ${TASK} -lmfn ${LM} -dictfn ${DICT} -noisedict ${HMM}/noisedict -phnfn ${HMM}/phone -mapfn ${HMM}/map -hmmdir ${HMM} -hmmdirlist ${HMM} -8bsen TRUE -sendumpfn ${HMM}/sendump -cbdir ${HMM}
      The error occurs here:

      #0  0x4003e4f8 in word_transition () at search.c:1649
      1649        if ((rhmm->active < cf) || (rhmm->score[0] < newscore)) {
      (gdb) where
      #0  0x4003e4f8 in word_transition () at search.c:1649
      #1  0x4003f5d5 in search_one_ply_fwd () at search.c:2196
      #2  0x4003f09e in search_fwd (cep=0x419970d8, dcep=0x419e40d8,
          dcep_80ms=0x41a310d8, pcep=0x83e5e00, ddcep=0x41a7e0d8) at search.c:2023
      #3  0x4004b59a in uttproc_frame () at uttproc.c:689
      #4  0x4004c0c5 in uttproc_rawdata (raw=0xbfffc7a0, len=0, block=0)
          at uttproc.c:1016
      #5  0x08048e6f in utterance_loop () at tty-continuous.c:166
      #6  0x080490e7 in main (argc=73, argv=0xbfffe834) at tty-continuous.c:224
      #7  0x40120507 in __libc_start_main (main=0x8049044 <main>, argc=73,
          ubp_av=0xbfffe834, init=0x8048934 <_init>, fini=0x8049150 <_fini>,
          rtld_fini=0x4000dc14 <_dl_fini>, stack_end=0xbfffe82c)
          at ../sysdeps/generic/libc-start.c:129
      (gdb) print rhmm
      $1 = (ROOT_CHAN_T *) 0x0
      (gdb)

      Any idea what happened?

       

Log in to post a comment.

Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.