I am recently trying to use a classbased LM for my previously trained acoustic
model.
I have created files:
a control file :
a def fie
corresponding lm and its dictionary.
When I run the decoder, it gives me error like:
INFO: lmclass.c(181): Reading LM Class file 'atis_sphinx.def'
SYSTEM_ERROR: "lmclass.c", line 183: Failed to open LM class file; No such
file or directory
Can any one please suggest me, what should be my steps to overcome this
problem?
Am I missing anything ?
I am using turtle class lm example to format my files.
Here is my log:
INFO: info.c(65): Host: 'passidin1'
INFO: info.c(66): Directory:
'/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx'
INFO: info.c(70):
/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/bin/.libs/lt-
sphinx3_decode Compiled on: Apr 18 2012, AT: 12:35:12
-agc none none
-agcthresh 2.0 2.000000e+00
-alpha 0.97 9.700000e-01
-cep2spec no no
-ceplen 13 13
-cmn current current
-cmninit 8.0 8.0
-dither no no
-doublebw no no
-feat 1s_c_d_dd 1s_c_d_dd
-frate 100 100
-input_endian little little
-lda
-ldadim 0 0
-lifter 0 0
-logspec no no
-lowerf 133.33334 1.333333e+02
-ncep 13 13
-nfft 512 512
-nfilt 40 40
-remove_dc no yes
-round_filters yes yes
-samprate 16000 1.600000e+04
-seed -1 -1
-smoothspec no no
-spec2cep no no
-svspec
-transform legacy legacy
-unit_area yes yes
-upperf 6855.4976 6.855498e+03
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wlen 0.025625 2.000000e-02
INFO: Initialization of the log add table
INFO: Log-Add table size = 29356 x 2 >> 0
INFO:
INFO: feat.c(848): 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: kbcore.c(486): .cont.
INFO: Initialization of feat_t, report:
INFO: Feature type = 1s_c_d_dd
INFO: Cepstral size = 13
INFO: Number of streams = 1
INFO: Vector size of stream: 39
INFO: Number of subvectors = 0
INFO: Whether CMN is used = 1
INFO: Whether AGC is used = 0
INFO: Whether variance is normalized = 0
INFO:
INFO: Reading HMM in Sphinx 3 Model format
INFO: Model Definition File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/
def/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/mdef
INFO: Mean File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sph
inx/model_parameters/atis_sphinx.cd_cont_3000/means
INFO: Variance File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis sphinx/model_parameters/atis_sphinx.cd_cont_3000/variances
INFO: Mixture Weight File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/de
f/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/mixture_weights
INFO: Transition Matrices File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspa
ce/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/transition_matric
es
INFO: mdef.c(682): Reading model definition: /home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/mdef
INFO: Initialization of mdef_t, report:
INFO: 40 CI-phone, 98785 CD-phone, 3 emitstate/phone, 120 CI-sen, 3120 Sen,
10327 Sen-Seq
INFO:
INFO: kbcore.c(298): Using optimized GMM computation for Continuous HMM, -topn
will be ignored
INFO: cont_mgau.c(163): Reading mixture gaussian file '/home/sachowdhury/ASR_S
IS_Lab_ATIS/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont
_3000/means'
INFO: cont_mgau.c(422): 3120 mixture Gaussians, 16 components, 1 streams,
veclen 39
INFO: cont_mgau.c(163): Reading mixture gaussian file '/home/sachowdhury/ASR_S
IS_Lab_ATIS/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont
_3000/variances'
INFO: cont_mgau.c(422): 3120 mixture Gaussians, 16 components, 1 streams,
veclen 39
INFO: cont_mgau.c(523): Reading mixture weights file '/home/sachowdhury/ASR_SI
S_Lab_ATIS/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont
3000/mixture_weights'
INFO: cont_mgau.c(678): Read 3120 x 16 mixture weights
INFO: cont_mgau.c(706): Removing uninitialized Gaussian densities
WARNING: "cont_mgau.c", line 780: 61 densities removed (0 mixtures removed
entirely)
INFO: cont_mgau.c(796): Applying variance floor
INFO: cont_mgau.c(814): 272 variance values floored
INFO: cont_mgau.c(862): Precomputing Mahalanobis distance invariants
INFO: tmat.c(169): Reading HMM transition probability matrices: /home/sachowdh
ury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphin
x.cd_cont_3000/transition_matrices
INFO: Initialization of tmat_t, report:
INFO: Read 40 transition matrices of size 3x4
INFO:
INFO: dict.c(383): Reading main dictionary: /home/sachowdhury/ASR_SIS_Lab_ATIS
/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.dic
INFO: dict.c(386): 133749 words read
INFO: dict.c(391): Reading filler dictionary: /home/sachowdhury/ASR_SIS_Lab_AT
IS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.filler
INFO: dict.c(394): 3 words read
INFO: dict.c(423): Added 0 fillers from mdef file
INFO: Initialization of dict_t, report:
INFO: No of CI phone: 0
INFO: Max word: 137848
INFO: No of word: 133752
INFO:
INFO: lmset.c(133): Reading LM ctl file
INFO: lmset.c(358): Reading LM control file '/home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.ctl'
INFO: lmclass.c(181): Reading LM Class file 'atis_sphinx.def'
SYSTEM_ERROR: "lmclass.c", line 183: Failed to open LM class file; No such
file or directory
Tue May 15 10:07:04 2012
Thank you.
The problem solved after I gave full path.
Another question I have is that if I want to map a word in multiple class
for example "four"
it can be inside class NUM
and also in TIME
in this type of case, error I am getting is :
ERROR: "wid.c", line 179: Multiple mappings of 'FOUR' in LM
Hope I got the problem right. So how can I define four in 2 classes?
If it is not the problem, can you please do suggest me what can i do .
Logs
INFO: lmset.c(133): Reading LM ctl file
INFO: lmset.c(358): Reading LM control file '/home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.ctl'
INFO: lmclass.c(181): Reading LM Class file '/home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.def'
INFO: lmclass.c(301): Loaded LM Class ''; 32 words
INFO: lmclass.c(301): Loaded LM Class ''; 18 words
INFO: lmclass.c(301): Loaded LM Class ''; 23 words
INFO: lmclass.c(301): Loaded LM Class ''; 17 words
INFO: lmclass.c(301): Loaded LM Class ''; 67 words
INFO: lmclass.c(301): Loaded LM Class ''; 7 words
INFO: lmclass.c(301): Loaded LM Class ''; 6 words
INFO: lmclass.c(301): Loaded LM Class ''; 10 words
INFO: lmclass.c(301): Loaded LM Class ''; 9 words
INFO: lmclass.c(301): Loaded LM Class ''; 74 words
INFO: lmclass.c(301): Loaded LM Class ''; 12 words
INFO: lmclass.c(301): Loaded LM Class ''; 22 words
INFO: lmclass.c(301): Loaded LM Class ''; 8 words
INFO: lmclass.c(301): Loaded LM Class ''; 1 words
INFO: lmclass.c(301): Loaded LM Class ''; 1 words
INFO: lmclass.c(301): Loaded LM Class ''; 175 words
INFO: lmclass.c(301): Loaded LM Class ''; 26 words
INFO: lmclass.c(301): Loaded LM Class ''; 26 words
INFO: lmclass.c(301): Loaded LM Class ''; 1 words
INFO: lmclass.c(301): Loaded LM Class ''; 9 words
INFO: lmclass.c(301): Loaded LM Class ''; 25 words
INFO: lmclass.c(301): Loaded LM Class ''; 219 words
INFO: lmclass.c(301): Loaded LM Class ''; 2 words
INFO: lmclass.c(301): Loaded LM Class ''; 7 words
INFO: lmset.c(418): Number of LM class specified 24 in file /home/sachowdhury/
ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.ctl
INFO: lm.c(606): LM read('/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def
/atis_sphinx/etc/atis_sphinx.lm.DMP', lw= 9.50, wip= 0.20, uw= 0.70)
INFO: lm.c(608): Reading LM file /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_worksp
ace/def/atis_sphinx/etc/atis_sphinx.lm.DMP (LM name "general")
INFO: lm_3g_dmp.c(630): Reading LM in 16 bits format
INFO: lm_3g_dmp.c(686): Read 1300 unigrams
INFO: lm_3g_dmp.c(759): 12040 bigrams
INFO: lm_3g_dmp.c(832): 11204 bigrams
INFO: lm_3g_dmp.c(902): 1343 bigram prob entries
INFO: lm_3g_dmp.c(936): 2848 trigram bowt entries
INFO: lm_3g_dmp.c(967): 2956 trigram prob entries
INFO: lm_3g_dmp.c(998): 24 trigram segtable entries (512 segsize)
INFO: lm_3g_dmp.c(1053): 1300 word strings
INFO: lm.c(691): The LM routine is operating at 16 bits mode
INFO: lmset.c(458): Did I enter here?
INFO: lmset.c(94): LM->inclass_ugscore size 133752
INFO: lmset.c(95): Number of class used 24
INFO: lmset.c(473): 16 1
INFO: lmset.c(478): No. of LM set allocated 16, no. of LM 1
INFO: wid.c(178): map = 2
ERROR: "wid.c", line 179: Multiple mappings of 'A' in LM
INFO: wid.c(178): map = 918
ERROR: "wid.c", line 179: Multiple mappings of 'S' in LM
INFO: wid.c(178): map = 390
ERROR: "wid.c", line 179: Multiple mappings of 'FIRST' in LM
INFO: wid.c(178): map = 393
ERROR: "wid.c", line 179: Multiple mappings of 'FIVE' in LM
INFO: wid.c(178): map = 698
.....
FATAL_ERROR: "wid.c", line 302: Error in mapping, please read the log to see
why
Thanks
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Just a question regarding the performance of Class Based LM.
Till my recent experiment, WER of my Wordbased LM experiment is 4.8 %
Where as WER for my class based is 5.1%
No matter how much the result improve WER of classbased is always lower than
my Word base.
But when I talk to other people, every one suggests that performance of Class
based should be more than using Word base language.
Is there is any reason for it ?
or is the problem with my work set?
I have also used other LM which is used to test performance of same data set
in HTK. It gives me more or less same performance as my LM.
Could you please be kind enough to give me any advice or suggestion on it
Really thank you for all your help
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
It's senseless to discuss accuracy issues without data in hands. There could
be dozen problems in your class-based lm which should affect the performance.
Starting from lower perplexity which you must test before running recognition
to overfitting. If you want help on accuracy issues you need to share your
data.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am extremely sorry for being so late. Please accept my apologies.
Here in the link below: I have attached my class based language model, its
defination file and its ctl file http://www.mediafire.com/?txnt8voe618db4d
For information related to perplexity:
CB - lm with uniform word probability has ppl 11
where as with non uniform word prob ppl = 23
my 3gram lm had ppl =14
Could you please mention what data will be useful for you to give me
suggestion.
I mean should I send you my Acoustic model or the full training set, as the
full data set is huge in size and can I email you the data in your email, as I
am working with ATIS data and I am not sure that am I allowed to upload data
over here.
Really sorry for so many troubles
and another question:
as my recent experiments:
I had wer = 4.71 and {4.14 } in s3 decoder
but when I used CB lm for my fwdtree mode my result is 5.20 but for fwdflat:
the decoder stops after initialising the search and no error is given in the
log.
Is there any specific reason?
-agc none none
-agcthresh 2.0 2.000000e+00
-alpha 0.97 9.700000e-01
-cep2spec no no
-ceplen 13 13
-cmn current current
-cmninit 8.0 8.0
-dither no no
-doublebw no no
-feat 1s_c_d_dd 1s_c_d_dd
-frate 100 100
-input_endian little little
-lda
-ldadim 0 0
-lifter 0 0
-logspec no no
-lowerf 133.33334 1.333333e+02
-ncep 13 13
-nfft 512 512
-nfilt 40 40
-remove_dc no yes
-round_filters yes yes
-samprate 16000 1.600000e+04
-seed -1 -1
-smoothspec no no
-spec2cep no no
-svspec
-transform legacy legacy
-unit_area yes yes
-upperf 6855.4976 6.855498e+03
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wlen 0.025625 2.000000e-02
INFO: Initialization of the log add table
INFO: Log-Add table size = 29356 x 2 >> 0
INFO:
INFO: feat.c(848): 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: kbcore.c(486): .cont.
INFO: Initialization of feat_t, report:
INFO: Feature type = 1s_c_d_dd
INFO: Cepstral size = 13
INFO: Number of streams = 1
INFO: Vector size of stream: 39
INFO: Number of subvectors = 0
INFO: Whether CMN is used = 1
INFO: Whether AGC is used = 0
INFO: Whether variance is normalized = 0
INFO:
INFO: Reading HMM in Sphinx 3 Model format
INFO: Model Definition File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/
Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/mdef
INFO: Mean File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_s
phinx/model_parameters/atis_sphinx.cd_cont_3000/means
INFO: Variance File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/at
is_sphinx/model_parameters/atis_sphinx.cd_cont_3000/variances
INFO: Mixture Weight File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/De
f5a/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/mixture_weights
INFO: Transition Matrices File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspa
ce/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/transition_matr
ices
INFO: mdef.c(682): Reading model definition: /home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/md
ef
INFO: Initialization of mdef_t, report:
INFO: 40 CI-phone, 134690 CD-phone, 3 emitstate/phone, 120 CI-sen, 3120 Sen,
12335 Sen-Seq
INFO:
INFO: kbcore.c(298): Using optimized GMM computation for Continuous HMM, -topn
will be ignored
INFO: cont_mgau.c(163): Reading mixture gaussian file '/home/sachowdhury/ASR_S
IS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_co
nt_3000/means'
INFO: cont_mgau.c(422): 3120 mixture Gaussians, 31 components, 1 streams,
veclen 39
INFO: cont_mgau.c(163): Reading mixture gaussian file '/home/sachowdhury/ASR_S
IS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_co
nt_3000/variances'
INFO: cont_mgau.c(422): 3120 mixture Gaussians, 31 components, 1 streams,
veclen 39
INFO: cont_mgau.c(523): Reading mixture weights file '/home/sachowdhury/ASR_SI
S_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_con
t_3000/mixture_weights'
INFO: cont_mgau.c(678): Read 3120 x 31 mixture weights
INFO: cont_mgau.c(706): Removing uninitialized Gaussian densities
117 118 119 3117 3118 3119
WARNING: "cont_mgau.c", line 780: 4031 densities removed (6 mixtures removed
entirely)
INFO: cont_mgau.c(796): Applying variance floor
INFO: cont_mgau.c(814): 5507 variance values floored
INFO: cont_mgau.c(862): Precomputing Mahalanobis distance invariants
INFO: tmat.c(169): Reading HMM transition probability matrices: /home/sachowdh
ury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sph
inx.cd_cont_3000/transition_matrices
INFO: Initialization of tmat_t, report:
INFO: Read 40 transition matrices of size 3x4
INFO:
INFO: dict.c(383): Reading main dictionary: /home/sachowdhury/ASR_SIS_Lab_ATIS
/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.dic
INFO: dict.c(386): 133839 words read
INFO: dict.c(391): Reading filler dictionary: /home/sachowdhury/ASR_SIS_Lab_AT
IS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.filler
INFO: dict.c(394): 3 words read
INFO: dict.c(423): Added 0 fillers from mdef file
INFO: Initialization of dict_t, report:
INFO: No of CI phone: 0
INFO: Max word: 137938
INFO: No of word: 133842
INFO:
INFO: lmset.c(133): Reading LM ctl file
INFO: lmset.c(358): Reading LM control file '/home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.ctl'
INFO: lmclass.c(181): Reading LM Class file '/home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.def'
INFO: lmclass.c(301): Loaded LM Class ''; 32 words
INFO: lmclass.c(301): Loaded LM Class ''; 18 words
INFO: lmclass.c(301): Loaded LM Class ''; 23 words
INFO: lmclass.c(301): Loaded LM Class ''; 17 words
INFO: lmclass.c(301): Loaded LM Class ''; 7 words
INFO: lmclass.c(301): Loaded LM Class ''; 6 words
INFO: lmclass.c(301): Loaded LM Class ''; 10 words
INFO: lmclass.c(301): Loaded LM Class ''; 9 words
INFO: lmclass.c(301): Loaded LM Class ''; 12 words
INFO: lmclass.c(301): Loaded LM Class ''; 175 words
INFO: lmclass.c(301): Loaded LM Class ''; 26 words
INFO: lmclass.c(301): Loaded LM Class ''; 26 words
INFO: lmclass.c(301): Loaded LM Class ''; 1 words
INFO: lmclass.c(301): Loaded LM Class ''; 25 words
INFO: lmclass.c(301): Loaded LM Class ''; 2 words
INFO: lmclass.c(301): Loaded LM Class ''; 7 words
INFO: lmset.c(418): Number of LM class specified 16 in file /home/sachowdhury/
ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.ctl
INFO: lm.c(606): LM read('/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def
5a/atis_sphinx/etc/class_based_model.lm.DMP', lw= 14.00, wip= 0.20, uw= 0.80)
INFO: lm.c(608): Reading LM file /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_worksp
ace/Def5a/atis_sphinx/etc/class_based_model.lm.DMP (LM name "general")
INFO: lm_3g_dmp.c(630): Reading LM in 16 bits format
INFO: lm_3g_dmp.c(686): Read 1868 unigrams
INFO: lm_3g_dmp.c(759): 17189 bigrams
INFO: lm_3g_dmp.c(832): 14981 bigrams
INFO: lm_3g_dmp.c(902): 9852 bigram prob entries
INFO: lm_3g_dmp.c(936): 2675 trigram bowt entries
INFO: lm_3g_dmp.c(967): 10487 trigram prob entries
INFO: lm_3g_dmp.c(998): 34 trigram segtable entries (512 segsize)
INFO: lm_3g_dmp.c(1053): 1868 word strings
INFO: lm.c(691): The LM routine is operating at 16 bits mode
INFO: lmset.c(458): Did I enter here?
INFO: lmset.c(94): LM->inclass_ugscore size 133842
INFO: lmset.c(95): Number of class used 16
INFO: lmset.c(473): 16 1
INFO: lmset.c(478): No. of LM set allocated 16, no. of LM 1
INFO: Initialization of fillpen_t, report:
INFO: Language weight =14.000000
INFO: Word Insertion Penalty =0.200000
INFO: Silence probability =0.800000
INFO: Filler probability =0.100000
INFO:
INFO: dict2pid.c(599): Building PID tables for dictionary
INFO: Initialization of dict2pid_t, report:
INFO: Dict2pid is in composite triphone mode
INFO: 2484 composite states; 1117 composite sseq
INFO:
INFO: kbcore.c(628): Inside kbcore: Verifying models consistency ......
INFO: kbcore.c(650): End of Initialization of Core Models:
INFO: Initialization of beam_t, report:
INFO: Parameters used in Beam Pruning of Viterbi Search:
INFO: Beam=-614114
INFO: PBeam=-614114
INFO: WBeam=-460586 (Skip=0)
INFO: WEndBeam=-614114
INFO: No of CI Phone assumed=40
INFO:
INFO: Initialization of fast_gmm_t, report:
INFO: Parameters used in Fast GMM computation:
INFO: Frame-level: Down Sampling Ratio 1, Conditional Down Sampling? 0,
Distance-based Down Sampling? 0
INFO: GMM-level: CI phone beam -614114. MAX CD 100000
INFO: Gaussian-level: GS map would be used for Gaussian Selection? =1, SVQ
would be used as Gaussian Score? =0 SubVQ Beam -19366
INFO:
INFO: Initialization of pl_t, report:
INFO: Parameters used in phoneme lookahead:
INFO: Phoneme look-ahead type = 0
INFO: Phoneme look-ahead beam size = -614114
INFO: No of CI Phones assumed=40
INFO:
INFO: Initialization of ascr_t, report:
INFO: No. of CI senone =120
INFO: No. of senone = 3120
INFO: No. of composite senone = 2484
INFO: No. of senone sequence = 12335
INFO: No. of composite senone sequence=1117
INFO: Parameters used in phoneme lookahead:
INFO: Phoneme lookahead window = 1
INFO:
INFO: kb.c(308): SEARCH MODE INDEX 3
INFO: srch.c(374): Search Initialization.
INFO: srch_flat_fwd.c(204): Initialization
Thu Jun 14 19:57:48 2012
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello,
I am recently trying to use a classbased LM for my previously trained acoustic
model.
I have created files:
a control file :
a def fie
corresponding lm and its dictionary.
When I run the decoder, it gives me error like:
INFO: lmclass.c(181): Reading LM Class file 'atis_sphinx.def'
SYSTEM_ERROR: "lmclass.c", line 183: Failed to open LM class file; No such
file or directory
Can any one please suggest me, what should be my steps to overcome this
problem?
Am I missing anything ?
I am using turtle class lm example to format my files.
Here is my log:
INFO: info.c(65): Host: 'passidin1'
INFO: info.c(66): Directory:
'/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx'
INFO: info.c(70):
/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/bin/.libs/lt-
sphinx3_decode Compiled on: Apr 18 2012, AT: 12:35:12
INFO: cmd_ln.c(510): Parsing command line:
/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/bin/.libs/lt-
sphinx3_decode \
-senmgau .cont. \
-hmm /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000 \
-lw 9.5 \
-feat 1s_c_d_dd \
-beam 1e-60 \
-wbeam 1e-30 \
-dict /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.dic \
-fdict /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.filler \
-lmctlfn /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.ctl \
-wip 0.2 \
-ctl /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx_nov93_test.fileids \
-ctloffset 0 \
-ctlcount 10 \
-cepdir /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/feat \
-cepext .mfc \
-hyp /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/result/atis_sphinx-1-1.match \
-agc none \
-varnorm no \
-cmn current
Current configuration:
-adchdr 0 0
-adcin no no
-agc none none
-agcthresh 2.0 2.000000e+00
-alpha 0.97 9.700000e-01
-backtrace yes yes
-beam 1.0e-55 1.000000e-60
-bestpath no no
-bestpathlw 0.000000e+00
-bestscoredir
-bestsenscrdir
-bghist no no
-bptbldir
-bptblsize 32768 32768
-cb2mllr .1cls. .1cls.
-cep2spec no no
-cepdir /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/feat
-cepext .mfc .mfc
-ceplen 13 13
-ci_pbeam 1e-80 1.000000e-80
-cmn current current
-cmninit 8.0 8.0
-cond_ds no no
-ctl /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx_nov93_test.fileids
-ctlcount 1000000000 10
-ctloffset 0 0
-ctl_lm
-ctl_mllr
-dagfudge 2 2
-debug 0
-dict /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.dic
-dist_ds no no
-dither no no
-doublebw no no
-ds 1 1
-epl 3 3
-fdict /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.filler
-feat 1s_c_d_dd 1s_c_d_dd
-featparams
-fillpen
-fillprob 0.1 1.000000e-01
-frate 100 100
-fsg
-fsgusealtpron yes yes
-fsgusefiller yes yes
-gs
-gs4gs yes yes
-hmm /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000
-hmmdump no no
-hmmdumpef 200000000 200000000
-hmmdumpsf 200000000 200000000
-hmmhistbinsize 5000 5000
-hyp /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/result/atis_sphinx-1-1.match
-hypseg
-hypsegscore_unscale yes yes
-inlatdir
-inlatwin 50 50
-input_endian little little
-kdmaxbbi -1 -1
-kdmaxdepth 0 0
-kdtree
-latcompress yes yes
-latext lat.gz lat.gz
-lda
-ldadim 0 0
-lextreedump 0 0
-lifter 0 0
-lm
-lmctlfn /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.ctl
-lmdumpdir
-lmname
-log3table yes yes
-logbase 1.0003 1.000300e+00
-logfn
-logspec no no
-lowerf 133.33334 1.333333e+02
-lts_mismatch no no
-lw 9.5 9.500000e+00
-maxcdsenpf 100000 100000
-maxedge 2000000 2000000
-maxhistpf 100 100
-maxhmmpf 20000 20000
-maxlmop 100000000 100000000
-maxlpf 40000 40000
-maxppath 1000000 1000000
-maxwpf 20 20
-mdef
-mean
-min_endfr 3 3
-mixw
-mixwfloor 0.0000001 1.000000e-07
-mllr
-mode fwdtree fwdtree
-nbest 200 200
-nbestdir
-nbestext nbest.gz nbest.gz
-ncep 13 13
-nfft 512 512
-nfilt 40 40
-Nlextree 3 3
-Nstalextree 25 25
-op_mode -1 -1
-outlatdir
-outlatfmt s3 s3
-pbeam 1.0e-50 1.000000e-50
-pheurtype 0 0
-phonepen 1.0 1.000000e+00
-phsegdir
-pl_beam 1.0e-80 1.000000e-80
-pl_window 1 1
-ppathdebug no no
-ptranskip 0 0
-remove_dc no no
-round_filters yes yes
-samprate 16000 1.600000e+04
-seed -1 -1
-sendump
-senmgau .cont. .cont.
-silprob 0.1 1.000000e-01
-smoothspec no no
-spec2cep no no
-subvq
-subvqbeam 3.0e-3 3.000000e-03
-svq4svq no no
-svspec
-tighten_factor 0.5 5.000000e-01
-tmat
-tmatfloor 0.0001 1.000000e-04
-topn 4 4
-topn_beam 0 0
-tracewhmm
-transform legacy legacy
-treeugprob yes yes
-unit_area yes yes
-upperf 6855.4976 6.855498e+03
-utt
-uw 0.7 7.000000e-01
-var
-varfloor 0.0001 1.000000e-04
-varnorm no no
-verbose no no
-vqeval 3 3
-warp_params
-warp_type inverse_linear inverse_linear
-wbeam 1.0e-35 1.000000e-30
-wend_beam 1.0e-80 1.000000e-80
-wip 0.7 2.000000e-01
-wlen 0.025625 2.562500e-02
-worddumpef 200000000 200000000
-worddumpsf 200000000 200000000
INFO: kbcore.c(439): Begin Initialization of Core Models:
INFO: cmd_ln.c(510): Parsing command line:
\
-alpha 0.97 \
-remove_dc yes \
-doublebw no \
-nfilt 40 \
-ncep 13 \
-lowerf 133.33334 \
-upperf 6855.4976 \
-nfft 512 \
-wlen 0.0200 \
-transform legacy \
-feat 1s_c_d_dd \
-agc none \
-cmn current \
-varnorm no
Current configuration:
-agc none none
-agcthresh 2.0 2.000000e+00
-alpha 0.97 9.700000e-01
-cep2spec no no
-ceplen 13 13
-cmn current current
-cmninit 8.0 8.0
-dither no no
-doublebw no no
-feat 1s_c_d_dd 1s_c_d_dd
-frate 100 100
-input_endian little little
-lda
-ldadim 0 0
-lifter 0 0
-logspec no no
-lowerf 133.33334 1.333333e+02
-ncep 13 13
-nfft 512 512
-nfilt 40 40
-remove_dc no yes
-round_filters yes yes
-samprate 16000 1.600000e+04
-seed -1 -1
-smoothspec no no
-spec2cep no no
-svspec
-transform legacy legacy
-unit_area yes yes
-upperf 6855.4976 6.855498e+03
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wlen 0.025625 2.000000e-02
INFO: Initialization of the log add table
INFO: Log-Add table size = 29356 x 2 >> 0
INFO:
INFO: feat.c(848): 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: kbcore.c(486): .cont.
INFO: Initialization of feat_t, report:
INFO: Feature type = 1s_c_d_dd
INFO: Cepstral size = 13
INFO: Number of streams = 1
INFO: Vector size of stream: 39
INFO: Number of subvectors = 0
INFO: Whether CMN is used = 1
INFO: Whether AGC is used = 0
INFO: Whether variance is normalized = 0
INFO:
INFO: Reading HMM in Sphinx 3 Model format
INFO: Model Definition File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/
def/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/mdef
INFO: Mean File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sph
inx/model_parameters/atis_sphinx.cd_cont_3000/means
INFO: Variance File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis
sphinx/model_parameters/atis_sphinx.cd_cont_3000/variances
INFO: Mixture Weight File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/de
f/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/mixture_weights
INFO: Transition Matrices File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspa
ce/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/transition_matric
es
INFO: mdef.c(682): Reading model definition: /home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/mdef
INFO: Initialization of mdef_t, report:
INFO: 40 CI-phone, 98785 CD-phone, 3 emitstate/phone, 120 CI-sen, 3120 Sen,
10327 Sen-Seq
INFO:
INFO: kbcore.c(298): Using optimized GMM computation for Continuous HMM, -topn
will be ignored
INFO: cont_mgau.c(163): Reading mixture gaussian file '/home/sachowdhury/ASR_S
IS_Lab_ATIS/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont
_3000/means'
INFO: cont_mgau.c(422): 3120 mixture Gaussians, 16 components, 1 streams,
veclen 39
INFO: cont_mgau.c(163): Reading mixture gaussian file '/home/sachowdhury/ASR_S
IS_Lab_ATIS/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont
_3000/variances'
INFO: cont_mgau.c(422): 3120 mixture Gaussians, 16 components, 1 streams,
veclen 39
INFO: cont_mgau.c(523): Reading mixture weights file '/home/sachowdhury/ASR_SI
S_Lab_ATIS/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphinx.cd_cont
3000/mixture_weights'
INFO: cont_mgau.c(678): Read 3120 x 16 mixture weights
INFO: cont_mgau.c(706): Removing uninitialized Gaussian densities
WARNING: "cont_mgau.c", line 780: 61 densities removed (0 mixtures removed
entirely)
INFO: cont_mgau.c(796): Applying variance floor
INFO: cont_mgau.c(814): 272 variance values floored
INFO: cont_mgau.c(862): Precomputing Mahalanobis distance invariants
INFO: tmat.c(169): Reading HMM transition probability matrices: /home/sachowdh
ury/ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/model_parameters/atis_sphin
x.cd_cont_3000/transition_matrices
INFO: Initialization of tmat_t, report:
INFO: Read 40 transition matrices of size 3x4
INFO:
INFO: dict.c(383): Reading main dictionary: /home/sachowdhury/ASR_SIS_Lab_ATIS
/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.dic
INFO: dict.c(386): 133749 words read
INFO: dict.c(391): Reading filler dictionary: /home/sachowdhury/ASR_SIS_Lab_AT
IS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.filler
INFO: dict.c(394): 3 words read
INFO: dict.c(423): Added 0 fillers from mdef file
INFO: Initialization of dict_t, report:
INFO: No of CI phone: 0
INFO: Max word: 137848
INFO: No of word: 133752
INFO:
INFO: lmset.c(133): Reading LM ctl file
INFO: lmset.c(358): Reading LM control file '/home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.ctl'
INFO: lmclass.c(181): Reading LM Class file 'atis_sphinx.def'
SYSTEM_ERROR: "lmclass.c", line 183: Failed to open LM class file; No such
file or directory
Tue May 15 10:07:04 2012
and my ctl file def file is attached in
http://www.mediafire.com/?ia19k3kzgiifhqq
http://www.mediafire.com/?t6aop9n60u3f9oe
Please any help or suggestion will be very helpful
Thank You
Watch this
http://www.youtube.com/watch?v=M3JirqAPg9g
Read this
http://www.doc.ic.ac.uk/~wjk/UnixIntro/
Thank you.
The problem solved after I gave full path.
Another question I have is that if I want to map a word in multiple class
for example "four"
it can be inside class NUM
and also in TIME
in this type of case, error I am getting is :
ERROR: "wid.c", line 179: Multiple mappings of 'FOUR' in LM
Hope I got the problem right. So how can I define four in 2 classes?
If it is not the problem, can you please do suggest me what can i do .
Logs
INFO: lmset.c(133): Reading LM ctl file
INFO: lmset.c(358): Reading LM control file '/home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.ctl'
INFO: lmclass.c(181): Reading LM Class file '/home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.def'
INFO: lmclass.c(301): Loaded LM Class ''; 32 words
INFO: lmclass.c(301): Loaded LM Class ''; 18 words
INFO: lmclass.c(301): Loaded LM Class ''; 23 words
INFO: lmclass.c(301): Loaded LM Class ''; 17 words
INFO: lmclass.c(301): Loaded LM Class ''; 67 words
INFO: lmclass.c(301): Loaded LM Class ''; 7 words
INFO: lmclass.c(301): Loaded LM Class ''; 6 words
INFO: lmclass.c(301): Loaded LM Class ''; 10 words
INFO: lmclass.c(301): Loaded LM Class ''; 9 words
INFO: lmclass.c(301): Loaded LM Class ''; 74 words
INFO: lmclass.c(301): Loaded LM Class ''; 12 words
INFO: lmclass.c(301): Loaded LM Class ''; 22 words
INFO: lmclass.c(301): Loaded LM Class ''; 8 words
INFO: lmclass.c(301): Loaded LM Class ''; 1 words
INFO: lmclass.c(301): Loaded LM Class ''; 1 words
INFO: lmclass.c(301): Loaded LM Class ''; 175 words
INFO: lmclass.c(301): Loaded LM Class ''; 26 words
INFO: lmclass.c(301): Loaded LM Class ''; 26 words
INFO: lmclass.c(301): Loaded LM Class ''; 1 words
INFO: lmclass.c(301): Loaded LM Class ''; 9 words
INFO: lmclass.c(301): Loaded LM Class ''; 25 words
INFO: lmclass.c(301): Loaded LM Class ''; 219 words
INFO: lmclass.c(301): Loaded LM Class ''; 2 words
INFO: lmclass.c(301): Loaded LM Class ''; 7 words
INFO: lmset.c(418): Number of LM class specified 24 in file /home/sachowdhury/
ASR_SIS_Lab_ATIS/ASR_workspace/def/atis_sphinx/etc/atis_sphinx.ctl
INFO: lm.c(606): LM read('/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/def
/atis_sphinx/etc/atis_sphinx.lm.DMP', lw= 9.50, wip= 0.20, uw= 0.70)
INFO: lm.c(608): Reading LM file /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_worksp
ace/def/atis_sphinx/etc/atis_sphinx.lm.DMP (LM name "general")
INFO: lm_3g_dmp.c(630): Reading LM in 16 bits format
INFO: lm_3g_dmp.c(686): Read 1300 unigrams
INFO: lm_3g_dmp.c(759): 12040 bigrams
INFO: lm_3g_dmp.c(832): 11204 bigrams
INFO: lm_3g_dmp.c(902): 1343 bigram prob entries
INFO: lm_3g_dmp.c(936): 2848 trigram bowt entries
INFO: lm_3g_dmp.c(967): 2956 trigram prob entries
INFO: lm_3g_dmp.c(998): 24 trigram segtable entries (512 segsize)
INFO: lm_3g_dmp.c(1053): 1300 word strings
INFO: lm.c(691): The LM routine is operating at 16 bits mode
INFO: lmset.c(458): Did I enter here?
INFO: lmset.c(94): LM->inclass_ugscore size 133752
INFO: lmset.c(95): Number of class used 24
INFO: lmset.c(473): 16 1
INFO: lmset.c(478): No. of LM set allocated 16, no. of LM 1
INFO: wid.c(178): map = 2
ERROR: "wid.c", line 179: Multiple mappings of 'A' in LM
INFO: wid.c(178): map = 918
ERROR: "wid.c", line 179: Multiple mappings of 'S' in LM
INFO: wid.c(178): map = 390
ERROR: "wid.c", line 179: Multiple mappings of 'FIRST' in LM
INFO: wid.c(178): map = 393
ERROR: "wid.c", line 179: Multiple mappings of 'FIVE' in LM
INFO: wid.c(178): map = 698
.....
FATAL_ERROR: "wid.c", line 302: Error in mapping, please read the log to see
why
Thanks
Can I give entry like this in dictionary?
You need to use distinct strings in this case, for example FOUR_TIME and
FOUR_NUM with the same phonetic transcription.
Thank you.
I am giving a try
Just a question regarding the performance of Class Based LM.
Till my recent experiment, WER of my Wordbased LM experiment is 4.8 %
Where as WER for my class based is 5.1%
No matter how much the result improve WER of classbased is always lower than
my Word base.
But when I talk to other people, every one suggests that performance of Class
based should be more than using Word base language.
Is there is any reason for it ?
or is the problem with my work set?
I have also used other LM which is used to test performance of same data set
in HTK. It gives me more or less same performance as my LM.
Could you please be kind enough to give me any advice or suggestion on it
Really thank you for all your help
It's senseless to discuss accuracy issues without data in hands. There could
be dozen problems in your class-based lm which should affect the performance.
Starting from lower perplexity which you must test before running recognition
to overfitting. If you want help on accuracy issues you need to share your
data.
I am extremely sorry for being so late. Please accept my apologies.
Here in the link below: I have attached my class based language model, its
defination file and its ctl file
http://www.mediafire.com/?txnt8voe618db4d
For information related to perplexity:
CB - lm with uniform word probability has ppl 11
where as with non uniform word prob ppl = 23
my 3gram lm had ppl =14
Could you please mention what data will be useful for you to give me
suggestion.
I mean should I send you my Acoustic model or the full training set, as the
full data set is huge in size and can I email you the data in your email, as I
am working with ATIS data and I am not sure that am I allowed to upload data
over here.
Really sorry for so many troubles
and another question:
as my recent experiments:
I had wer = 4.71 and {4.14 } in s3 decoder
but when I used CB lm for my fwdtree mode my result is 5.20 but for fwdflat:
the decoder stops after initialising the search and no error is given in the
log.
Is there any specific reason?
Here is another post I looked but it is seen they might face same problem like
me:
https://sourceforge.net/projects/cmusphinx/forums/forum/5471/topic/1686557
https://sourceforge.net/projects/cmusphinx/forums/forum/5471/topic/1690251
Here is my log:
INFO: info.c(65): Host: 'passidin5.science.unitn.it'
INFO: info.c(66): Directory:
'/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx'
INFO: info.c(70):
/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/bin/.libs
/lt-sphinx3_decode Compiled on: Apr 18 2012, AT: 12:35:12
INFO: cmd_ln.c(510): Parsing command line:
/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/bin/.libs
/lt-sphinx3_decode \
-senmgau .cont. \
-hmm /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000 \
-lw 14 \
-feat 1s_c_d_dd \
-beam 1e-80 \
-wbeam 1e-60 \
-dict /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.dic \
-fdict /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.filler \
-lmctlfn /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.ctl \
-uw 0.8 \
-maxhmmpf 20000 \
-maxwpf 20 \
-maxhistpf 100 \
-subvqbeam 3e-3 \
-mode fwdflat \
-silprob 0.8 \
-pl_beam 1e-80 \
-pbeam 1e-80 \
-wip 0.2 \
-ctl /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx_test.fileids \
-ctloffset 0 \
-ctlcount 981 \
-cepdir /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/feat \
-cepext .mfc \
-hyp /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/result/atis_sphinx-1-1.match \
-agc none \
-varnorm no \
-cmn current
Current configuration:
-adchdr 0 0
-adcin no no
-agc none none
-agcthresh 2.0 2.000000e+00
-alpha 0.97 9.700000e-01
-backtrace yes yes
-beam 1.0e-55 1.000000e-80
-bestpath no no
-bestpathlw 0.000000e+00
-bestscoredir
-bestsenscrdir
-bghist no no
-bptbldir
-bptblsize 32768 32768
-cb2mllr .1cls. .1cls.
-cep2spec no no
-cepdir /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/feat
-cepext .mfc .mfc
-ceplen 13 13
-ci_pbeam 1e-80 1.000000e-80
-cmn current current
-cmninit 8.0 8.0
-cond_ds no no
-ctl /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx_test.fileids
-ctlcount 1000000000 981
-ctloffset 0 0
-ctl_lm
-ctl_mllr
-dagfudge 2 2
-debug 0
-dict /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.dic
-dist_ds no no
-dither no no
-doublebw no no
-ds 1 1
-epl 3 3
-fdict /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.filler
-feat 1s_c_d_dd 1s_c_d_dd
-featparams
-fillpen
-fillprob 0.1 1.000000e-01
-frate 100 100
-fsg
-fsgusealtpron yes yes
-fsgusefiller yes yes
-gs
-gs4gs yes yes
-hmm /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000
-hmmdump no no
-hmmdumpef 200000000 200000000
-hmmdumpsf 200000000 200000000
-hmmhistbinsize 5000 5000
-hyp /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/result/atis_sphinx-1-1.match
-hypseg
-hypsegscore_unscale yes yes
-inlatdir
-inlatwin 50 50
-input_endian little little
-kdmaxbbi -1 -1
-kdmaxdepth 0 0
-kdtree
-latcompress yes yes
-latext lat.gz lat.gz
-lda
-ldadim 0 0
-lextreedump 0 0
-lifter 0 0
-lm
-lmctlfn /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.ctl
-lmdumpdir
-lmname
-log3table yes yes
-logbase 1.0003 1.000300e+00
-logfn
-logspec no no
-lowerf 133.33334 1.333333e+02
-lts_mismatch no no
-lw 9.5 1.400000e+01
-maxcdsenpf 100000 100000
-maxedge 2000000 2000000
-maxhistpf 100 100
-maxhmmpf 20000 20000
-maxlmop 100000000 100000000
-maxlpf 40000 40000
-maxppath 1000000 1000000
-maxwpf 20 20
-mdef
-mean
-min_endfr 3 3
-mixw
-mixwfloor 0.0000001 1.000000e-07
-mllr
-mode fwdtree fwdflat
-nbest 200 200
-nbestdir
-nbestext nbest.gz nbest.gz
-ncep 13 13
-nfft 512 512
-nfilt 40 40
-Nlextree 3 3
-Nstalextree 25 25
-op_mode -1 -1
-outlatdir
-outlatfmt s3 s3
-pbeam 1.0e-50 1.000000e-80
-pheurtype 0 0
-phonepen 1.0 1.000000e+00
-phsegdir
-pl_beam 1.0e-80 1.000000e-80
-pl_window 1 1
-ppathdebug no no
-ptranskip 0 0
-remove_dc no no
-round_filters yes yes
-samprate 16000 1.600000e+04
-seed -1 -1
-sendump
-senmgau .cont. .cont.
-silprob 0.1 8.000000e-01
-smoothspec no no
-spec2cep no no
-subvq
-subvqbeam 3.0e-3 3.000000e-03
-svq4svq no no
-svspec
-tighten_factor 0.5 5.000000e-01
-tmat
-tmatfloor 0.0001 1.000000e-04
-topn 4 4
-topn_beam 0 0
-tracewhmm
-transform legacy legacy
-treeugprob yes yes
-unit_area yes yes
-upperf 6855.4976 6.855498e+03
-utt
-uw 0.7 8.000000e-01
-var
-varfloor 0.0001 1.000000e-04
-varnorm no no
-verbose no no
-vqeval 3 3
-warp_params
-warp_type inverse_linear inverse_linear
-wbeam 1.0e-35 1.000000e-60
-wend_beam 1.0e-80 1.000000e-80
-wip 0.7 2.000000e-01
-wlen 0.025625 2.562500e-02
-worddumpef 200000000 200000000
-worddumpsf 200000000 200000000
INFO: kbcore.c(439): Begin Initialization of Core Models:
INFO: cmd_ln.c(510): Parsing command line:
\
-alpha 0.97 \
-remove_dc yes \
-doublebw no \
-nfilt 40 \
-ncep 13 \
-lowerf 133.33334 \
-upperf 6855.4976 \
-nfft 512 \
-wlen 0.0200 \
-transform legacy \
-feat 1s_c_d_dd \
-agc none \
-cmn current \
-varnorm no
Current configuration:
-agc none none
-agcthresh 2.0 2.000000e+00
-alpha 0.97 9.700000e-01
-cep2spec no no
-ceplen 13 13
-cmn current current
-cmninit 8.0 8.0
-dither no no
-doublebw no no
-feat 1s_c_d_dd 1s_c_d_dd
-frate 100 100
-input_endian little little
-lda
-ldadim 0 0
-lifter 0 0
-logspec no no
-lowerf 133.33334 1.333333e+02
-ncep 13 13
-nfft 512 512
-nfilt 40 40
-remove_dc no yes
-round_filters yes yes
-samprate 16000 1.600000e+04
-seed -1 -1
-smoothspec no no
-spec2cep no no
-svspec
-transform legacy legacy
-unit_area yes yes
-upperf 6855.4976 6.855498e+03
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wlen 0.025625 2.000000e-02
INFO: Initialization of the log add table
INFO: Log-Add table size = 29356 x 2 >> 0
INFO:
INFO: feat.c(848): 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: kbcore.c(486): .cont.
INFO: Initialization of feat_t, report:
INFO: Feature type = 1s_c_d_dd
INFO: Cepstral size = 13
INFO: Number of streams = 1
INFO: Vector size of stream: 39
INFO: Number of subvectors = 0
INFO: Whether CMN is used = 1
INFO: Whether AGC is used = 0
INFO: Whether variance is normalized = 0
INFO:
INFO: Reading HMM in Sphinx 3 Model format
INFO: Model Definition File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/
Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/mdef
INFO: Mean File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_s
phinx/model_parameters/atis_sphinx.cd_cont_3000/means
INFO: Variance File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/at
is_sphinx/model_parameters/atis_sphinx.cd_cont_3000/variances
INFO: Mixture Weight File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/De
f5a/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/mixture_weights
INFO: Transition Matrices File: /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspa
ce/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/transition_matr
ices
INFO: mdef.c(682): Reading model definition: /home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_cont_3000/md
ef
INFO: Initialization of mdef_t, report:
INFO: 40 CI-phone, 134690 CD-phone, 3 emitstate/phone, 120 CI-sen, 3120 Sen,
12335 Sen-Seq
INFO:
INFO: kbcore.c(298): Using optimized GMM computation for Continuous HMM, -topn
will be ignored
INFO: cont_mgau.c(163): Reading mixture gaussian file '/home/sachowdhury/ASR_S
IS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_co
nt_3000/means'
INFO: cont_mgau.c(422): 3120 mixture Gaussians, 31 components, 1 streams,
veclen 39
INFO: cont_mgau.c(163): Reading mixture gaussian file '/home/sachowdhury/ASR_S
IS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_co
nt_3000/variances'
INFO: cont_mgau.c(422): 3120 mixture Gaussians, 31 components, 1 streams,
veclen 39
INFO: cont_mgau.c(523): Reading mixture weights file '/home/sachowdhury/ASR_SI
S_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sphinx.cd_con
t_3000/mixture_weights'
INFO: cont_mgau.c(678): Read 3120 x 31 mixture weights
INFO: cont_mgau.c(706): Removing uninitialized Gaussian densities
117 118 119 3117 3118 3119
WARNING: "cont_mgau.c", line 780: 4031 densities removed (6 mixtures removed
entirely)
INFO: cont_mgau.c(796): Applying variance floor
INFO: cont_mgau.c(814): 5507 variance values floored
INFO: cont_mgau.c(862): Precomputing Mahalanobis distance invariants
INFO: tmat.c(169): Reading HMM transition probability matrices: /home/sachowdh
ury/ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/model_parameters/atis_sph
inx.cd_cont_3000/transition_matrices
INFO: Initialization of tmat_t, report:
INFO: Read 40 transition matrices of size 3x4
INFO:
INFO: dict.c(383): Reading main dictionary: /home/sachowdhury/ASR_SIS_Lab_ATIS
/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.dic
INFO: dict.c(386): 133839 words read
INFO: dict.c(391): Reading filler dictionary: /home/sachowdhury/ASR_SIS_Lab_AT
IS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.filler
INFO: dict.c(394): 3 words read
INFO: dict.c(423): Added 0 fillers from mdef file
INFO: Initialization of dict_t, report:
INFO: No of CI phone: 0
INFO: Max word: 137938
INFO: No of word: 133842
INFO:
INFO: lmset.c(133): Reading LM ctl file
INFO: lmset.c(358): Reading LM control file '/home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.ctl'
INFO: lmclass.c(181): Reading LM Class file '/home/sachowdhury/ASR_SIS_Lab_ATI
S/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.def'
INFO: lmclass.c(301): Loaded LM Class ''; 32 words
INFO: lmclass.c(301): Loaded LM Class ''; 18 words
INFO: lmclass.c(301): Loaded LM Class ''; 23 words
INFO: lmclass.c(301): Loaded LM Class ''; 17 words
INFO: lmclass.c(301): Loaded LM Class ''; 7 words
INFO: lmclass.c(301): Loaded LM Class ''; 6 words
INFO: lmclass.c(301): Loaded LM Class ''; 10 words
INFO: lmclass.c(301): Loaded LM Class ''; 9 words
INFO: lmclass.c(301): Loaded LM Class ''; 12 words
INFO: lmclass.c(301): Loaded LM Class ''; 175 words
INFO: lmclass.c(301): Loaded LM Class ''; 26 words
INFO: lmclass.c(301): Loaded LM Class ''; 26 words
INFO: lmclass.c(301): Loaded LM Class ''; 1 words
INFO: lmclass.c(301): Loaded LM Class ''; 25 words
INFO: lmclass.c(301): Loaded LM Class ''; 2 words
INFO: lmclass.c(301): Loaded LM Class ''; 7 words
INFO: lmset.c(418): Number of LM class specified 16 in file /home/sachowdhury/
ASR_SIS_Lab_ATIS/ASR_workspace/Def5a/atis_sphinx/etc/atis_sphinx.ctl
INFO: lm.c(606): LM read('/home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_workspace/Def
5a/atis_sphinx/etc/class_based_model.lm.DMP', lw= 14.00, wip= 0.20, uw= 0.80)
INFO: lm.c(608): Reading LM file /home/sachowdhury/ASR_SIS_Lab_ATIS/ASR_worksp
ace/Def5a/atis_sphinx/etc/class_based_model.lm.DMP (LM name "general")
INFO: lm_3g_dmp.c(630): Reading LM in 16 bits format
INFO: lm_3g_dmp.c(686): Read 1868 unigrams
INFO: lm_3g_dmp.c(759): 17189 bigrams
INFO: lm_3g_dmp.c(832): 14981 bigrams
INFO: lm_3g_dmp.c(902): 9852 bigram prob entries
INFO: lm_3g_dmp.c(936): 2675 trigram bowt entries
INFO: lm_3g_dmp.c(967): 10487 trigram prob entries
INFO: lm_3g_dmp.c(998): 34 trigram segtable entries (512 segsize)
INFO: lm_3g_dmp.c(1053): 1868 word strings
INFO: lm.c(691): The LM routine is operating at 16 bits mode
INFO: lmset.c(458): Did I enter here?
INFO: lmset.c(94): LM->inclass_ugscore size 133842
INFO: lmset.c(95): Number of class used 16
INFO: lmset.c(473): 16 1
INFO: lmset.c(478): No. of LM set allocated 16, no. of LM 1
INFO: Initialization of fillpen_t, report:
INFO: Language weight =14.000000
INFO: Word Insertion Penalty =0.200000
INFO: Silence probability =0.800000
INFO: Filler probability =0.100000
INFO:
INFO: dict2pid.c(599): Building PID tables for dictionary
INFO: Initialization of dict2pid_t, report:
INFO: Dict2pid is in composite triphone mode
INFO: 2484 composite states; 1117 composite sseq
INFO:
INFO: kbcore.c(628): Inside kbcore: Verifying models consistency ......
INFO: kbcore.c(650): End of Initialization of Core Models:
INFO: Initialization of beam_t, report:
INFO: Parameters used in Beam Pruning of Viterbi Search:
INFO: Beam=-614114
INFO: PBeam=-614114
INFO: WBeam=-460586 (Skip=0)
INFO: WEndBeam=-614114
INFO: No of CI Phone assumed=40
INFO:
INFO: Initialization of fast_gmm_t, report:
INFO: Parameters used in Fast GMM computation:
INFO: Frame-level: Down Sampling Ratio 1, Conditional Down Sampling? 0,
Distance-based Down Sampling? 0
INFO: GMM-level: CI phone beam -614114. MAX CD 100000
INFO: Gaussian-level: GS map would be used for Gaussian Selection? =1, SVQ
would be used as Gaussian Score? =0 SubVQ Beam -19366
INFO:
INFO: Initialization of pl_t, report:
INFO: Parameters used in phoneme lookahead:
INFO: Phoneme look-ahead type = 0
INFO: Phoneme look-ahead beam size = -614114
INFO: No of CI Phones assumed=40
INFO:
INFO: Initialization of ascr_t, report:
INFO: No. of CI senone =120
INFO: No. of senone = 3120
INFO: No. of composite senone = 2484
INFO: No. of senone sequence = 12335
INFO: No. of composite senone sequence=1117
INFO: Parameters used in phoneme lookahead:
INFO: Phoneme lookahead window = 1
INFO:
INFO: kb.c(308): SEARCH MODE INDEX 3
INFO: srch.c(374): Search Initialization.
INFO: srch_flat_fwd.c(204): Initialization
Thu Jun 14 19:57:48 2012