k, right, with logspec no it cares about ncep, with logspec yes it should not
matter
as for word_align, check log for details, check result transcription in result
folder, try to run word_align from a command line to see what is failed.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I tried with -nfft 512 but it doesnt transform all my .wav files as I have
4579 .wav files. Thats why when I put -nfft as 8192, it transformed all the
4576 files.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
UttID is ignored but file size mismatch between /CLUSTERHOMES/LMS_AUDIO/maas/v
ij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_cle
an_log_test.transcription3120 and /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK34
1_project/ASR_systems/pocketSphinx/train_clean_log/result/train_clean_log.matc
h3120 at ./scripts_pl/decode/word_align.pl line 58.
This is written in the .align file in results. I checked and there is no
.trancription3120 in etc folder
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
UttID is ignored but file size mismatch between /CLUSTERHOMES/LMS_AUDIO/maas
/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_c
lean_log_test.transcription3120 and /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK
341_project/ASR_systems/pocketSphinx/train_clean_log/result/train_clean_log.ma
tch3120 at ./scripts_pl/decode/word_align.pl line 58.
File size mismatch means that some of the files were not decoded properly. You
need to check the size of the both reference and match files (compare number
of lines) and check the decoder log to see why not all utterances were
decoded.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I just observed that the match files being created in the result folder are
empty. Can you please tell if you have an idea, why this is happening? That is
why it is saying, the file size mismatch
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
This error means that your model did not train properly. Transition matrix for
some phone have only zeros. You can check transition matrix with printp
command. The most probable reason for this issue is insufficient data for the
training.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
For this error, the number of training and testing files are sufficient in my
case. I have 4579 files for training which is much more than the an4 corpus.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thanks for your help. I was able to remove the above error. I just trained
again with a few changes.
But the following error still remains while decoding:
Decoding 513 segments starting at 0 (part 1 of 1)
0%
Aligning results to find error rate
word_align.pl failed with error code 65280 at ./scripts_pl/decode/slave.pl
line 173.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Very low recognition accuracy.This might happen if there is a mismatch
in the audio files and the parameters of training, or between the training and
the testing.
To ask for help
If you want to ask for help about training, try to provide the training folder
or at least logdir. We would be glad to help you.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thank you very much for your help.
As I was getting a lot of errors, I started the training of logspec features
from scratch. I have been able to extract them successfully. But when I run
./scripts_pl/RunAll.pl, it gives the following error:
Configuration (e.g. etc/sphinx_train.cfg) not defined
Compilation failed in require at ./scripts_pl/RunAll.pl line 48.
BEGIN failed--compilation aborted at ./scripts_pl/RunAll.pl line 48.
Please help as I don't know what to do.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
k, right, with logspec no it cares about ncep, with logspec yes it should not
matter
as for word_align, check log for details, check result transcription in result
folder, try to run word_align from a command line to see what is failed.
I tried with -nfft 512 but it doesnt transform all my .wav files as I have
4579 .wav files. Thats why when I put -nfft as 8192, it transformed all the
4576 files.
UttID is ignored but file size mismatch between /CLUSTERHOMES/LMS_AUDIO/maas/v
ij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_cle
an_log_test.transcription3120 and /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK34
1_project/ASR_systems/pocketSphinx/train_clean_log/result/train_clean_log.matc
h3120 at ./scripts_pl/decode/word_align.pl line 58.
This is written in the .align file in results. I checked and there is no
.trancription3120 in etc folder
Running just word_align was not of much use.. Do you have any idea how can I
proceed to debug this error..
File size mismatch means that some of the files were not decoded properly. You
need to check the size of the both reference and match files (compare number
of lines) and check the decoder log to see why not all utterances were
decoded.
I just observed that the match files being created in the result folder are
empty. Can you please tell if you have an idea, why this is happening? That is
why it is saying, the file size mismatch
Please check decoder log in logdir for details.
Thanks a lot for ur help. I will try it.
This my logdir. It shows nothing related to the error.
INFO: cmd_ln.c(691): Parsing command line:
/CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx
/train_clean_log/bin/pocketsphinx_batch \
-hmm /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/model_parameters/train_clean_log.cd_semi_1000 \
-lw 10 \
-feat cep \
-beam 1e-80 \
-wbeam 1e-40 \
-dict /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_clean_log.dic \
-lm /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_clean_log.lm \
-wip 0.2 \
-ctl /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_clean_log_test.fileids \
-ctloffset 0 \
-ctlcount 513 \
-cepdir /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/feat/test \
-cepext .lmc \
-hyp /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/result/train_clean_log-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
-argfile
-ascale 20.0 2.000000e+01
-aw 1 1
-backtrace no no
-beam 1e-48 1.000000e-80
-bestpath yes yes
-bestpathlw 9.5 9.500000e+00
-bghist no no
-build_outdirs yes yes
-cepdir /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/feat/test
-cepext .mfc .lmc
-ceplen 13 13
-cmn current current
-cmninit 8.0 8.0
-compallsen no no
-ctl /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_clean_log_test.fileids
-ctlcount -1 513
-ctlincr 1 1
-ctloffset 0 0
-ctm
-debug 0
-dict /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_clean_log.dic
-dictcase no no
-dither no no
-doublebw no no
-ds 1 1
-fdict
-feat 1s_c_d_dd cep
-featparams
-fillprob 1e-8 1.000000e-08
-frate 100 100
-fsg
-fsgctl
-fsgdir
-fsgext
-fsgusealtpron yes yes
-fsgusefiller yes yes
-fwdflat yes yes
-fwdflatbeam 1e-64 1.000000e-64
-fwdflatefwid 4 4
-fwdflatlw 8.5 8.500000e+00
-fwdflatsfwin 25 25
-fwdflatwbeam 7e-29 7.000000e-29
-fwdtree yes yes
-hmm /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/model_parameters/train_clean_log.cd_semi_1000
-hyp /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/result/train_clean_log-1-1.match
-hypseg
-input_endian little little
-jsgf
-kdmaxbbi -1 -1
-kdmaxdepth 0 0
-kdtree
-latsize 5000 5000
-lda
-ldadim 0 0
-lextreedump 0 0
-lifter 0 0
-lm /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_clean_log.lm
-lmctl
-lmname default default
-lmnamectl
-logbase 1.0001 1.000100e+00
-logfn
-logspec no no
-lowerf 133.33334 1.333333e+02
-lpbeam 1e-40 1.000000e-40
-lponlybeam 7e-29 7.000000e-29
-lw 6.5 1.000000e+01
-maxhmmpf -1 -1
-maxnewoov 20 20
-maxwpf -1 -1
-mdef
-mean
-mfclogdir
-min_endfr 0 0
-mixw
-mixwfloor 0.0000001 1.000000e-07
-mllr
-mllrctl
-mllrdir
-mllrext
-mmap yes yes
-nbest 0 0
-nbestdir
-nbestext .hyp .hyp
-ncep 13 13
-nfft 512 512
-nfilt 40 40
-nwpen 1.0 1.000000e+00
-outlatbeam 1e-5 1.000000e-05
-outlatdir
-outlatext .lat .lat
-outlatfmt s3 s3
-pbeam 1e-48 1.000000e-48
-pip 1.0 1.000000e+00
-pl_beam 1e-10 1.000000e-10
-pl_pbeam 1e-5 1.000000e-05
-pl_window 0 0
-rawlogdir
-remove_dc no no
-round_filters yes yes
-samprate 16000 1.600000e+04
-seed -1 -1
-sendump
-senin no no
-senlogdir
-senmgau
-silprob 0.005 5.000000e-03
-smoothspec no no
-svspec
-tmat
-tmatfloor 0.0001 1.000000e-04
-topn 4 4
-topn_beam 0 0
-toprule
-transform legacy legacy
-unit_area yes yes
-upperf 6855.4976 6.855498e+03
-usewdphones no no
-uw 1.0 1.000000e+00
-var
-varfloor 0.0001 1.000000e-04
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wbeam 7e-29 1.000000e-40
-wip 0.65 2.000000e-01
-wlen 0.025625 2.562500e-02
INFO: cmd_ln.c(691): Parsing command line:
\
-alpha 0.97 \
-doublebw no \
-nfilt 24 \
-ncep 13 \
-lowerf 133.33334 \
-upperf 6855.4976 \
-nfft 1024 \
-ceplen 24 \
-logspec yes \
-wlen 0.0256 \
-transform legacy \
-feat 1s_c \
-agc none \
-cmn current \
-varnorm no
Current configuration:
-agc none none
-agcthresh 2.0 2.000000e+00
-alpha 0.97 9.700000e-01
-ceplen 13 24
-cmn current current
-cmninit 8.0 8.0
-dither no no
-doublebw no no
-feat 1s_c_d_dd 1s_c
-frate 100 100
-input_endian little little
-lda
-ldadim 0 0
-lifter 0 0
-logspec no yes
-lowerf 133.33334 1.333333e+02
-ncep 13 13
-nfft 512 1024
-nfilt 40 24
-remove_dc no no
-round_filters yes yes
-samprate 16000 1.600000e+04
-seed -1 -1
-smoothspec 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 2.560000e-02
INFO: acmod.c(242): Parsed model-specific feature parameters from /CLUSTERHOME
S/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean
log/model_parameters/train_clean_log.cd_semi_1000/feat.params
INFO: feat.c(684): Initializing feature stream to type: '1s_c', ceplen=24,
CMN='current', VARNORM='no', AGC='none'
INFO: cmn.c(142): mean= 12.00, mean= 0.0
INFO: mdef.c(520): Reading model definition: /CLUSTERHOMES/LMS_AUDIO/maas/vij/
REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/model_parameters
/train_clean_log.cd_semi_1000/mdef
INFO: bin_mdef.c(173): Allocating 381 * 8 bytes (2 KiB) for CD tree
INFO: tmat.c(205): Reading HMM transition probability matrices: /CLUSTERHOMES/
LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_l
og/model_parameters/train_clean_log.cd_semi_1000/transition_matrices
WARNING: "tmat.c", line 278: Normalization failed for transition matrix 0 from
state 0
WARNING: "tmat.c", line 278: Normalization failed for transition matrix 0 from
state 1
WARNING: "tmat.c", line 278: Normalization failed for transition matrix 0 from
state 2
INFO: acmod.c(117): Attempting to use SCHMM computation module
INFO: ms_gauden.c(198): Reading mixture gaussian parameter: /CLUSTERHOMES/LMS
AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/m
odel_parameters/train_clean_log.cd_semi_1000/means
INFO: ms_gauden.c(292): 1 codebook, 1 feature, size:
INFO: ms_gauden.c(294): 256x24
INFO: ms_gauden.c(198): Reading mixture gaussian parameter: /CLUSTERHOMES/LMS_
AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/m
odel_parameters/train_clean_log.cd_semi_1000/variances
INFO: ms_gauden.c(292): 1 codebook, 1 feature, size:
INFO: ms_gauden.c(294): 256x24
INFO: ms_gauden.c(354): 0 variance values floored
INFO: s2_semi_mgau.c(908): Loading senones from dump file /CLUSTERHOMES/LMS_AU
DIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/mod
el_parameters/train_clean_log.cd_semi_1000/sendump
INFO: s2_semi_mgau.c(932): BEGIN FILE FORMAT DESCRIPTION
INFO: s2_semi_mgau.c(995): Rows: 256, Columns: 438
INFO: s2_semi_mgau.c(1027): Using memory-mapped I/O for senones
INFO: s2_semi_mgau.c(1304): Maximum top-N: 4 Top-N beams: 0
INFO: dict.c(306): Allocating 4114 * 32 bytes (128 KiB) for word entries
INFO: dict.c(321): Reading main dictionary: /CLUSTERHOMES/LMS_AUDIO/maas/vij/R
EMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_clean_l
og.dic
INFO: dict.c(212): Allocated 0 KiB for strings, 0 KiB for phones
INFO: dict.c(324): 15 words read
INFO: dict.c(330): Reading filler dictionary: /CLUSTERHOMES/LMS_AUDIO/maas/vij
/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/model_parameter
s/train_clean_log.cd_semi_1000/noisedict
INFO: dict.c(212): Allocated 0 KiB for strings, 0 KiB for phones
INFO: dict.c(333): 3 words read
INFO: dict2pid.c(396): Building PID tables for dictionary
INFO: dict2pid.c(404): Allocating 21^3 * 2 bytes (18 KiB) for word-initial
triphones
INFO: dict2pid.c(131): Allocated 10752 bytes (10 KiB) for word-final triphones
INFO: dict2pid.c(195): Allocated 10752 bytes (10 KiB) for single-phone word
triphones
INFO: ngram_model_arpa.c(477): ngrams 1=13, 2=22, 3=11
INFO: ngram_model_arpa.c(135): Reading unigrams
INFO: ngram_model_arpa.c(516): 13 = #unigrams created
INFO: ngram_model_arpa.c(195): Reading bigrams
INFO: ngram_model_arpa.c(533): 22 = #bigrams created
INFO: ngram_model_arpa.c(534): 3 = #prob2 entries
INFO: ngram_model_arpa.c(542): 3 = #bo_wt2 entries
INFO: ngram_model_arpa.c(292): Reading trigrams
INFO: ngram_model_arpa.c(555): 11 = #trigrams created
INFO: ngram_model_arpa.c(556): 2 = #prob3 entries
INFO: ngram_search_fwdtree.c(99): 12 unique initial diphones
INFO: ngram_search_fwdtree.c(147): 0 root, 0 non-root channels, 7 single-phone
words
INFO: ngram_search_fwdtree.c(186): Creating search tree
INFO: ngram_search_fwdtree.c(191): before: 0 root, 0 non-root channels, 7
single-phone words
INFO: ngram_search_fwdtree.c(326): after: max nonroot chan increased to 153
INFO: ngram_search_fwdtree.c(338): after: 12 root, 25 non-root channels, 3
single-phone words
INFO: ngram_search_fwdflat.c(156): fwdflat: min_ef_width = 4, max_sf_win = 25
INFO: cmn.c(175): CMN: 10.75 10.36 10.20 10.34 9.80 9.34 9.53 9.29 8.69 8.47
8.59 9.58 9.47 8.64 8.34 8.73 8.31 6.53 5.10 4.45 5.08 5.53 5.25 4.90
INFO: ngram_search_fwdtree.c(1549): 1239 words recognized (4/fr)
INFO: ngram_search_fwdtree.c(1551): 40475 senones evaluated (142/fr)
INFO: ngram_search_fwdtree.c(1553): 16662 channels searched (58/fr), 3384 1st,
7532 last
INFO: ngram_search_fwdtree.c(1557): 1467 words for which last channels
evaluated (5/fr)
INFO: ngram_search_fwdtree.c(1560): 1048 candidate words for entering last
phone (3/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.01 CPU 0.002 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.00 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 13 words
INFO: ngram_search_fwdflat.c(940): 776 words recognized (3/fr)
INFO: ngram_search_fwdflat.c(942): 38658 senones evaluated (135/fr)
INFO: ngram_search_fwdflat.c(944): 18066 channels searched (63/fr)
INFO: ngram_search_fwdflat.c(946): 2220 words searched (7/fr)
INFO: ngram_search_fwdflat.c(948): 1370 word transitions (4/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.00 wall 0.002 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.278INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 107 nodes, 91 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:278:284) = -722576
INFO: ps_lattice.c(1390): Joint P(O,S) = -735820 P(S|O) = -13244
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 1115ZA: 2.85 seconds speech, 0.01 seconds CPU, 0.01
seconds wall
INFO: batch.c(762): 1115ZA: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 9.52 9.25 9.45 8.60 8.38 7.86 7.99 7.72 7.55 7.54 7.30
7.25 6.82 6.11 5.71 5.43 5.84 6.68 6.64 6.24 5.12 4.16 4.14 4.65
INFO: ngram_search_fwdtree.c(1549): 925 words recognized (4/fr)
INFO: ngram_search_fwdtree.c(1551): 32102 senones evaluated (153/fr)
INFO: ngram_search_fwdtree.c(1553): 12480 channels searched (59/fr), 2472 1st,
5818 last
INFO: ngram_search_fwdtree.c(1557): 1100 words for which last channels
evaluated (5/fr)
INFO: ngram_search_fwdtree.c(1560): 969 candidate words for entering last
phone (4/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.00 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 11 words
INFO: ngram_search_fwdflat.c(940): 680 words recognized (3/fr)
INFO: ngram_search_fwdflat.c(942): 31939 senones evaluated (152/fr)
INFO: ngram_search_fwdflat.c(944): 13065 channels searched (62/fr)
INFO: ngram_search_fwdflat.c(946): 1613 words searched (7/fr)
INFO: ngram_search_fwdflat.c(948): 980 word transitions (4/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.00 wall 0.002 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.207INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 77 nodes, 91 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:207:208) = -559470
INFO: ps_lattice.c(1390): Joint P(O,S) = -586987 P(S|O) = -27517
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 1144A: 2.09 seconds speech, 0.01 seconds CPU, 0.01 seconds
wall
INFO: batch.c(762): 1144A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 9.61 9.49 9.33 9.00 8.95 8.57 8.31 8.30 8.39 8.68 8.66
8.61 8.98 9.20 9.34 8.88 8.05 8.17 7.98 7.35 6.56 6.42 6.57 7.24
INFO: ngram_search_fwdtree.c(1549): 1362 words recognized (3/fr)
INFO: ngram_search_fwdtree.c(1551): 52431 senones evaluated (129/fr)
INFO: ngram_search_fwdtree.c(1553): 20262 channels searched (49/fr), 4812 1st,
8265 last
INFO: ngram_search_fwdtree.c(1557): 1714 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 1164 candidate words for entering last
phone (2/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.01 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.01 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 14 words
INFO: ngram_search_fwdflat.c(940): 781 words recognized (2/fr)
INFO: ngram_search_fwdflat.c(942): 45185 senones evaluated (111/fr)
INFO: ngram_search_fwdflat.c(944): 18891 channels searched (46/fr)
INFO: ngram_search_fwdflat.c(946): 2600 words searched (6/fr)
INFO: ngram_search_fwdflat.c(948): 1368 word transitions (3/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.01 CPU 0.001 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.01 wall 0.001 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.405INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 114 nodes, 62 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:405:406) = -1019182
INFO: ps_lattice.c(1390): Joint P(O,S) = -1026316 P(S|O) = -7134
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 1167489A: 4.07 seconds speech, 0.01 seconds CPU, 0.01
seconds wall
INFO: batch.c(762): 1167489A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 8.38 8.08 8.16 8.03 7.19 6.87 6.76 6.35 6.21 6.30 6.75
8.17 8.32 7.07 6.41 6.53 6.02 5.62 4.52 4.61 5.68 5.85 5.41 5.33
INFO: ngram_search_fwdtree.c(1549): 599 words recognized (3/fr)
INFO: ngram_search_fwdtree.c(1551): 22219 senones evaluated (124/fr)
INFO: ngram_search_fwdtree.c(1553): 8244 channels searched (46/fr), 2100 1st,
3447 last
INFO: ngram_search_fwdtree.c(1557): 738 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 475 candidate words for entering last
phone (2/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.00 wall 0.001 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 9 words
INFO: ngram_search_fwdflat.c(940): 322 words recognized (2/fr)
INFO: ngram_search_fwdflat.c(942): 19772 senones evaluated (110/fr)
INFO: ngram_search_fwdflat.c(944): 7897 channels searched (44/fr)
INFO: ngram_search_fwdflat.c(946): 1116 words searched (6/fr)
INFO: ngram_search_fwdflat.c(948): 578 word transitions (3/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.002 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.00 wall 0.001 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.169INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 31 nodes, 18 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:169:177) = -459619
INFO: ps_lattice.c(1390): Joint P(O,S) = -461349 P(S|O) = -1730
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 118A: 1.78 seconds speech, 0.00 seconds CPU, 0.01 seconds
wall
INFO: batch.c(762): 118A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 9.41 9.90 9.94 9.50 9.04 8.86 8.66 8.22 7.63 7.25 7.33
7.10 7.02 7.28 7.25 7.14 7.37 7.26 6.75 5.90 5.80 5.33 5.81 6.12
INFO: ngram_search_fwdtree.c(1549): 1265 words recognized (5/fr)
INFO: ngram_search_fwdtree.c(1551): 39317 senones evaluated (168/fr)
INFO: ngram_search_fwdtree.c(1553): 15715 channels searched (67/fr), 2760 1st,
7640 last
INFO: ngram_search_fwdtree.c(1557): 1369 words for which last channels
evaluated (5/fr)
INFO: ngram_search_fwdtree.c(1560): 1318 candidate words for entering last
phone (5/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.00 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 11 words
INFO: ngram_search_fwdflat.c(940): 957 words recognized (4/fr)
INFO: ngram_search_fwdflat.c(942): 36454 senones evaluated (156/fr)
INFO: ngram_search_fwdflat.c(944): 15521 channels searched (66/fr)
INFO: ngram_search_fwdflat.c(946): 1923 words searched (8/fr)
INFO: ngram_search_fwdflat.c(948): 959 word transitions (4/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.00 wall 0.002 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.231INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 102 nodes, 151 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:231:232) = -608199
INFO: ps_lattice.c(1390): Joint P(O,S) = -650315 P(S|O) = -42116
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 11O92A: 2.33 seconds speech, 0.01 seconds CPU, 0.01
seconds wall
INFO: batch.c(762): 11O92A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 8.15 8.23 8.03 7.60 7.42 7.15 6.56 6.03 5.74 5.61 5.66
6.29 7.00 6.78 6.64 6.55 6.78 6.93 6.45 5.56 5.31 5.04 4.94 5.40
INFO: ngram_search_fwdtree.c(1549): 1544 words recognized (4/fr)
INFO: ngram_search_fwdtree.c(1551): 53770 senones evaluated (135/fr)
INFO: ngram_search_fwdtree.c(1553): 19863 channels searched (50/fr), 4716 1st,
8274 last
INFO: ngram_search_fwdtree.c(1557): 1845 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 1386 candidate words for entering last
phone (3/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.01 wall 0.001 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 11 words
INFO: ngram_search_fwdflat.c(940): 985 words recognized (2/fr)
INFO: ngram_search_fwdflat.c(942): 47227 senones evaluated (119/fr)
INFO: ngram_search_fwdflat.c(944): 17732 channels searched (44/fr)
INFO: ngram_search_fwdflat.c(946): 2431 words searched (6/fr)
INFO: ngram_search_fwdflat.c(948): 1077 word transitions (2/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.01 wall 0.001 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.394INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 79 nodes, 61 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:394:395) = -993930
INFO: ps_lattice.c(1390): Joint P(O,S) = -1000472 P(S|O) = -6542
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 12184A: 3.96 seconds speech, 0.01 seconds CPU, 0.01
seconds wall
INFO: batch.c(762): 12184A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 9.83 8.93 9.12 9.10 7.94 7.89 7.70 7.47 7.59 7.94 8.20
8.58 8.79 8.80 8.89 8.44 7.88 8.02 7.45 6.12 5.54 5.79 6.42 6.60
INFO: ngram_search_fwdtree.c(1549): 1478 words recognized (4/fr)
INFO: ngram_search_fwdtree.c(1551): 56244 senones evaluated (152/fr)
INFO: ngram_search_fwdtree.c(1553): 21155 channels searched (57/fr), 4404 1st,
9373 last
INFO: ngram_search_fwdtree.c(1557): 1822 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 1379 candidate words for entering last
phone (3/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.01 CPU 0.002 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.01 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 14 words
INFO: ngram_search_fwdflat.c(940): 930 words recognized (3/fr)
INFO: ngram_search_fwdflat.c(942): 52225 senones evaluated (141/fr)
INFO: ngram_search_fwdflat.c(944): 20686 channels searched (55/fr)
INFO: ngram_search_fwdflat.c(946): 2745 words searched (7/fr)
INFO: ngram_search_fwdflat.c(948): 1499 word transitions (4/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.01 wall 0.002 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.365INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 88 nodes, 102 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:365:369) = -929175
INFO: ps_lattice.c(1390): Joint P(O,S) = -936665 P(S|O) = -7490
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 1231374A: 3.70 seconds speech, 0.01 seconds CPU, 0.01
seconds wall
INFO: batch.c(762): 1231374A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 8.08 8.09 8.14 7.64 7.26 7.05 7.06 6.92 6.41 6.05 6.33
6.65 6.59 6.62 6.86 6.96 5.73 4.29 3.84 3.77 4.28 4.66 4.47 4.47
INFO: ngram_search_fwdtree.c(1549): 608 words recognized (3/fr)
INFO: ngram_search_fwdtree.c(1551): 24863 senones evaluated (135/fr)
INFO: ngram_search_fwdtree.c(1553): 8894 channels searched (48/fr), 2160 1st,
3525 last
INFO: ngram_search_fwdtree.c(1557): 783 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 584 candidate words for entering last
phone (3/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.00 wall 0.001 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 10 words
INFO: ngram_search_fwdflat.c(940): 399 words recognized (2/fr)
INFO: ngram_search_fwdflat.c(942): 24705 senones evaluated (134/fr)
INFO: ngram_search_fwdflat.c(944): 8707 channels searched (47/fr)
INFO: ngram_search_fwdflat.c(946): 1227 words searched (6/fr)
INFO: ngram_search_fwdflat.c(948): 667 word transitions (3/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.002 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.00 wall 0.001 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.180INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 55 nodes, 31 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:180:182) = -458097
INFO: ps_lattice.c(1390): Joint P(O,S) = -458994 P(S|O) = -897
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 124A: 1.83 seconds speech, 0.00 seconds CPU, 0.01 seconds
wall
INFO: batch.c(762): 124A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 8.26 8.20 8.58 7.87 7.31 7.24 7.17 7.11 7.15 7.31 7.61
7.96 8.30 8.47 8.35 8.21 8.37 7.96 7.33 6.87 7.38 7.78 7.91 7.96
INFO: ngram_search_fwdtree.c(1549): 931 words recognized (3/fr)
INFO: ngram_search_fwdtree.c(1551): 36012 senones evaluated (133/fr)
INFO: ngram_search_fwdtree.c(1553): 13151 channels searched (48/fr), 3192 1st,
5077 last
INFO: ngram_search_fwdtree.c(1557): 1084 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 737 candidate words for entering last
phone (2/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.00 wall 0.001 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 11 words
INFO: ngram_search_fwdflat.c(940): 540 words recognized (2/fr)
INFO: ngram_search_fwdflat.c(942): 30704 senones evaluated (114/fr)
INFO: ngram_search_fwdflat.c(944): 11339 channels searched (41/fr)
INFO: ngram_search_fwdflat.c(946): 1546 words searched (5/fr)
INFO: ngram_search_fwdflat.c(948): 823 word transitions (3/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.00 wall 0.001 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.266INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 75 nodes, 62 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:266:268) = -671582
INFO: ps_lattice.c(1390): Joint P(O,S) = -672593 P(S|O) = -1011
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 1286A: 2.69 seconds speech, 0.01 seconds CPU, 0.01 seconds
wall
INFO: batch.c(762): 1286A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 9.29 8.66 8.59 8.39 8.54 8.26 7.80 7.34 7.21 6.72 6.30
6.05 6.14 6.65 6.60 5.80 6.51 6.63 5.64 4.64 4.69 5.81 6.38 6.25
INFO: ngram_search_fwdtree.c(1549): 812 words recognized (5/fr)
INFO: ngram_search_fwdtree.c(1551): 25503 senones evaluated (150/fr)
INFO: ngram_search_fwdtree.c(1553): 10773 channels searched (63/fr), 1992 1st,
5285 last
INFO: ngram_search_fwdtree.c(1557): 951 words for which last channels
evaluated (5/fr)
INFO: ngram_search_fwdtree.c(1560): 776 candidate words for entering last
phone (4/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.002 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.00 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 13 words
INFO: ngram_search_fwdflat.c(940): 515 words recognized (3/fr)
INFO: ngram_search_fwdflat.c(942): 28224 senones evaluated (166/fr)
INFO: ngram_search_fwdflat.c(944): 13875 channels searched (81/fr)
INFO: ngram_search_fwdflat.c(946): 1715 words searched (10/fr)
INFO: ngram_search_fwdflat.c(948): 893 word transitions (5/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.002 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.00 wall 0.002 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.167INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 69 nodes, 45 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:167:168) = -440122
INFO: ps_lattice.c(1390): Joint P(O,S) = -445066 P(S|O) = -4944
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 12OA: 1.69 seconds speech, 0.01 seconds CPU, 0.01 seconds
wall
INFO: batch.c(762): 12OA: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 8.89 8.98 8.68 8.19 7.54 7.07 7.15 7.28 7.50 8.11 8.26
8.15 8.33 8.05 8.19 7.93 7.48 7.43 7.13 6.79 6.24 5.66 6.31 6.86
INFO: ngram_search_fwdtree.c(1549): 1211 words recognized (4/fr)
INFO: ngram_search_fwdtree.c(1551): 44349 senones evaluated (145/fr)
INFO: ngram_search_fwdtree.c(1553): 16852 channels searched (55/fr), 3624 1st,
6910 last
INFO: ngram_search_fwdtree.c(1557): 1462 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 1306 candidate words for entering last
phone (4/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.00 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 14 words
INFO: ngram_search_fwdflat.c(940): 796 words recognized (3/fr)
INFO: ngram_search_fwdflat.c(942): 37071 senones evaluated (121/fr)
INFO: ngram_search_fwdflat.c(944): 14489 channels searched (47/fr)
INFO: ngram_search_fwdflat.c(946): 2025 words searched (6/fr)
INFO: ngram_search_fwdflat.c(948): 1041 word transitions (3/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.002 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.00 wall 0.001 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.303INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 94 nodes, 91 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:303:304) = -759359
INFO: ps_lattice.c(1390): Joint P(O,S) = -768740 P(S|O) = -9381
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 1332ZA: 3.05 seconds speech, 0.01 seconds CPU, 0.01
seconds wall
INFO: batch.c(762): 1332ZA: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 8.79 8.18 8.35 8.35 8.11 7.91 7.54 7.34 7.27 7.97 8.84
9.21 8.97 9.03 9.21 8.57 8.08 8.75 8.10 6.72 7.29 8.23 8.78 8.81
INFO: ngram_search_fwdtree.c(1549): 1419 words recognized (4/fr)
INFO: ngram_search_fwdtree.c(1551): 54754 senones evaluated (145/fr)
INFO: ngram_search_fwdtree.c(1553): 20918 channels searched (55/fr), 4464 1st,
8788 last
INFO: ngram_search_fwdtree.c(1557): 1748 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 1369 candidate words for entering last
phone (3/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.01 CPU 0.002 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.01 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 14 words
INFO: ngram_search_fwdflat.c(940): 784 words recognized (2/fr)
INFO: ngram_search_fwdflat.c(942): 59242 senones evaluated (157/fr)
INFO: ngram_search_fwdflat.c(944): 23763 channels searched (62/fr)
INFO: ngram_search_fwdflat.c(946): 3017 words searched (7/fr)
INFO: ngram_search_fwdflat.c(948): 1598 word transitions (4/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.01 CPU 0.002 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.01 wall 0.002 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.369INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 91 nodes, 65 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:369:376) = -966800
INFO: ps_lattice.c(1390): Joint P(O,S) = -973126 P(S|O) = -6326
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 1369O76A: 3.77 seconds speech, 0.01 seconds CPU, 0.01
seconds wall
INFO: batch.c(762): 1369O76A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 9.29 9.14 8.92 8.14 7.43 6.85 6.71 6.75 6.97 7.19 7.29
7.93 8.39 7.96 7.55 7.85 7.70 6.88 6.08 6.40 7.09 7.04 6.34 6.73
INFO: ngram_search_fwdtree.c(1549): 1462 words recognized (4/fr)
INFO: ngram_search_fwdtree.c(1551): 56496 senones evaluated (142/fr)
INFO: ngram_search_fwdtree.c(1553): 21279 channels searched (53/fr), 4692 1st,
8915 last
INFO: ngram_search_fwdtree.c(1557): 1804 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 1271 candidate words for entering last
phone (3/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.01 CPU 0.002 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.01 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 14 words
INFO: ngram_search_fwdflat.c(940): 833 words recognized (2/fr)
INFO: ngram_search_fwdflat.c(942): 53852 senones evaluated (135/fr)
INFO: ngram_search_fwdflat.c(944): 21391 channels searched (53/fr)
INFO: ngram_search_fwdflat.c(946): 2893 words searched (7/fr)
INFO: ngram_search_fwdflat.c(948): 1677 word transitions (4/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.01 CPU 0.002 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.01 wall 0.001 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.392INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 94 nodes, 58 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:392:396) = -988594
INFO: ps_lattice.c(1390): Joint P(O,S) = -992576 P(S|O) = -3982
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 1384377A: 3.97 seconds speech, 0.01 seconds CPU, 0.01
seconds wall
INFO: batch.c(762): 1384377A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 8.61 9.04 8.87 7.75 7.32 7.35 7.45 7.25 7.38 7.53 7.91
8.36 7.99 7.58 7.25 7.07 6.43 5.98 5.84 5.60 5.93 6.03 6.17 6.65
INFO: ngram_search_fwdtree.c(1549): 502 words recognized (4/fr)
INFO: ngram_search_fwdtree.c(1551): 19147 senones evaluated (146/fr)
INFO: ngram_search_fwdtree.c(1553): 7058 channels searched (53/fr), 1524 1st,
2779 last
INFO: ngram_search_fwdtree.c(1557): 609 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 548 candidate words for entering last
phone (4/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.002 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.00 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 9 words
INFO: ngram_search_fwdflat.c(940): 337 words recognized (3/fr)
INFO: ngram_search_fwdflat.c(942): 17198 senones evaluated (131/fr)
INFO: ngram_search_fwdflat.c(944): 6530 channels searched (49/fr)
INFO: ngram_search_fwdflat.c(946): 935 words searched (7/fr)
INFO: ngram_search_fwdflat.c(948): 463 word transitions (3/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.002 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.00 wall 0.001 xRT
INFO: ngram_search.c(1201): not found in last frame, using three.129
instead
INFO: ngram_search.c(1253): lattice start node
.0 end node three.53:405:408) = -899718INFO: ngram_search.c(1281): Eliminated 10 nodes before end node
INFO: ngram_search.c(1386): Lattice has 54 nodes, 24 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(three:53:129) = -301635
INFO: ps_lattice.c(1390): Joint P(O,S) = -304717 P(S|O) = -3082
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 13A: 1.30 seconds speech, 0.00 seconds CPU, 0.00 seconds
wall
INFO: batch.c(762): 13A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 8.79 9.14 8.85 8.20 7.87 7.75 7.61 7.34 7.48 7.82 8.50
8.97 8.92 8.38 8.27 7.88 7.76 8.72 8.70 8.04 7.64 7.51 7.99 8.52
INFO: ngram_search_fwdtree.c(1549): 1224 words recognized (3/fr)
INFO: ngram_search_fwdtree.c(1551): 48694 senones evaluated (119/fr)
INFO: ngram_search_fwdtree.c(1553): 18219 channels searched (44/fr), 4869 1st,
6555 last
INFO: ngram_search_fwdtree.c(1557): 1508 words for which last channels
evaluated (3/fr)
INFO: ngram_search_fwdtree.c(1560): 1004 candidate words for entering last
phone (2/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.01 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.01 wall 0.001 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 11 words
INFO: ngram_search_fwdflat.c(940): 761 words recognized (2/fr)
INFO: ngram_search_fwdflat.c(942): 39292 senones evaluated (96/fr)
INFO: ngram_search_fwdflat.c(944): 15097 channels searched (36/fr)
INFO: ngram_search_fwdflat.c(946): 2069 words searched (5/fr)
INFO: ngram_search_fwdflat.c(948): 974 word transitions (2/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.01 wall 0.001 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.405INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 120 nodes, 83 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(
INFO: ps_lattice.c(1390): Joint P(O,S) = -901248 P(S|O) = -1530
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 14266A: 4.09 seconds speech, 0.01 seconds CPU, 0.01
seconds wall
INFO: batch.c(762): 14266A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 10.12 9.91 10.01 9.54 9.03 8.59 8.54 9.01 9.36 9.58
9.68 9.71 10.24 10.29 9.68 9.78 9.76 9.37 8.52 7.73 7.07 7.32 8.20 8.37
INFO: ngram_search_fwdtree.c(1549): 1276 words recognized (4/fr)
INFO: ngram_search_fwdtree.c(1551): 49958 senones evaluated (138/fr)
INFO: ngram_search_fwdtree.c(1553): 18629 channels searched (51/fr), 4308 1st,
7392 last
INFO: ngram_search_fwdtree.c(1557): 1580 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 1136 candidate words for entering last
phone (3/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.01 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 14 words
INFO: ngram_search_fwdflat.c(940): 754 words recognized (2/fr)
INFO: ngram_search_fwdflat.c(942): 43820 senones evaluated (121/fr)
INFO: ngram_search_fwdflat.c(944): 16961 channels searched (46/fr)
INFO: ngram_search_fwdflat.c(946): 2276 words searched (6/fr)
INFO: ngram_search_fwdflat.c(948): 1356 word transitions (3/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.01 CPU 0.002 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.01 wall 0.001 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.354INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
INFO: ngram_search.c(1386): Lattice has 88 nodes, 49 links
INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(:354:361) = -898854
INFO: ps_lattice.c(1390): Joint P(O,S) = -901696 P(S|O) = -2842
INFO: ngram_search.c(875): bestpath 0.00 CPU 0.000 xRT
INFO: ngram_search.c(878): bestpath 0.00 wall 0.000 xRT
INFO: batch.c(760): 14552Z3A: 3.62 seconds speech, 0.01 seconds CPU, 0.01
seconds wall
INFO: batch.c(762): 14552Z3A: 0.00 xRT (CPU), 0.00 xRT (elapsed)
INFO: cmn.c(175): CMN: 9.48 9.68 9.47 9.62 9.20 8.40 7.85 7.71 8.08 8.49 8.63
8.39 8.18 8.22 8.39 8.53 7.78 6.53 5.86 5.24 5.14 5.15 5.21 5.08
INFO: ngram_search_fwdtree.c(1549): 1490 words recognized (4/fr)
INFO: ngram_search_fwdtree.c(1551): 56036 senones evaluated (150/fr)
INFO: ngram_search_fwdtree.c(1553): 21236 channels searched (56/fr), 4440 1st,
9044 last
INFO: ngram_search_fwdtree.c(1557): 1822 words for which last channels
evaluated (4/fr)
INFO: ngram_search_fwdtree.c(1560): 1319 candidate words for entering last
phone (3/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.00 CPU 0.001 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 0.01 wall 0.002 xRT
INFO: ngram_search_fwdflat.c(305): Utterance vocabulary contains 13 words
INFO: ngram_search_fwdflat.c(940): 852 words recognized (2/fr)
INFO: ngram_search_fwdflat.c(942): 52251 senones evaluated (140/fr)
INFO: ngram_search_fwdflat.c(944): 21067 channels searched (56/fr)
INFO: ngram_search_fwdflat.c(946): 2806 words searched (7/fr)
INFO: ngram_search_fwdflat.c(948): 1525 word transitions (4/fr)
INFO: ngram_search_fwdflat.c(951): fwdflat 0.01 CPU 0.002 xRT
INFO: ngram_search_fwdflat.c(954): fwdflat 0.01 wall 0.002 xRT
INFO: ngram_search.c(1253): lattice start node
.0 end node.370INFO: ngram_search.c(1281): Eliminated 0 nodes before end node
pocketsphinx_batch: ngram_search.c:707: ngram_compute_seg_score: Assertion
`start_score > ((int)0xE0000000)' failed.
Wed Jun 27 16:34:14 2012
This error means that your model did not train properly. Transition matrix for
some phone have only zeros. You can check transition matrix with printp
command. The most probable reason for this issue is insufficient data for the
training.
It's also recommended to remove CMN normalization because it doesn't make much
sense for logspec.
pocketsphinx_batch: ngram_search.c:707: ngram_compute_seg_score: Assertion
`start_score > ((int)0xE0000000)' failed.
For this error, the number of training and testing files are sufficient in my
case. I have 4579 files for training which is much more than the an4 corpus.
Thanks for your help. I was able to remove the above error. I just trained
again with a few changes.
But the following error still remains while decoding:
Decoding 513 segments starting at 0 (part 1 of 1)
0%
Aligning results to find error rate
word_align.pl failed with error code 65280 at ./scripts_pl/decode/slave.pl
line 173.
And the empty .match files in the results folder still remain.
Check the decoder log in logdir again, it should contain all the details
Share your training folder if you need more detailed help
For more information see the FAQ
http://cmusphinx.sourceforge.net/wiki/faq
What exactly of the training folder so I have to share?
And the decode log does not show any error or warning
Current configuration:
-adchdr 0 0
-adcin no no
-agc none none
-agcthresh 2.0 2.000000e+00
-alpha 0.97 9.700000e-01
-argfile
-ascale 20.0 2.000000e+01
-aw 1 1
-backtrace no no
-beam 1e-48 1.000000e-80
-bestpath yes yes
-bestpathlw 9.5 9.500000e+00
-bghist no no
-build_outdirs yes yes
-cepdir /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/feat/test
-cepext .mfc .lmc
-ceplen 13 13
-cmn current current
-cmninit 8.0 8.0
-compallsen no no
-ctl /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_clean_log_test.fileids
-ctlcount -1 513
-ctlincr 1 1
-ctloffset 0 0
-ctm
-debug 0
-dict /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_clean_log.dic
-dictcase no no
-dither no no
-doublebw no no
-ds 1 1
-fdict
-feat 1s_c_d_dd cep
-featparams
-fillprob 1e-8 1.000000e-08
-frate 100 100
-fsg
-fsgctl
-fsgdir
-fsgext
-fsgusealtpron yes yes
-fsgusefiller yes yes
-fwdflat yes yes
-fwdflatbeam 1e-64 1.000000e-64
-fwdflatefwid 4 4
-fwdflatlw 8.5 8.500000e+00
-fwdflatsfwin 25 25
-fwdflatwbeam 7e-29 7.000000e-29
-fwdtree yes yes
-hmm /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/model_parameters/train_clean_log.cd_semi_1000
-hyp /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/result/train_clean_log-1-1.match
-hypseg
-input_endian little little
-jsgf
-kdmaxbbi -1 -1
-kdmaxdepth 0 0
-kdtree
-latsize 5000 5000
-lda
-ldadim 0 0
-lextreedump 0 0
-lifter 0 0
-lm /CLUSTERHOMES/LMS_AUDIO/maas/vij/REMOS_HTK341_project/ASR_systems/pocketSphinx/train_clean_log/etc/train_clean_log.lm.DMP
-lmctl
-lmname default default
-lmnamectl
-logbase 1.0001 1.000100e+00
-logfn
-logspec no no
-lowerf 133.33334 1.333333e+02
-lpbeam 1e-40 1.000000e-40
-lponlybeam 7e-29 7.000000e-29
-lw 6.5 1.000000e+01
-maxhmmpf -1 -1
-maxnewoov 20 20
-maxwpf -1 -1
-mdef
-mean
-mfclogdir
-min_endfr 0 0
-mixw
-mixwfloor 0.0000001 1.000000e-07
-mllr
-mllrctl
-mllrdir
-mllrext
-mmap yes yes
-nbest 0 0
-nbestdir
-nbestext .hyp .hyp
-ncep 13 13
-nfft 512 512
-nfilt 40 40
-nwpen 1.0 1.000000e+00
-outlatbeam 1e-5 1.000000e-05
-outlatdir
-outlatext .lat .lat
-outlatfmt s3 s3
-pbeam 1e-48 1.000000e-48
-pip 1.0 1.000000e+00
-pl_beam 1e-10 1.000000e-10
-pl_pbeam 1e-5 1.000000e-05
-pl_window 0 0
-rawlogdir
-remove_dc no no
-round_filters yes yes
-samprate 16000 1.600000e+04
-seed -1 -1
-sendump
-senin no no
-senlogdir
-senmgau
-silprob 0.005 5.000000e-03
-smoothspec no no
-svspec
-tmat
-tmatfloor 0.0001 1.000000e-04
-topn 4 4
-topn_beam 0 0
-toprule
-transform legacy legacy
-unit_area yes yes
-upperf 6855.4976 6.855498e+03
-usewdphones no no
-uw 1.0 1.000000e+00
-var
-varfloor 0.0001 1.000000e-04
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wbeam 7e-29 1.000000e-40
-wip 0.65 2.000000e-01
-wlen 0.025625 2.562500e-02
INFO: cmd_ln.c(691): Parsing command line:
\
-alpha 0.97 \
-doublebw no \
-nfilt 24 \
-ncep 13 \
-lowerf 133.33334 \
-upperf 6855.4976 \
-nfft 8192 \
-ceplen 24 \
-logspec yes \
-wlen 0.0256 \
-transform legacy \
-feat 1s_c \
-agc none \
-cmn none \
-varnorm no
Current configuration:
-agc none none
-agcthresh 2.0 2.000000e+00
-alpha 0.97 9.700000e-01
-ceplen 13 24
-cmn current none
-cmninit 8.0 8.0
-dither no no
-doublebw no no
-feat 1s_c_d_dd 1s_c
-frate 100 100
-input_endian little little
-lda
-ldadim 0 0
-lifter 0 0
-logspec no yes
-lowerf 133.33334 1.333333e+02
-ncep 13 13
-nfft 512 8192
-nfilt 40 24
-remove_dc no no
-round_filters yes yes
-samprate 16000 1.600000e+04
-seed -1 -1
-smoothspec 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 2.560000e-02
After decoding this is the result shown. Do you have any idea why would such a
result come?
MODULE: DECODE Decoding using models previously trained
Decoding 513 segments starting at 0 (part 1 of 1)
0%
Aligning results to find error rate
SENTENCE ERROR: 100.0% (513/513) WORD ERROR RATE: 91.3% (2228/2439)
Please see the troubleshooting section in the acoustic model training
tutorial:
http://cmusphinx.sourceforge.net/wiki/tutorialam#troubleshooting
To ask for help
If you want to ask for help about training, try to provide the training folder
or at least logdir. We would be glad to help you.
Thank you very much for your help.
As I was getting a lot of errors, I started the training of logspec features
from scratch. I have been able to extract them successfully. But when I run
./scripts_pl/RunAll.pl, it gives the following error:
Configuration (e.g. etc/sphinx_train.cfg) not defined
Compilation failed in require at ./scripts_pl/RunAll.pl line 48.
BEGIN failed--compilation aborted at ./scripts_pl/RunAll.pl line 48.
Please help as I don't know what to do.
The sphinx_train.cfg file is there in etc
I got it. Thanks any ways.