Kaldi / Discussion / Open Discussion: problem when running kaldi with TIMIT database

Mao-Chang - 2015-01-14

Hello,

I got some problem running timit s5. It always stops in the process of tri2:LDA+MLLT .
It goes like this:
============================================================================
tri2 : LDA + MLLT Training & Decoding
============================================================================
steps/align_si.sh --nj 16 --cmd run.pl data/train data/lang exp/tri1 exp/tri1_ali
steps/align_si.sh: feature type is delta
steps/align_si.sh: aligning data in data/train using model from exp/tri1, putting alignments in exp/tri1_ali
steps/align_si.sh: done aligning data.
steps/train_lda_mllt.sh --cmd run.pl --splice-opts --left-context=3 --right-context=3 2500 15000 data/train data/lang exp/tri1_ali exp/tri2
Accumulating LDA statistics.
root@lab711:/home/jiaming/kaldi-trunk/egs/timit/s5#

I don't know why it stops here.

Thank you!

Moa-Chang

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- creatorscan - 2015-01-14
  
  Hi Mao-Chang,
  
  Can you paste your output log in "tri2/log/acc.1.1.log"
  
  On Wed, Jan 14, 2015 at 4:12 PM, Mao-Chang kevin79577@users.sf.net wrote:
  
  Hello,
  
  I got some problem running timit s5. It always stops in the process of
  tri2:LDA+MLLT .
  It goes like this:
  
  ============================================================================
  tri2 : LDA + MLLT Training & Decoding
  
  ============================================================================
  steps/align_si.sh --nj 16 --cmd run.pl data/train data/lang exp/tri1
  exp/tri1_ali
  steps/align_si.sh: feature type is delta
  steps/align_si.sh: aligning data in data/train using model from exp/tri1,
  putting alignments in exp/tri1_ali
  steps/align_si.sh: done aligning data.
  steps/train_lda_mllt.sh --cmd run.pl --splice-opts --left-context=3
  --right-context=3 2500 15000 data/train data/lang exp/tri1_ali exp/tri2
  Accumulating LDA statistics.
  root@lab711:/home/jiaming/kaldi-trunk/egs/timit/s5#
  
  I don't know why it stops here.
  
  Thank you!
  
  Moa-Chang
  
  problem when running kaldi with TIMIT database
  https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29
  
  Sent from sourceforge.net because you indicated interest in
  https://sourceforge.net/p/kaldi/discussion/1355347/
  
  To unsubscribe from further messages, please visit
  https://sourceforge.net/auth/subscriptions/
  
  alternate
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
  - Mao-Chang - 2015-01-15
    
    Hear is the output log file.
    
    ali-to-post "ark:gunzip -c exp/tri1_ali/ali.11.gz|" ark:- | weight-silence-post 0.0 1 exp/tri1_ali/final.mdl ark:- ark:- | acc-lda --rand-prune=4.0 exp/tri1_ali/final.mdl "ark,s,cs:apply-cmvn --utt2spk=ark:data/train/split16/11/utt2spk scp:data/train/split16/11/cmvn.scp scp:data/train/split16/11/feats.scp ark:- | splice-feats --left-context=3 --right-context=3 ark:- ark:- |" ark,s,cs:- exp/tri2/lda.11.acc
    
    Started at Mon Jan 12 17:38:21 CST 2015
    
    weight-silence-post 0.0 1 exp/tri1_ali/final.mdl ark:- ark:-
    acc-lda --rand-prune=4.0 exp/tri1_ali/final.mdl 'ark,s,cs:apply-cmvn --utt2spk=ark:data/train/split16/11/utt2spk scp:data/train/split16/11/cmvn.scp scp:data/train/split16/11/feats.scp ark:- | splice-feats --left-context=3 --right-context=3 ark:- ark:- |' ark,s,cs:- exp/tri2/lda.11.acc
    ali-to-post 'ark:gunzip -c exp/tri1_ali/ali.11.gz|' ark:-
    splice-feats --left-context=3 --right-context=3 ark:- ark:-
    apply-cmvn --utt2spk=ark:data/train/split16/11/utt2spk scp:data/train/split16/11/cmvn.scp scp:data/train/split16/11/feats.scp ark:-
    LOG (acc-lda:main():acc-lda.cc:113) Done 100 utterances.
    LOG (acc-lda:main():acc-lda.cc:113) Done 200 utterances.
    LOG (ali-to-post:main():ali-to-post.cc:65) Converted 232 alignments.
    LOG (weight-silence-post:main():weight-silence-post.cc:95) Done 232 posteriors.
    LOG (apply-cmvn:main():apply-cmvn.cc:132) Applied cepstral mean normalization to 232 utterances, errors on 0
    LOG (acc-lda:main():acc-lda.cc:116) Done 232 files, failed for 0
    LOG (acc-lda:main():acc-lda.cc:121) Written statistics.
    
    Accounting: time=1 threads=1
    
    Ended (code 0) at Mon Jan 12 17:38:22 CST 2015, elapsed time 1 seconds
    
    Thank you very much!
    
    Mao-Chang
    
    If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
    - Mao-Chang - 2015-01-15
      
      By the way, I paste the output in "s5/log/lda_est.log" file:
      
      est-lda: symbol lookup error: est-lda: undefined symbol: clapack_dgetrf
      
      Maybe there is the problem here?
      
      Mao-Chang
      
      If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
      - Daniel Povey - 2015-01-15
        
        Yes that must be the problem.
        in src/, do "make test". probably it will fail. means it did not install
        correctly.
        Did you just do "configure" and make? You could show us the output of
        running "configure", and also the result of "ldd matrix/matrix-lib-test"
        
        Dan
        
        On Wed, Jan 14, 2015 at 8:39 PM, Mao-Chang kevin79577@users.sf.net wrote:
        
        By the way, I paste the output in "s5/log/lda_est.log" file:
        
        est-lda: symbol lookup error: est-lda: undefined symbol: clapack_dgetrf
        
        Maybe there is the problem here?
        
        Mao-Chang
        
        problem when running kaldi with TIMIT database
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Daniel Povey - 2015-01-15
        
        Hi,
        I realized that there was an old thread with the same issue:
        
        http://sourceforge.net/p/kaldi/discussion/1355348/thread/5b275ab6/
        
        although in the end, that user was not able to replicate the problem when
        re-running the installation, so we could not fix the issue at that time.
        Again, please run "configure", show me the output,
        show me the output of
        nm --dynamic /usr/lib/liblapack.so.3 | grep cgetrf
        and in src/, type
        ldd ./matrix/matrix-lib-test
        and show us the output.
        Thanks,
        Dan
        
        On Thu, Jan 15, 2015 at 12:01 AM, Daniel Povey danielpovey@users.sf.net
        wrote:
        
        Yes that must be the problem.
        in src/, do "make test". probably it will fail. means it did not install
        correctly.
        Did you just do "configure" and make? You could show us the output of
        running "configure", and also the result of "ldd matrix/matrix-lib-test"
        
        Dan
        
        On Wed, Jan 14, 2015 at 8:39 PM, Mao-Chang kevin79577@users.sf.net wrote:
        
        By the way, I paste the output in "s5/log/lda_est.log" file:
        
        est-lda: symbol lookup error: est-lda: undefined symbol: clapack_dgetrf
        
        Maybe there is the problem here?
        Mao-Chang
        
        problem when running kaldi with TIMIT database
        
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        problem when running kaldi with TIMIT database
        http://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        alternate
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Mao-Chang - 2015-01-15
        
        The result of "make test":
        Running matrix-lib-test ...... FAIL matrix-lib-test
        Running kaldi-gpsr-test ...... SUCCESS
        make[1]: [test] Error 1
        make[1]: Leaving directory `/home/jiaming/kaldi-trunk/src/matrix'
        make: [matrix/test] Error 2
        
        ====================================
        
        The result of "ldd matrix/matrix-lib-test":
        root@lab711:/home/jiaming/kaldi-trunk/src# ldd matrix/matrix-lib-test
        linux-vdso.so.1 => (0x00007fff2c3ff000)
        libfst.so.1 => /home/jiaming/kaldi-trunk/tools/openfst/lib/libfst.so.1 (0x00007f035d921000)
        liblapack.so.3 => /usr/lib/liblapack.so.3 (0x00007f035cd37000)
        libcblas.so.3 => /usr/lib/libcblas.so.3 (0x00007f035cb19000)
        libatlas.so.3 => /usr/lib/libatlas.so.3 (0x00007f035c51e000)
        libf77blas.so.3 => /usr/lib/libf77blas.so.3 (0x00007f035c301000)
        libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f035c0e4000)
        libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f035bee0000)
        libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f035bbd9000)
        libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f035b956000)
        libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f035b740000)
        libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f035b3b4000)
        libblas.so.3 => /usr/lib/libblas.so.3 (0x00007f035ad84000)
        libgfortran.so.3 => /usr/lib/x86_64-linux-gnu/libgfortran.so.3 (0x00007f035aa6e000)
        libquadmath.so.0 => /usr/lib/x86_64-linux-gnu/libquadmath.so.0 (0x00007f035a839000)
        /lib64/ld-linux-x86-64.so.2 (0x00007f035dd36000)
        
        =====================================
        
        The result of
        "nm --dynamic /usr/lib/liblapack.so.3 | grep cgetrf":
        000000000007c860 T cgetrf_
        
        =====================================
        
        The result of "./configure":
        
        Configuring ...
        Checking OpenFST library in /home/jiaming/kaldi-trunk/tools/openfst ...
        Checking OpenFst library was patched.
        Backing up kaldi.mk to kaldi.mk.bak
        Doing OS specific configurations ...
        On Linux: Checking for linear algebra header files ...
        Using ATLAS as the linear algebra library.
        Atlas found in /usr/lib
        ... no libatlas.so in /usr/lib/atlas
        ... no libatlas.so in /usr/lib/atlas-sse2
        ... no libatlas.so in /usr/lib/atlas-sse3
        ... no libatlas.so in /usr/lib64
        ... no libatlas.so in /usr/lib64/atlas
        ... no libatlas.so in /usr/lib64/atlas-sse2
        ... no libatlas.so in /usr/lib64/atlas-sse3
        ... no libatlas.so in /usr/local/lib
        ... no libatlas.so in /usr/local/lib/atlas
        ... no libatlas.so in /usr/local/lib/atlas-sse2
        ... no libatlas.so in /usr/local/lib/atlas-sse3
        ... no libatlas.so in /usr/local/lib64
        ... no libatlas.so in /usr/local/lib64/atlas
        ... no libatlas.so in /usr/local/lib64/atlas-sse2
        ... no libatlas.so in /usr/local/lib64/atlas-sse3
        ... no libatlas.so in /home/jiaming/kaldi-trunk/src/../tools/ATLAS/build/install/lib/
        ... no libatlas.so in /home/jiaming/kaldi-trunk/tools/ATLAS/lib
        Looking for ATLAS libs in /usr/lib
        Using library -L/usr/lib -llapack as ATLAS's CLAPACK library.
        CUDA will not be used! If you have already installed cuda drivers and cuda toolkit, try using --cudatk-dir=... option. Note: this is only relevant for neural net experiments
        Static=[false] Speex library not found: You can still build Kaldi without Speex.
        Successfully configured for Linux [dynamic libraries] with ATLASLIBS =-L/usr/lib -llapack -lcblas -latlas -lf77blas
        exp() time: 0.00278187
        expf() time: 0.023715
        WARNING: expf() seems to be slower than exp() on your machine. This is a known bug in old versions of glibc. Please consider updating glibc.
        Kaldi will be configured to use exp() instead of expf() in base/kaldi-math.h Exp() routine for single-precision floats.
        
        These are the output of those command.
        
        Thanks!
        
        Mao-Chang
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Daniel Povey - 2015-01-15
        
        OK,
        That's odd because the configure script checks that that library contains
        the symbol ATL_cgetrf but the output of your grep command seems to show
        the opposite.
        
        Could you please do the following commands:
        
        nm --dynamic /usr/lib/liblapack.so | grep cgetrf
        nm --dynamic /usr/lib/liblapack.so.3 | grep cgetrf
        ls -l /usr/lib/liblapack.so
        
        and show me the output?
        Dan
        
        On Thu, Jan 15, 2015 at 4:23 AM, Mao-Chang kevin79577@users.sf.net wrote:
        
        The result of "make test":
        Running matrix-lib-test ...... FAIL matrix-lib-test
        Running kaldi-gpsr-test ...... SUCCESS
        make[1]:
        
        [test] Error 1 make[1]: Leaving directory
        `/home/jiaming/kaldi-trunk/src/matrix' make: * [matrix/test] Error 2
        
        ====================================
        
        The result of "ldd matrix/matrix-lib-test":
        root@lab711:/home/jiaming/kaldi-trunk/src# ldd matrix/matrix-lib-test
        linux-vdso.so.1 => (0x00007fff2c3ff000)
        libfst.so.1 => /home/jiaming/kaldi-trunk/tools/openfst/lib/libfst.so.1
        (0x00007f035d921000)
        liblapack.so.3 => /usr/lib/liblapack.so.3 (0x00007f035cd37000)
        libcblas.so.3 => /usr/lib/libcblas.so.3 (0x00007f035cb19000)
        libatlas.so.3 => /usr/lib/libatlas.so.3 (0x00007f035c51e000)
        libf77blas.so.3 => /usr/lib/libf77blas.so.3 (0x00007f035c301000)
        libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0
        (0x00007f035c0e4000)
        libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f035bee0000)
        libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6
        (0x00007f035bbd9000)
        libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f035b956000)
        libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f035b740000)
        libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f035b3b4000)
        libblas.so.3 => /usr/lib/libblas.so.3 (0x00007f035ad84000)
        libgfortran.so.3 => /usr/lib/x86_64-linux-gnu/libgfortran.so.3
        (0x00007f035aa6e000)
        libquadmath.so.0 => /usr/lib/x86_64-linux-gnu/libquadmath.so.0
        (0x00007f035a839000)
        /lib64/ld-linux-x86-64.so.2 (0x00007f035dd36000)
        
        =====================================
        
        The result of
        "nm --dynamic /usr/lib/liblapack.so.3 | grep cgetrf":
        000000000007c860 T cgetrf_
        
        =====================================
        
        The result of "./configure":
        
        Configuring ...
        Checking OpenFST library in /home/jiaming/kaldi-trunk/tools/openfst ...
        Checking OpenFst library was patched.
        Backing up kaldi.mk to kaldi.mk.bak
        Doing OS specific configurations ...
        On Linux: Checking for linear algebra header files ...
        Using ATLAS as the linear algebra library.
        Atlas found in /usr/lib
        ... no libatlas.so in /usr/lib/atlas
        ... no libatlas.so in /usr/lib/atlas-sse2
        ... no libatlas.so in /usr/lib/atlas-sse3
        ... no libatlas.so in /usr/lib64
        ... no libatlas.so in /usr/lib64/atlas
        ... no libatlas.so in /usr/lib64/atlas-sse2
        ... no libatlas.so in /usr/lib64/atlas-sse3
        ... no libatlas.so in /usr/local/lib
        ... no libatlas.so in /usr/local/lib/atlas
        ... no libatlas.so in /usr/local/lib/atlas-sse2
        ... no libatlas.so in /usr/local/lib/atlas-sse3
        ... no libatlas.so in /usr/local/lib64
        ... no libatlas.so in /usr/local/lib64/atlas
        ... no libatlas.so in /usr/local/lib64/atlas-sse2
        ... no libatlas.so in /usr/local/lib64/atlas-sse3
        ... no libatlas.so in
        /home/jiaming/kaldi-trunk/src/../tools/ATLAS/build/install/lib/
        ... no libatlas.so in /home/jiaming/kaldi-trunk/tools/ATLAS/lib
        Looking for ATLAS libs in /usr/lib
        Using library -L/usr/lib -llapack as ATLAS's CLAPACK library.
        CUDA will not be used! If you have already installed cuda drivers and cuda
        toolkit, try using --cudatk-dir=... option. Note: this is only relevant for
        neural net experiments
        Static=[false] Speex library not found: You can still build Kaldi without
        Speex.
        Successfully configured for Linux [dynamic libraries] with ATLASLIBS
        =-L/usr/lib -llapack -lcblas -latlas -lf77blas
        exp() time: 0.00278187
        expf() time: 0.023715
        WARNING: expf() seems to be slower than exp() on your machine. This is a
        known bug in old versions of glibc. Please consider updating glibc.
        * Kaldi will be configured to use exp() instead of expf() in
        base/kaldi-math.h Exp() routine for single-precision floats. *
        
        These are the output of those command.
        
        Thanks!
        
        Mao-Chang
        
        problem when running kaldi with TIMIT database
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Mao-Chang - 2015-01-18
        
        "nm --dynamic /usr/lib/liblapack.so | grep cgetrf"
        0000000000026e50 T ATL_cgetrf
        0000000000026e80 T ATL_cgetrfC
        0000000000027180 T ATL_cgetrfR
        000000000002d214 T atl_f77wrap_cgetrf_
        00000000000a4904 T cgetrf_
        000000000015e8c0 T clapack_cgetrf
        
        "nm --dynamic /usr/lib/liblapack.so.3 | grep cgetrf"
        000000000007c860 T cgetrf_
        
        "ls -l /usr/lib/liblapack.so"
        lrwxrwxrwx 1 root root 30 1月 10 20:09 /usr/lib/liblapack.so -> /etc/alternatives/liblapack.so
        
        Thanks
        
        Mao-Chang
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Daniel Povey - 2015-01-18
        
        Please update Kaldi, rerun configure, and try to compile.
        I've attempted to fix the issue.
        Dan
        
        On Sat, Jan 17, 2015 at 10:30 PM, Mao-Chang kevin79577@users.sf.net wrote:
        
        "nm --dynamic /usr/lib/liblapack.so | grep cgetrf"
        0000000000026e50 T ATL_cgetrf
        0000000000026e80 T ATL_cgetrfC
        0000000000027180 T ATL_cgetrfR
        000000000002d214 T atl_f77wrap_cgetrf_
        00000000000a4904 T cgetrf_
        000000000015e8c0 T clapack_cgetrf
        
        "nm --dynamic /usr/lib/liblapack.so.3 | grep cgetrf"
        000000000007c860 T cgetrf_
        
        "ls -l /usr/lib/liblapack.so"
        lrwxrwxrwx 1 root root 30 1月 10 20:09 /usr/lib/liblapack.so ->
        /etc/alternatives/liblapack.so
        
        Thanks
        
        Mao-Chang
        
        problem when running kaldi with TIMIT database
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Mao-Chang - 2015-01-18
        
        I enter this command line to download Kaldi.
        svn co svn://svn.code.sf.net/p/kaldi/code/trunk kaldi-trunk
        
        If I want to update kaldi, what can I do?
        
        Thanks!
        
        Mao-Chang
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Daniel Povey - 2015-01-18
        
        type "svn up"
        Dan
        
        On Sun, Jan 18, 2015 at 5:52 AM, Mao-Chang kevin79577@users.sf.net wrote:
        
        I enter this command line to download Kaldi.
        svn co svn://svn.code.sf.net/p/kaldi/code/trunk kaldi-trunk
        
        If I want to update kaldi, what can I do?
        
        Thanks!
        
        Mao-Chang
        
        problem when running kaldi with TIMIT database
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        alternate
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Mao-Chang - 2015-01-19
        
        I have typed "svn up svn://svn.code.sf.net/p/kaldi/code/trunk kaldi-trunk" to update kaldi.
        And the output of "./configure" is below.
        Configuring ...
        Checking OpenFST library in /home/jiaming/kaldi-trunk/tools/openfst ...
        Checking OpenFst library was patched.
        Backing up kaldi.mk to kaldi.mk.bak
        Doing OS specific configurations ...
        On Linux: Checking for linear algebra header files ...
        Using ATLAS as the linear algebra library.
        Atlas found in /usr/lib
        ... no libatlas.so in /usr/lib/atlas
        ... no libatlas.so in /usr/lib/atlas-sse2
        ... no libatlas.so in /usr/lib/atlas-sse3
        ... no libatlas.so in /usr/lib64
        ... no libatlas.so in /usr/lib64/atlas
        ... no libatlas.so in /usr/lib64/atlas-sse2
        ... no libatlas.so in /usr/lib64/atlas-sse3
        ... no libatlas.so in /usr/local/lib
        ... no libatlas.so in /usr/local/lib/atlas
        ... no libatlas.so in /usr/local/lib/atlas-sse2
        ... no libatlas.so in /usr/local/lib/atlas-sse3
        ... no libatlas.so in /usr/local/lib64
        ... no libatlas.so in /usr/local/lib64/atlas
        ... no libatlas.so in /usr/local/lib64/atlas-sse2
        ... no libatlas.so in /usr/local/lib64/atlas-sse3
        ... no libatlas.so in /home/jiaming/kaldi-trunk/src/../tools/ATLAS/build/install/lib/
        ... no libatlas.so in /home/jiaming/kaldi-trunk/tools/ATLAS/lib
        Validating presence of ATLAS libs in /usr/lib
        Using library /usr/lib/liblapack.so as ATLAS's CLAPACK library.
        CUDA will not be used! If you have already installed cuda drivers and cuda toolkit, try using --cudatk-dir=... option. Note: this is only relevant for neural net experiments
        Static=[false] Speex library not found: You can still build Kaldi without Speex.
        Successfully configured for Linux [dynamic libraries] with ATLASLIBS =/usr/lib/liblapack.so /usr/lib/libcblas.so /usr/lib/libatlas.so /usr/lib/libf77blas.so
        exp() time: 0.00278783
        expf() time: 0.0241911
        WARNING: expf() seems to be slower than exp() on your machine. This is a known bug in old versions of glibc. Please consider updating glibc.
        Kaldi will be configured to use exp() instead of expf() in base/kaldi-math.h Exp() routine for single-precision floats.
        root@lab711:/home/jiaming/kaldi-trunk/src#
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Daniel Povey - 2015-01-19
        
        Thanks. I think it should work now-- try to compile and run and see if it
        works; let us know if there is still a problem.
        Dan
        
        On Mon, Jan 19, 2015 at 2:47 AM, Mao-Chang kevin79577@users.sf.net wrote:
        
        I have typed "svn up svn://svn.code.sf.net/p/kaldi/code/trunk
        kaldi-trunk" to update kaldi.
        And the output of "./configure" is below.
        
        Configuring ...
        Checking OpenFST library in /home/jiaming/kaldi-trunk/tools/openfst ...
        Checking OpenFst library was patched.
        Backing up kaldi.mk to kaldi.mk.bak
        Doing OS specific configurations ...
        On Linux: Checking for linear algebra header files ...
        Using ATLAS as the linear algebra library.
        Atlas found in /usr/lib
        ... no libatlas.so in /usr/lib/atlas
        ... no libatlas.so in /usr/lib/atlas-sse2
        ... no libatlas.so in /usr/lib/atlas-sse3
        ... no libatlas.so in /usr/lib64
        ... no libatlas.so in /usr/lib64/atlas
        ... no libatlas.so in /usr/lib64/atlas-sse2
        ... no libatlas.so in /usr/lib64/atlas-sse3
        ... no libatlas.so in /usr/local/lib
        ... no libatlas.so in /usr/local/lib/atlas
        ... no libatlas.so in /usr/local/lib/atlas-sse2
        ... no libatlas.so in /usr/local/lib/atlas-sse3
        ... no libatlas.so in /usr/local/lib64
        ... no libatlas.so in /usr/local/lib64/atlas
        ... no libatlas.so in /usr/local/lib64/atlas-sse2
        ... no libatlas.so in /usr/local/lib64/atlas-sse3
        ... no libatlas.so in
        /home/jiaming/kaldi-trunk/src/../tools/ATLAS/build/install/lib/
        ... no libatlas.so in /home/jiaming/kaldi-trunk/tools/ATLAS/lib
        Validating presence of ATLAS libs in /usr/lib
        Using library /usr/lib/liblapack.so as ATLAS's CLAPACK library.
        CUDA will not be used! If you have already installed cuda drivers and cuda
        toolkit, try using --cudatk-dir=... option. Note: this is only relevant for
        neural net experiments
        Static=[false] Speex library not found: You can still build Kaldi without
        Speex.
        Successfully configured for Linux [dynamic libraries] with ATLASLIBS
        =/usr/lib/liblapack.so /usr/lib/libcblas.so /usr/lib/libatlas.so
        /usr/lib/libf77blas.so
        exp() time: 0.00278783
        expf() time: 0.0241911
        WARNING: expf() seems to be slower than exp() on your machine. This is a
        known bug in old versions of glibc. Please consider updating glibc.
        Kaldi will be configured to use exp() instead of expf() in
        base/kaldi-math.h Exp() routine for single-precision floats.
        root@lab711:/home/jiaming/kaldi-trunk/src#
        
        problem when running kaldi with TIMIT database
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579/dff3/0fed
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Mao-Chang - 2015-01-20
        
        Oh....I run TIMIT again and it stops in the same way.
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Daniel Povey - 2015-01-20
        
        You probably need to "make clean" and then "make" again.
        Dan
        
        On Tue, Jan 20, 2015 at 12:49 AM, Mao-Chang kevin79577@users.sf.net wrote:
        
        Oh....I run TIMIT again and it stops in the same way.
        
        problem when running kaldi with TIMIT database
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579/dff3/0fed/49c7/d509
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        alternate
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Mao-Chang - 2015-01-26
        
        I have done "make clean"and "make" again, but it still says no libatlas after entering "./configure". Is that any problem?
        And it stops again running TIMIT.
        
        Mao-Chang
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Daniel Povey - 2015-01-26
        
        Please show the complete output from "configure", and also ensure that you
        typed "configure" before typing "make clean" and "make".
        Dan
        
        On Mon, Jan 26, 2015 at 10:15 AM, Mao-Chang kevin79577@users.sf.net wrote:
        
        I have done "make clean"and "make" again, but it still says no libatlas
        after entering "./configure". Is that any problem?
        And it stops again running TIMIT.
        
        Mao-Chang
        
        problem when running kaldi with TIMIT database
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579/dff3/0fed/49c7/d509/a4f3/f8db
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        alternate
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Mao-Chang - 2015-01-27
        
        I'm sure that I have already typed "make clean" before "make".
        These are the output of "./configure".
        root@lab711:/home/jiaming/kaldi-trunk/src# ./configure
        Configuring ...
        Checking OpenFST library in /home/jiaming/kaldi-trunk/tools/openfst ...
        Checking OpenFst library was patched.
        Backing up kaldi.mk to kaldi.mk.bak
        Doing OS specific configurations ...
        On Linux: Checking for linear algebra header files ...
        Using ATLAS as the linear algebra library.
        Atlas found in /usr/lib
        ... no libatlas.so in /usr/lib/atlas
        ... no libatlas.so in /usr/lib/atlas-sse2
        ... no libatlas.so in /usr/lib/atlas-sse3
        ... no libatlas.so in /usr/lib64
        ... no libatlas.so in /usr/lib64/atlas
        ... no libatlas.so in /usr/lib64/atlas-sse2
        ... no libatlas.so in /usr/lib64/atlas-sse3
        ... no libatlas.so in /usr/local/lib
        ... no libatlas.so in /usr/local/lib/atlas
        ... no libatlas.so in /usr/local/lib/atlas-sse2
        ... no libatlas.so in /usr/local/lib/atlas-sse3
        ... no libatlas.so in /usr/local/lib64
        ... no libatlas.so in /usr/local/lib64/atlas
        ... no libatlas.so in /usr/local/lib64/atlas-sse2
        ... no libatlas.so in /usr/local/lib64/atlas-sse3
        ... no libatlas.so in /home/jiaming/kaldi-trunk/src/../tools/ATLAS/build/install/lib/
        ... no libatlas.so in /home/jiaming/kaldi-trunk/tools/ATLAS/lib
        Validating presence of ATLAS libs in /usr/lib
        Using library /usr/lib/liblapack.so as ATLAS's CLAPACK library.
        CUDA will not be used! If you have already installed cuda drivers and cuda toolkit, try using --cudatk-dir=... option. Note: this is only relevant for neural net experiments
        Static=[false] Speex library not found: You can still build Kaldi without Speex.
        Successfully configured for Linux [dynamic libraries] with ATLASLIBS =/usr/lib/liblapack.so /usr/lib/libcblas.so /usr/lib/libatlas.so /usr/lib/libf77blas.so
        exp() time: 0.00384402
        expf() time: 0.023818
        WARNING: expf() seems to be slower than exp() on your machine. This is a known bug in old versions of glibc. Please consider updating glibc.
        Kaldi will be configured to use exp() instead of expf() in base/kaldi-math.h Exp() routine for single-precision floats.
        root@lab711:/home/jiaming/kaldi-trunk/src#
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Daniel Povey - 2015-01-27
        
        OK, in that case please go into the matrix/ subdirectory and
        (1) show me the output of ldd matrix-lib-test
        (2) see if matrix-lib-test runs.
        Dan
        
        On Mon, Jan 26, 2015 at 9:15 PM, Mao-Chang kevin79577@users.sf.net wrote:
        
        I'm sure that I have already typed "make clean" before "make".
        These are the output of "./configure".
        root@lab711:/home/jiaming/kaldi-trunk/src# ./configure
        
        Configuring ...
        Checking OpenFST library in /home/jiaming/kaldi-trunk/tools/openfst ...
        Checking OpenFst library was patched.
        Backing up kaldi.mk to kaldi.mk.bak
        Doing OS specific configurations ...
        On Linux: Checking for linear algebra header files ...
        Using ATLAS as the linear algebra library.
        Atlas found in /usr/lib
        ... no libatlas.so in /usr/lib/atlas
        ... no libatlas.so in /usr/lib/atlas-sse2
        ... no libatlas.so in /usr/lib/atlas-sse3
        ... no libatlas.so in /usr/lib64
        ... no libatlas.so in /usr/lib64/atlas
        ... no libatlas.so in /usr/lib64/atlas-sse2
        ... no libatlas.so in /usr/lib64/atlas-sse3
        ... no libatlas.so in /usr/local/lib
        ... no libatlas.so in /usr/local/lib/atlas
        ... no libatlas.so in /usr/local/lib/atlas-sse2
        ... no libatlas.so in /usr/local/lib/atlas-sse3
        ... no libatlas.so in /usr/local/lib64
        ... no libatlas.so in /usr/local/lib64/atlas
        ... no libatlas.so in /usr/local/lib64/atlas-sse2
        ... no libatlas.so in /usr/local/lib64/atlas-sse3
        ... no libatlas.so in
        /home/jiaming/kaldi-trunk/src/../tools/ATLAS/build/install/lib/
        ... no libatlas.so in /home/jiaming/kaldi-trunk/tools/ATLAS/lib
        Validating presence of ATLAS libs in /usr/lib
        Using library /usr/lib/liblapack.so as ATLAS's CLAPACK library.
        CUDA will not be used! If you have already installed cuda drivers and cuda
        toolkit, try using --cudatk-dir=... option. Note: this is only relevant for
        neural net experiments
        Static=[false] Speex library not found: You can still build Kaldi without
        Speex.
        Successfully configured for Linux [dynamic libraries] with ATLASLIBS
        =/usr/lib/liblapack.so /usr/lib/libcblas.so /usr/lib/libatlas.so
        /usr/lib/libf77blas.so
        exp() time: 0.00384402
        expf() time: 0.023818
        WARNING: expf() seems to be slower than exp() on your machine. This is a
        known bug in old versions of glibc. Please consider updating glibc.
        Kaldi will be configured to use exp() instead of expf() in
        base/kaldi-math.h Exp() routine for single-precision floats.
        root@lab711:/home/jiaming/kaldi-trunk/src#
        
        problem when running kaldi with TIMIT database
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579/dff3/0fed/49c7/d509/a4f3/f8db/5ef2/d7e8
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        alternate
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Mao-Chang - 2015-01-27
        
        the ouput of " matrix-lib-test"
        
        root@lab711:/home/jiaming/kaldi-trunk/src/matrix# ./matrix-lib-test
        LOG (UnitTestTridiagonalize():matrix-lib-test.cc:1530) S trace 432, T trace 432
        ./matrix-lib-test: symbol lookup error: ./matrix-lib-test: undefined symbol: clapack_sgetrf
        root@lab711:/home/jiaming/kaldi-trunk/src/matrix#
        
        Mao-Chang
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Jan "yenda" Trmal - 2015-01-27
        
        CCing sedgerepawk, as he had the same issues and he managed to resolve it
        -- I'm not sure how, however.
        
        y.
        
        On Tue, Jan 27, 2015 at 8:26 AM, Mao-Chang kevin79577@users.sf.net wrote:
        
        the ouput of " matrix-lib-test"
        
        root@lab711:/home/jiaming/kaldi-trunk/src/matrix# ./matrix-lib-test
        LOG (UnitTestTridiagonalize():matrix-lib-test.cc:1530) S trace 432, T
        trace 432
        ./matrix-lib-test: symbol lookup error: ./matrix-lib-test: undefined
        symbol: clapack_sgetrf
        root@lab711:/home/jiaming/kaldi-trunk/src/matrix#
        
        Mao-Chang
        
        problem when running kaldi with TIMIT database
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579/dff3/0fed/49c7/d509/a4f3/f8db/5ef2/d7e8/a3dd/ac2d
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        alternate
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Daniel Povey - 2015-01-27
        
        Mao-Chang: I wanted to see the output of "ldd matrix-lib-test", not
        "matrix-lib-test".
        
        BTW, Yenda: I think the root cause of this problem is that in /usr/lib, he
        has two files liblapack.so and liblapack.so.3 that point to different
        libraries entirely, and I think the whole Linux linking system probably
        assumes that the .so and .so.3 with the same should point to the same
        library. This is really a problem with his installation, where he has two
        different lapack libraries that have tried to install in the same place.
        To deal with it gracefully, I was trying to fix it so that instead of
        adding "-L/usr/lib -llapack" to the command line, we'd add
        "/usr/lib/liblapack.so" to the command line, thinking that this might force
        it to use the .so version of the library. But it's possible that g++ just
        assumes the .so and .so.3 are different versions of the same thing, and
        prefers the .so.3. I could not find a good explanation of these issues
        online. If that doesn't work, we could fix this by modifying the configure
        script to do "readlink" on /usr/lib/liblapack.so and add the destination
        file to the linking line instead.
        
        Dan
        
        On Tue, Jan 27, 2015 at 4:55 AM, Jan Trmal jtrmal@users.sf.net wrote:
        
        CCing sedgerepawk, as he had the same issues and he managed to resolve it
        -- I'm not sure how, however.
        
        y.
        
        On Tue, Jan 27, 2015 at 8:26 AM, Mao-Chang kevin79577@users.sf.net wrote:
        
        the ouput of " matrix-lib-test"
        
        root@lab711:/home/jiaming/kaldi-trunk/src/matrix# ./matrix-lib-test
        LOG (UnitTestTridiagonalize():matrix-lib-test.cc:1530) S trace 432, T
        trace 432
        ./matrix-lib-test: symbol lookup error: ./matrix-lib-test: undefined
        symbol: clapack_sgetrf
        root@lab711:/home/jiaming/kaldi-trunk/src/matrix#
        Mao-Chang
        
        problem when running kaldi with TIMIT database
        
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579/dff3/0fed/49c7/d509/a4f3/f8db/5ef2/d7e8/a3dd/ac2d
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        problem when running kaldi with TIMIT database
        http://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579/dff3/0fed/49c7/d509/a4f3/f8db/5ef2/d7e8/a3dd/ac2d/e0ff
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Mao-Chang - 2015-01-28
        
        I don't have the file "ldd matrix-lib-test".
        
        Mao-Chang
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
        
        Jan "yenda" Trmal - 2015-01-28
        
        the program is just "ldd" and matrix-lib-test is another program that is
        used as a parameter. What ldd is that it tries to run the program and
        displays what dependencies are missing.
        Please send the complete command line you executed including the error you
        obtained.
        y.
        
        On Wed, Jan 28, 2015 at 4:10 PM, Mao-Chang kevin79577@users.sf.net wrote:
        
        I don't have the file "ldd matrix-lib-test".
        
        Mao-Chang
        
        problem when running kaldi with TIMIT database
        https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579/dff3/0fed/49c7/d509/a4f3/f8db/5ef2/d7e8/a3dd/ac2d/e0ff/d150/8a22
        
        Sent from sourceforge.net because you indicated interest in
        https://sourceforge.net/p/kaldi/discussion/1355347/
        
        To unsubscribe from further messages, please visit
        https://sourceforge.net/auth/subscriptions/
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

problem when running kaldi with TIMIT database

Forums

Help

problem when running kaldi with TIMIT database

Moa-Chang

Started at Mon Jan 12 17:38:21 CST 2015

Accounting: time=1 threads=1

Ended (code 0) at Mon Jan 12 17:38:22 CST 2015, elapsed time 1 seconds

Mao-Chang

https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb

Mao-Chang

Mao-Chang

Mao-Chang

Oh....I run TIMIT again and it stops in the same way.

Mao-Chang

Mao-Chang

https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579/dff3/0fed/49c7/d509/a4f3/f8db/5ef2/d7e8/a3dd/ac2d

Mao-Chang

problem when running kaldi with TIMIT database

Forums

Help

problem when running kaldi with TIMIT database document.SUBSCRIPTION_OPTIONS = { "thing": "topic", "subscribed": false, "url": "subscribe", "icon": { "css": "fa fa-envelope-o" } };

Moa-Chang

Started at Mon Jan 12 17:38:21 CST 2015

Accounting: time=1 threads=1

Ended (code 0) at Mon Jan 12 17:38:22 CST 2015, elapsed time 1 seconds

Mao-Chang

https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb

Mao-Chang

Mao-Chang

Mao-Chang

Oh....I run TIMIT again and it stops in the same way.

Mao-Chang

Mao-Chang

https://sourceforge.net/p/kaldi/discussion/1355347/thread/51787e1f/?limit=25#0f29/182a/9527/8ffb/ab96/b222/72ce/a651/1774/b1fe/2579/dff3/0fed/49c7/d509/a4f3/f8db/5ef2/d7e8/a3dd/ac2d

Mao-Chang

problem when running kaldi with TIMIT database