Menu ▾ ▴

sphinx3_align.exe

Help
numb skull
2010-09-11
2012-09-22
1 2 > >> (Page 1 of 2)

  • numb skull

    numb skull - 2010-09-11

    Hi im new in sphinx. I need to force align my data however i couldn't locate
    sphinx3_align.exe.

    where can i find the file? thanks!

     

  • Nickolay V. Shmyrev

    Nickolay V. Shmyrev - 2010-09-11

    Hello

    sphinx3_align is a part of sphinx3 package. You also need sphinxbase.

     

  • numb skull

    numb skull - 2010-09-12

    im having a bit of trouble building the generated .sln file using microsoft
    visual studio 10 (for windows7). It gives the following error:

    1>InitializeBuildStatus:
    1> Touching ".\Debug\sphinx3_align.unsuccessfulbuild".
    1>ClCompile:
    1> s3_align.c
    1>j:\sphinx3\src\programs\s3_align.c(98): fatal error C1083: Cannot open
    include file: 'feat.h': No such file or directory
    1> main_align.c
    1>j:\sphinx3\include\s3types.h(87): fatal error C1083: Cannot open include
    file: 'prim_type.h': No such file or directory
    1> Generating Code...
    1>
    1>Build FAILED.

    have i missed a step in the process? thanks!

     

  • Nickolay V. Shmyrev

    Nickolay V. Shmyrev - 2010-09-12

    Yes, you need to compile sphinxbase first. See README file

     

  • numb skull

    numb skull - 2010-09-13

    I have already compiled sphinxbase. Could you please give me a step by step
    procedure on how to build it to output sphinx3_align.exe? thanks so much!!!

     

  • numb skull

    numb skull - 2010-09-19

    Im getting the following error when im trying to build sphinx3_align. i dont
    understand where the error comes from. what should i do to resolve the
    following errors? please help. thanks!

    1>------ Build started: Project: sphinx3_align, Configuration: Debug Win32

    1>Linking...
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__ctl_process referenced in function _main
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__model_set_mllr referenced in function _main
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__adapt_set_mllr referenced in function _main
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__adapt_am_init referenced in function _models_init
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__fast_gmm_init referenced in function _models_init
    1>main_align.obj : error LNK2019: unresolved external symbol __imp__ascr_init
    referenced in function _models_init
    1>main_align.obj : error LNK2019: unresolved external symbol __imp__dict_init
    referenced in function _models_init
    1>main_align.obj : error LNK2019: unresolved external symbol __imp__s3_am_init
    referenced in function _models_init
    1>main_align.obj : error LNK2019: unresolved external symbol __imp__logs3_init
    referenced in function _models_init
    1>main_align.obj : error LNK2019: unresolved external symbol __imp__New_kbcore
    referenced in function _models_init
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__kbcore_free referenced in function _models_free
    1>main_align.obj : error LNK2019: unresolved external symbol __imp__dict_free
    referenced in function _models_free
    1>main_align.obj : error LNK2019: unresolved external symbol __imp__ascr_free
    referenced in function _models_free
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__fast_gmm_free referenced in function _models_free
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__adapt_am_free referenced in function _models_free
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__approx_cont_mgau_frame_eval referenced in function _align_utt
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__approx_cont_mgau_ci_eval referenced in function _align_utt
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__s2_semi_mgau_frame_eval referenced in function _align_utt
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__ms_cont_mgau_frame_eval referenced in function _align_utt
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__mdef_phone_components referenced in function _write_s2stseg
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__mdef_ciphone_str referenced in function _write_stseg
    1>main_align.obj : error LNK2019: unresolved external symbol
    __imp__mdef_phone_str referenced in function _write_phseg
    1>s3_align.obj : error LNK2019: unresolved external symbol __imp__dict_wordid
    referenced in function _align_build_sent_hmm
    1>s3_align.obj : error LNK2019: unresolved external symbol
    __imp__mdef_phone_id_nearest referenced in function _alloc_pnode
    1>s3_align.obj : error LNK2019: unresolved external symbol __imp__logs3
    referenced in function _align_init
    1>.........\bin\Debug/sphinx3_align.exe : fatal error LNK1120: 25
    unresolved externals
    1>Build log was saved at "file://c:\cygwin\home\tutorial\sphinx3\win32\msdev\p
    rograms\align\Debug\BuildLog.htm"
    1>sphinx3_align - 26 error(s), 0 warning(s)

     

  • numb skull

    numb skull - 2010-09-20

    Hi ive done what you told me but it doesn't resolve the errors above. would be
    alright if you could send me the sphinx3_align.exe file?

    are there lattice rescoring in sphinx?

     

  • numb skull

    numb skull - 2010-09-20

    and a way to view the alignment from the trained acoustic models?

     

  • numb skull

    numb skull - 2010-09-22

    Hi, im starting to force align my data however i keep getting the following
    error:

    "$CFG_VECTOR_LENGTH requires explicit package name at
    /cygdrive/cygwin/home/.../scripts_pl/11.force_align/force_align.pl line 132
    $CFG_FEAT_WINDOW requires explicit package name at
    /cygdrive/cygwin/home/.../scripts_pl/11.force_align/force_align.pl line 132
    Execution of
    /cygdrive/cygwin/home/.../scripts_pl/11.force_align/force_align.pl aborted due
    to compilation errors
    Failed to open /cygdrive/cygwin/home/.../logdir/11.force_align/--falign.log:
    No such file or directory at scripts_pl/11.force_align/slave_align.pl at line
    222."

    ive already tried adding $CFG_FEAT_WINDOW=3; to the config file and editing
    the force_align.pl. though the force alignment has initialized, however it
    stops in part 1 ("Failed in part 1") and fails to continue.

    Please help. Thanks!

     

  • Nickolay V. Shmyrev

    Nickolay V. Shmyrev - 2010-09-22

    though the force alignment has initialized, however it stops in part 1
    ("Failed in part 1") and fails to continue.

    You need to copy sphinxbase.dll into bin folder I suppose. Otherwise the
    executable can't start.

     

  • numb skull

    numb skull - 2010-09-24

    Hi thanks for the help! ive finally generated forced aligned transcript. but i
    have a few questions

    1. where can i find the forced aligned transcript with timestamps?

    2. In the force_align.pl ive umcommented -cepwin => $ST::CFG_FEAT_WINDOW cos it couldnt locate cepwin. what do you mean by(from your previous post) "Sphinx3_align doens't support cepwin features out-of-box yet. You need to apply cepwin patch to sphinxbase in order to read cepwin features. "

    thanks!

     

  • Nickolay V. Shmyrev

    Nickolay V. Shmyrev - 2010-09-24
    1. where can i find the forced aligned transcript with timestamps?

    Run sphinx3_align without arguments to see the list of possible options. Run
    sphinx3_align with -wdsegdir, -phsegdir or -phlabdir to make it dump the word
    or phone labels

     

  • numb skull

    numb skull - 2010-09-28

    ok thanks!

    im trying to apply lda transformation ive already tried training my data
    before with the application of lda/mllt with a small amount of training data,
    and it worked out fine. however when i increased my training data, the
    training stops at module 06 mllt transformation. i cant figure out what this
    error means.

    MODULE: 06 Train MLLT transformation (2010-09-28 18:19)

    Phase 1: Cleaning up directories:
    accumulator... logs... qmanager... completed

    Phase 2: Flat initialize

    mk_flat Log File
    completed

    init_gau Log File

    FATAL_ERROR: "main.c", line 98: Failed to read LDA matrix

    This step had 1 ERROR messages and 0 WARNING messages. Please check the log
    file for details.

    from mllt logdir:

    WARN: "s3io.c", line 256: Unable to open
    /home/dsp/Desktop/cc/model_parameters/cc.lda for reading; No such file or
    directory
    ERROR: "lda.c", line 63: s3open(/home/dsp/Desktop/cc/model_parameters/cc.lda,
    rb) failed; No such file or directory
    FATAL_ERROR: "main.c", line 98: Failed to read LDA matrix

    please help thanks!

     

  • Nickolay V. Shmyrev

    Nickolay V. Shmyrev - 2010-09-28

    Hello

    If lda matrix is missing it means it was not created on a previous stage (lda
    step). You need to check previous stage logs to find out what was wrong.

     

  • numb skull

    numb skull - 2010-09-28

    yes, but it only says

    Only 0 parts of 1 of Baum Welch were successfully completed
    Parts 1 failed to run!

    in one of the log files. . . Could you please tell me what could be the
    possible reasons why it failed to run? Thanks for the reply!

     

  • Nickolay V. Shmyrev

    Nickolay V. Shmyrev - 2010-09-28

    Logfiles are located in logdir folder. You need to read them carefully or
    share them. You can pack them into archive and upload to public file hosting
    service. Don't forget to give a link on the files here after that.

     

  • numb skull

    numb skull - 2010-09-28

    here's the link to the logdir folder:

    http://www.mediafire.com/?sybd1c2vwzuaopy

    please help me figure. . . thanks!

     

  • Nickolay V. Shmyrev

    Nickolay V. Shmyrev - 2010-09-28

    Hello

    According to the log bw process crashed when it reached line number 21 or line
    number 22 in your transcription file. It means something is wrong with that
    line. It can be incorrect format of the transcription file, wrong windows-
    style line ending or issue with the transcription of the dictionary words from
    that line.

    To quickly solve this problem you can try to remove this line from fileids and
    transcription and proceed further.

    Proper way to fix this issue would be to update to the latest Sphinxtrain code
    and try to reproduce an issue, collect stack trace and fix the bug. If you'll
    share the whole training folder I can try to reproduce this issue myself.

     

  • numb skull

    numb skull - 2010-09-28

    yes it does stop there. ive checked the transcript and dictionary, the format
    seems to be correct since i have successfully trained ci and cd models. ill
    try to do what you suggested. . . thanks so much!

    oh by the way im now using ubuntu, and it seems that that error keeps popping
    when i use it. i asked my partner to run it using cygwin, and it proceeded to
    the succeeding modules. are there issues on sphinx regarding the use of ubuntu
    as a platform? and i sometimes get errors during sphinx3 installlation.

    ill try do that as well. . . ok, ill be back in a while to upload the training
    folder. it may take a while, my internet connection here is really slow. . . i
    appreciate your help! thanks!

     

  • Nickolay V. Shmyrev

    Nickolay V. Shmyrev - 2010-09-29

    Hm, I'm unable to reproduce this issue with your files. Everything works here.
    Can you try the following:

    1. Which sphinxtrain version are you using

    2. Install valgrind. Run this command from terminal from the training folder:

      valgrind /home/dsp/cc/bin/bw \
      -moddeffn /home/dsp/cc/model_architecture/cc.ci.mdef \
      -ts2cbfn .cont. \
      -mixwfn /home/dsp/cc/model_parameters/cc.ci_lda_flatinitial/mixture_weights \
      -mwfloor 1e-05 \
      -tpfloor 1e-05 \
      -tmatfn /home/dsp/cc/model_parameters/cc.ci_lda_flatinitial/transition_matrices \
      -meanfn /home/dsp/cc/model_parameters/cc.ci_lda_flatinitial/means \
      -varfn /home/dsp/cc/model_parameters/cc.ci_lda_flatinitial/variances \
      -ltsoov no \
      -dictfn /home/dsp/cc/etc/cc.dic \
      -fdictfn /home/dsp/cc/etc/cc.filler \
      -ctlfn /home/dsp/cc/etc/cc_train.fileids \
      -part 1 \
      -npart 1 \
      -cepdir /home/dsp/cc/feat \
      -cepext mfc \
      -lsnfn /home/dsp/cc/etc/cc_train.transcription \
      -accumdir /home/dsp/cc/bwaccumdir/cc_buff_1 \
      -varfloor 0.0001 \
      -topn 8 \
      -abeam 1e-90 \
      -bbeam 1e-10 \
      -agc none \
      -cmn current \
      -varnorm no \
      -meanreest yes \
      -varreest yes -2passvar no \
      -fullvar no \
      -diagfull no \
      -feat 1s_c_d_dd \
      -ceplen 13 \
      -cepwin 0 \
      -timing no

    Report back valgrind log please.

     

  • numb skull

    numb skull - 2010-09-29

    1. SphinxTrain.nightly from robust group's tutorial

    2. I get the an error:

    ==23290== Memcheck, a memory error detector.
    ==23290== Copyright (C) 2002-2008, and GNU GPL'd, by Julian Seward et al.
    ==23290== Using LibVEX rev 1884, a library for dynamic binary translation.
    ==23290== Copyright (C) 2004-2008, and GNU GPL'd, by OpenWorks LLP.
    ==23290== Using valgrind-3.4.1-Debian, a dynamic binary instrumentation
    framework.
    ==23290== Copyright (C) 2000-2008, and GNU GPL'd, by Julian Seward et al.
    ==23290== For more details, rerun with: -v
    ==23290==
    INFO: main.c(196): Compiled on Sep 28 2010 at 12:31:17
    /home/dsp/cc/bin/bw \
    -moddeffn /home/dsp/cc/model_architecture/cc.ci.mdef \
    -ts2cbfn .cont. \
    -mixwfn /home/dsp/cc/model_parameters/cc.ci_lda_flatinitial/mixture_weights \
    -mwfloor 1e-05 \
    -tpfloor 1e-05 \
    -tmatfn /home/dsp/cc/model_parameters/cc.ci_lda_flatinitial/transition_matrices

    -help no no
    -example no no
    -hmmdir
    -moddeffn /home/dsp/cc/model_architecture/cc.ci.mdef
    -tmatfn /home/dsp/cc/model_parameters/cc.ci_lda_flatinitial/transition_matrices
    -mixwfn /home/dsp/cc/model_parameters/cc.ci_lda_flatinitial/mixture_weights
    -meanfn
    -varfn
    -fullvar no no
    -diagfull no no
    -mwfloor 0.00001 1.000000e-05
    -tpfloor 0.0001 1.000000e-05
    -varfloor 0.00001 1.000000e-05
    -topn 4 4
    -dictfn
    -fdictfn
    -ltsoov no no
    -ctlfn
    -nskip
    -runlen -1 -1
    -part
    -npart
    -cepext mfc mfc
    -cepdir
    -phsegext phseg phseg
    -phsegdir
    -outphsegdir
    -sentdir
    -sentext sent sent
    -lsnfn
    -accumdir
    -ceplen 13 13
    -cepwin 0 0
    -agc max max
    -cmn current current
    -varnorm no no
    -silcomp none none
    -sildel no no
    -siltag SIL SIL
    -abeam 1e-100 1.000000e-100
    -bbeam 1e-100 1.000000e-100
    -varreest yes yes
    -meanreest yes yes
    -mixwreest yes yes
    -tmatreest yes yes
    -mllrmat
    -cb2mllrfn .1cls. .1cls.
    -ts2cbfn .cont.
    -feat 1s_c_d_dd 1s_c_d_dd
    -svspec
    -ldafn
    -ldadim 29 29
    -ldaaccum no no
    -timing yes yes
    -viterbi no no
    -2passvar no no
    -sildelfn
    -spthresh 0.0 0.000000e+00
    -maxuttlen 0 0
    -ckptintv
    -outputfullpath no no
    -fullsuffixmatch no no
    -pdumpdir
    INFO: main.c(255): Reading /home/dsp/cc/model_architecture/cc.ci.mdef
    INFO: model_def_io.c(587): Model definition info:
    INFO: model_def_io.c(588): 44 total models defined (44 base, 0 tri)
    INFO: model_def_io.c(589): 176 total states
    INFO: model_def_io.c(590): 132 total tied states
    INFO: model_def_io.c(591): 132 total tied CI states
    INFO: model_def_io.c(592): 44 total tied transition matrices
    INFO: model_def_io.c(593): 4 max state/model
    INFO: model_def_io.c(594): 4 min state/model
    INFO: s3mixw_io.c(116): Read
    /home/dsp/cc/model_parameters/cc.ci_lda_flatinitial/mixture_weights
    INFO: s3tmat_io.c(115): Read
    /home/dsp/cc/model_parameters/cc.ci_lda_flatinitial/transition_matrices
    INFO: mod_inv.c(297): inserting tprob floor 1.000000e-05 and renormalizing
    ==23290== Syscall param open(filename) points to unaddressable byte(s)
    ==23290== at 0x40007F2: (within /lib/ld-2.9.so)
    ==23290== by 0x40BFA3C: (within /lib/tls/i686/cmov/libc-2.9.so)
    ==23290== by 0x40BFA9B: fopen (in /lib/tls/i686/cmov/libc-2.9.so)
    ==23290== by 0x80648E8: s3open (s3io.c:254)
    ==23290== by 0x8061A80: s3gau_read (s3gau_io.c:76)
    ==23290== by 0x8070D46: mod_inv_read_gauden (mod_inv.c:164)
    ==23290== by 0x804F13A: main_initialize (main.c:303)
    ==23290== by 0x8051549: main (main.c:1052)
    ==23290== Address 0x0 is not stack'd, malloc'd or (recently) free'd
    WARN: "s3io.c", line 256: Unable to open (null) for reading; Bad address
    FATAL_ERROR: "main.c", line 1054: initialization failed
    ==23290==
    ==23290== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 13 from 1)
    ==23290== malloc/free: in use at exit: 8,470 bytes in 220 blocks.
    ==23290== malloc/free: 275 allocs, 55 frees, 10,432 bytes allocated.
    ==23290== For counts of detected errors, rerun with: -v
    ==23290== searching for pointers to 220 not-freed blocks.
    ==23290== checked 199,892 bytes.
    ==23290==
    ==23290== LEAK SUMMARY:
    ==23290== definitely lost: 175 bytes in 43 blocks.
    ==23290== possibly lost: 0 bytes in 0 blocks.
    ==23290== still reachable: 8,295 bytes in 177 blocks.
    ==23290== suppressed: 0 bytes in 0 blocks.
    ==23290== Rerun with --leak-check=full to see details of leaked memory.
    ./test.sh: line 9: -meanfn: command not found

     

  • Nickolay V. Shmyrev

    Nickolay V. Shmyrev - 2010-09-29

    ./test.sh: line 9: -meanfn: command not found

    Your forgot one backslash in the command I pasted you. Just before -meanfn.
    Each line should end with backslash.

     

  • numb skull

    numb skull - 2010-09-29

    i sent you a message containing the log file. sorry im working in the lab,
    hosting sites are forbidden here.

     

  • Nickolay V. Shmyrev

    Nickolay V. Shmyrev - 2010-09-29

    I haven't got anything yet.

     
1 2 > >> (Page 1 of 2)

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.