model adaptation

[Muneer]

$ /home/hiyassat/tutorial/long-audio-training/an4/bin/bw -hmmdir hub4wsj_sc_8k
-moddeffn hub4wsj_sc_8k/mdef.txt -ts2cbfn .semi. -feat 1s_c_d_dd -svspec
0-12/13-25/26-38 -cmn current -agc none -dictfn an4.dic
INFO: main.c(194): Compiled on Nov 1 2011 at 15:57:25
INFO: cmd_ln.c(691): Parsing command line:
/home/hiyassat/tutorial/long-audio-training/an4/bin/bw \
-hmmdir hub4wsj_sc_8k \
-moddeffn hub4wsj_sc_8k/mdef.txt \
-ts2cbfn .semi. \
-feat 1s_c_d_dd \
-svspec 0-12/13-25/26-38 \
-cmn current \
-agc none \
-dictfn an4.dic

Current configuration:

-2passvar no no
-abeam 1e-100 1.000000e-100
-accumdir
-agc none none
-agcthresh 2.0 2.000000e+00
-bbeam 1e-100 1.000000e-100
-cb2mllrfn .1cls. .1cls.
-cepdir
-cepext mfc mfc
-ceplen 13 13
-ckptintv 0
-cmn current current
-cmninit 8.0 8.0
-ctlfn
-diagfull no no
-dictfn an4.dic
-example no no
-fdictfn
-feat 1s_c_d_dd 1s_c_d_dd
-fullsuffixmatch no no
-fullvar no no
-help no no
-hmmdir hub4wsj_sc_8k
-latdir
-latext
-lda
-ldaaccum no no
-ldadim 0 0
-lsnfn
-ltsoov no no
-lw 11.5 1.150000e+01
-maxuttlen 0 0
-meanfn
-meanreest yes yes
-mixwfn
-mixwreest yes yes
-mllrmat
-mmie no no
-mmie_type rand rand
-moddeffn hub4wsj_sc_8k/mdef.txt
-mwfloor 0.00001 1.000000e-05
-npart 0
-nskip 0
-outphsegdir
-outputfullpath no no
-part 0
-pdumpdir
-phsegdir
-phsegext phseg phseg
-runlen -1 -1
-sentdir
-sentext sent sent
-spthresh 0.0 0.000000e+00
-svspec 0-12/13-25/26-38
-timing yes yes
-tmatfn
-tmatreest yes yes
-topn 4 4
-tpfloor 0.0001 1.000000e-04
-ts2cbfn .semi.
-varfloor 0.00001 1.000000e-05
-varfn
-varnorm no no
-varreest yes yes
-viterbi no no

INFO: feat.c(684): Initializing feature stream to type: '1s_c_d_dd',
ceplen=13, CMN='current', VARNORM='no', AGC='none'
INFO: cmn.c(142): mean= 12.00, mean= 0.0
INFO: main.c(219): Using subvector specification 0-12/13-25/26-38
INFO: main.c(283): Reading hub4wsj_sc_8k/mdef.txt
INFO: model_def_io.c(577): Model definition info:
INFO: model_def_io.c(578): 13997 total models defined (48 base, 13949 tri)
INFO: model_def_io.c(579): 55988 total states
INFO: model_def_io.c(580): 1144 total tied states
INFO: model_def_io.c(581): 144 total tied CI states
INFO: model_def_io.c(582): 48 total tied transition matrices
INFO: model_def_io.c(583): 4 max state/model
INFO: model_def_io.c(584): 4 min state/model
INFO: s3mixw_io.c(116): Read hub4wsj_sc_8k/mixture_weights
FATAL_ERROR: "mod_inv.c", line 357: Number of feature streams in
mixture_weights file 1 differs from the configured value 3, check the command
line options

[Muneer]

hi all,
i need to solve this error
thanks for help me .

[Nickolay V. Shmyrev]

What model are you trying to adapt?

[Muneer]

acoustic model

[Nickolay V. Shmyrev]

I understand that. I'm asking what exact model is it. Is it a hub4wsj_sc_8k
from the pocketsphinx distribution or some other model you created yourself.
Or you downloaded this model somewhere.

[Muneer]

yes its a hub4wsj_sc_8k from the pocketsphinx

[Nickolay V. Shmyrev]

Where did you take mixture_weights for this model? It doesn't seem compatible

What sphinxtrain are you using?

[Muneer]

im using long audio sphinx train and the from my acoustic model

[Nickolay V. Shmyrev]

and the from my acoustic model

Sorry, I have hard times to understand you. You said you adapting hub4wsj. Now
you are saying about your acoustic model. Again, are you using hub4wsj or your
own model? Can you clarify?

[Muneer]

thanks nshmyrev i solve this error but a new error generate its
ERROR: "s3acc_io.c", line 339: Unable to access ./gauden_counts

[Nickolay V. Shmyrev]

I see that gauden_counts file was not created on previous step.

Please see previous step messages for details.

Please provide full logs.

Long audio training branch is highly experimental. I suggest you to use
released version for your experiments.

[Muneer]

this is an error with segmentation fault in the previous step

./mllr_solve -meanfn /home//hiyassat/tutorial/long-audio-
training/an4/shreem/hub4wsj_sc_8k/means -varfn /home//hiyassat/tutorial/long-
audio-training/an4/shreem/hub4wsj_sc_8k/variances -outmllrfn mllr_matrix
-accumdir .
INFO: cmd_ln.c(691): Parsing command line:
./mllr_solve \
-meanfn /home//hiyassat/tutorial/long-audio-training/an4/shreem/hub4wsj_sc_8k/means \
-varfn /home//hiyassat/tutorial/long-audio-training/an4/shreem/hub4wsj_sc_8k/variances \
-outmllrfn mllr_matrix \
-accumdir .

Current configuration:

-accumdir .,
-cb2mllrfn .1cls. .1cls.
-cdonly no no
-example no no
-fullvar no no
-help no no
-meanfn /home//hiyassat/tutorial/long-audio-training/an4/shreem/hub4wsj_sc_8k/means
-mllradd yes yes
-mllrmult yes yes
-moddeffn
-outmllrfn mllr_matrix
-varfloor 1e-3 1.000000e-03
-varfn /home//hiyassat/tutorial/long-audio-training/an4/shreem/hub4wsj_sc_8k/variances

INFO: main.c(387): -- 1. Read input mean, (var) and accumulation.
INFO: s3gau_io.c(166): Read /home//hiyassat/tutorial/long-audio-
training/an4/shreem/hub4wsj_sc_8k/means
INFO: main.c(402): Reading and accumulating counts from .
ERROR: "s3acc_io.c", line 339: Unable to access ./gauden_counts
FATAL_ERROR: "main.c", line 412: Error in reading .

[Nickolay V. Shmyrev]

this is an error with segmentation fault in the previous step

I do not see segmentation fault error here. Are you sure you provided the
correct log?