Help with SphinxTrain

[z x]

Hi,

I tried to build a acoustic model using SphinxTrain and got all the data
prepared. I executed RunAll.pl and got the message:

MODULE: 00 verify training files
O.S. is case sensitive ("A" != "a").
Phones will be treated as case sensitive.
Phase 1: DICT - Checking to see if the dict and filler dict agrees with the
phonelist file.
Found 58426 words using 185 phones
Phase 2: DICT - Checking to make sure there are not duplicate entries in the
dictionary
Phase 3: CTL - Check general format; utterance length (must be positive);
files exist
Phase 4: CTL - Checking number of lines in the transcript should match lines
in control file
Phase 5: CTL - Determine amount of training data, see if n_tied_states seems
reasonable.
Estimated Total Hours Training: 55.1176555555556
Rule of thumb suggests 3000, however there is no correct answer
Phase 6: TRANSCRIPT - Checking that all the words in the transcript are in the
dictionary
Words in dictionary: 58423
Words in filler dictionary: 3
Phase 7: TRANSCRIPT - Checking that all the phones in the transcript are in
the phonelist, and all phones in the phonelist appear at least once
MODULE: 01 Train LDA transformation
Skipped (set $CFG_LDA_MLLT = 'yes' to enable)
MODULE: 02 Train MLLT transformation
Skipped (set $CFG_LDA_MLLT = 'yes' to enable)
MODULE: 05 Vector Quantization
Skipped for continuous models
MODULE: 10 Training Context Independent models for forced alignment and VTLN
Skipped: $ST::CFG_FORCEDALIGN set to 'no' in sphinx_train.cfg
Skipped: $ST::CFG_VTLN set to 'no' in sphinx_train.cfg
MODULE: 11 Force-aligning transcripts
Skipped: $ST::CFG_FORCEDALIGN set to 'no' in sphinx_train.cfg
MODULE: 12 Force-aligning data for VTLN
Skipped: $ST::CFG_VTLN set to 'no' in sphinx_train.cfg
MODULE: 20 Training Context Independent models
Phase 1: Cleaning up directories:
accumulator...logs...qmanager...models...
Phase 2: Flat initialize
This step had 3 ERROR messages and 0 WARNING messages. Please check the log
file for details.
This step had 1 ERROR messages and 1 WARNING messages. Please check the log
file for details.
This step had 3 ERROR messages and 0 WARNING messages. Please check the log
file for details.
This step had 1 ERROR messages and 1 WARNING messages. Please check the log
file for details.
Phase 3: Forward-Backward
Baum welch starting for 1 Gaussian(s), iteration: 1 (1 of 1)
0%
This step had 1 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
Something failed: (/root/mytraining/scripts_pl/20.ci_hmm/slave_convg.pl)

I checked the logs and found this one most suspicious:

INFO: cmd_ln.c(557): Parsing command line:
/root/mytraining/bin/init_gau \
-ctlfn /root/mytraining/etc/mytraining_train.fileids \
-part 1 \
-npart 1 \
-cepdir /root/mytraining/feat \
-cepext mfc \
-accumdir /root/mytraining/bwaccumdir/mytraining_buff_1 \
-agc none \
-cmn current \
-varnorm no \
-feat 1s_c_d_dd \
-ceplen 13 \
-cepwin 0

ERROR: "cmd_ln.c", line 612: Unknown argument name '-cepwin'
ERROR: "cmd_ln.c", line 703: cmd_ln_parse_r failed
ERROR: "cmd_ln.c", line 752: cmd_ln_parse failed, forced exit
Sat Apr 2 02:11:20 2011

Could you help me figure out the problem here? Thank you.

[Nickolay V. Shmyrev]

This is a bug that was just fixed in trunk. Please update.

[z x]

Thanks for your reply, but I'm afraid the bug is still there.

[Nickolay V. Shmyrev]

Maybe you forgot to update the binary in model training folder.

[z x]

Yes, you are right. Thank you very much.

[z x]

Hi Nickolay, I got some other error. Please help.

MODULE: 50 Training Context dependent models
Phase 1: Cleaning up directories:
accumulator...logs...qmanager...
Phase 2: Copy CI to CD initialize
Phase 3: Forward-Backward
Baum welch starting for 1 Gaussian(s), iteration: 1 (1 of 1)
0%
This step had 1 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

The log says:

INFO: main.c(202): Compiled on Apr 4 2011 at 20:12:52
INFO: cmd_ln.c(557): Parsing command line:
/root/mytraining/bin/bw \
-moddeffn /root/mytraining/model_architecture/mytraining.1000.mdef \
-ts2cbfn .cont. \
-mixwfn /root/mytraining/model_parameters/mytraining.cd_cont_initial/mixture_weights \
-mwfloor 1e-05 \
-tpfloor 1e-05 \
-tmatfn /root/mytraining/model_parameters/mytraining.cd_cont_initial/transition_matrices \
-meanfn /root/mytraining/model_parameters/mytraining.cd_cont_initial/means \
-varfn /root/mytraining/model_parameters/mytraining.cd_cont_initial/variances \
-ltsoov no \
-dictfn /root/mytraining/etc/mytraining.dic \
-fdictfn /root/mytraining/etc/mytraining.filler \
-ctlfn /root/mytraining/etc/mytraining_train.fileids \
-part 1 \
-npart 1 \
-cepdir /root/mytraining/feat \
-cepext mfc \
-lsnfn /root/mytraining/etc/mytraining_train.transcription \
-accumdir /root/mytraining/bwaccumdir/mytraining_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 \
-tmatreest yes \
-fullvar no \
-diagfull no \
-feat 1s_c_d_dd \
-ceplen 13 \
-cepwin 0 \
-timing no

Current configuration:

-2passvar no no
-abeam 1e-100 1.000000e-90
-accumdir /root/mytraining/bwaccumdir/mytraining_buff_1
-agc none none
-agcthresh 2.0 2.000000e+00
-bbeam 1e-100 1.000000e-10
-cb2mllrfn .1cls. .1cls.
-cepdir /root/mytraining/feat
-cepext .mfc mfc
-ceplen 13 13
-cepwin 0 0
-ckptintv 0
-cmn current current
-cmninit 8.0 8.0
-ctlfn /root/mytraining/etc/mytraining_train.fileids
-diagfull no no
-dictfn /root/mytraining/etc/mytraining.dic
-example no no
-fdictfn /root/mytraining/etc/mytraining.filler
-feat 1s_c_d_dd 1s_c_d_dd
-fullsuffixmatch no no
-fullvar no no
-help no no
-hmmdir
-latdir
-latext
-lda
-ldaaccum no no
-ldadim 0 0
-lsnfn /root/mytraining/etc/mytraining_train.transcription
-ltsoov no no
-lw 11.5 1.150000e+01
-maxuttlen 0 0
-meanfn /root/mytraining/model_parameters/mytraining.cd_cont_initial/means
-meanreest yes yes
-mixwfn /root/mytraining/model_parameters/mytraining.cd_cont_initial/mixture_weights
-mixwreest yes yes
-mllrmat
-mmie no no
-mmie_type rand rand
-moddeffn /root/mytraining/model_architecture/mytraining.1000.mdef
-mwfloor 0.00001 1.000000e-05
-npart 1
-nskip 0
-outphsegdir
-outputfullpath no no
-part 1
-pdumpdir
-phsegdir
-phsegext phseg phseg
-runlen -1 -1
-sentdir
-sentext sent sent
-silcomp none none
-sildel no no
-sildelfn
-siltag SIL SIL
-spthresh 0.0 0.000000e+00
-svspec
-timing yes no
-tmatfn /root/mytraining/model_parameters/mytraining.cd_cont_initial/transition_matrices
-tmatreest yes yes
-topn 4 8
-tpfloor 0.0001 1.000000e-05
-ts2cbfn .cont.
-varfloor 0.00001 1.000000e-04
-varfn /root/mytraining/model_parameters/mytraining.cd_cont_initial/variances
-varnorm no no
-varreest yes yes
-viterbi no no

INFO: main.c(261): Reading
/root/mytraining/model_architecture/mytraining.1000.mdef
WARN: "model_def_io.c", line 436: Unable to open
/root/mytraining/model_architecture/mytraining.1000.mdef for reading; No such
file or directory
FATAL_ERROR: "main.c", line 1942: initialization failed
Wed Apr 6 06:03:01 2011