bejimed
July 2nd, 2010, 11:55 AM
I had the ubuntu 10.04 and iwant to install HTK 3
I run succesfly ./configure , make all , make install but I had a problem testing the HTKDemo like showed below :
Number of owners = 1
SegLab : S
maxIter : 10
epsilon : 0.000100
minSeg : 3
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
16 Observation Sequences Loaded
Starting Estimation Process
Iteration 1: Average LogP = -804.22485
Iteration 2: Average LogP = -786.82153 Change = 17.40332
Iteration 3: Average LogP = -786.21497 Change = 0.60657
Iteration 4: Average LogP = -786.21503 Change = -0.00006
Estimation converged at iteration 5
Output written to directory hmms/hmm.0
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HInit for HMM C
HInit -A -i 10 -L labels/bcplabs/mon -l C -o C -C toolconfs/hinit.conf -D -M hmms/hmm.0 -T 1 proto/C data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Initialising HMM proto/C . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : C
maxIter : 10
epsilon : 0.000100
minSeg : 3
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
90 Observation Sequences Loaded
Starting Estimation Process
Iteration 1: Average LogP = -531.87134
Iteration 2: Average LogP = -527.31677 Change = 4.55458
Iteration 3: Average LogP = -526.60242 Change = 0.71439
Iteration 4: Average LogP = -526.32343 Change = 0.27898
Iteration 5: Average LogP = -526.26605 Change = 0.05737
Iteration 6: Average LogP = -526.24451 Change = 0.02152
Iteration 7: Average LogP = -526.23560 Change = 0.00892
Iteration 8: Average LogP = -526.23364 Change = 0.00196
Iteration 9: Average LogP = -526.22852 Change = 0.00513
Iteration 10: Average LogP = -526.22852 Change = 0.00000
Estimation converged at iteration 11
Output written to directory hmms/hmm.0
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HInit for HMM V
HInit -A -i 10 -L labels/bcplabs/mon -l V -o V -C toolconfs/hinit.conf -D -M hmms/hmm.0 -T 1 proto/V data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Initialising HMM proto/V . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : V
maxIter : 10
epsilon : 0.000100
minSeg : 3
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
68 Observation Sequences Loaded
Starting Estimation Process
Iteration 1: Average LogP = -599.68671
Iteration 2: Average LogP = -592.98328 Change = 6.70342
Iteration 3: Average LogP = -592.21442 Change = 0.76883
Iteration 4: Average LogP = -591.48163 Change = 0.73280
Iteration 5: Average LogP = -591.14313 Change = 0.33848
Iteration 6: Average LogP = -591.08777 Change = 0.05535
Iteration 7: Average LogP = -591.05005 Change = 0.03773
Iteration 8: Average LogP = -590.98999 Change = 0.06004
Iteration 9: Average LogP = -590.98560 Change = 0.00441
Iteration 10: Average LogP = -590.98558 Change = 0.00001
Estimation converged at iteration 11
Output written to directory hmms/hmm.0
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HInit for HMM N
HInit -A -i 10 -L labels/bcplabs/mon -l N -o N -C toolconfs/hinit.conf -D -M hmms/hmm.0 -T 1 proto/N data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Initialising HMM proto/N . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : N
maxIter : 10
epsilon : 0.000100
minSeg : 3
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
16 Observation Sequences Loaded
Starting Estimation Process
Iteration 1: Average LogP = -398.18735
Iteration 2: Average LogP = -395.48718 Change = 2.70016
Iteration 3: Average LogP = -395.26947 Change = 0.21771
Iteration 4: Average LogP = -395.17661 Change = 0.09286
Iteration 5: Average LogP = -395.17661 Change = 0.00000
Estimation converged at iteration 6
Output written to directory hmms/hmm.0
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HInit for HMM L
HInit -A -i 10 -L labels/bcplabs/mon -l L -o L -C toolconfs/hinit.conf -D -M hmms/hmm.0 -T 1 proto/L data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Initialising HMM proto/L . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : L
maxIter : 10
epsilon : 0.000100
minSeg : 3
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
25 Observation Sequences Loaded
Starting Estimation Process
Iteration 1: Average LogP = -424.39636
Iteration 2: Average LogP = -420.31281 Change = 4.08355
Iteration 3: Average LogP = -419.43475 Change = 0.87804
Iteration 4: Average LogP = -419.35437 Change = 0.08038
Iteration 5: Average LogP = -419.32385 Change = 0.03050
Iteration 6: Average LogP = -419.32387 Change = -0.00001
Estimation converged at iteration 7
Output written to directory hmms/hmm.0
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HRest for HMM S
HRest -A -u tmvw -w 3 -v 0.05 -i 10 -L labels/bcplabs/mon -l S -C toolconfs/hrest.conf -D -M hmms/hmm.1 -T 1 hmms/hmm.0/S data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Reestimating HMM hmms/hmm.0/S . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : S
MaxIter : 10
Epsilon : 0.000100
Updating : Transitions Means Variances
- system is PLAIN
16 Examples loaded, Max length = 20, Min length = 13
Ave LogProb at iter 1 = -785.97754 using 16 examples
Ave LogProb at iter 2 = -799.75970 using 16 examples change = -13.78217
Ave LogProb at iter 3 = -799.59235 using 16 examples change = 0.16736
Ave LogProb at iter 4 = -799.49408 using 16 examples change = 0.09827
Ave LogProb at iter 5 = -799.43396 using 16 examples change = 0.06012
Ave LogProb at iter 6 = -799.36908 using 16 examples change = 0.06488
Ave LogProb at iter 7 = -799.31201 using 16 examples change = 0.05707
Ave LogProb at iter 8 = -799.28625 using 16 examples change = 0.02576
Ave LogProb at iter 9 = -799.26813 using 16 examples change = 0.01813
Ave LogProb at iter 10 = -798.93494 using 16 examples change = 0.33319
Estimation aborted at iteration 10
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HRest for HMM C
HRest -A -u tmvw -w 3 -v 0.05 -i 10 -L labels/bcplabs/mon -l C -C toolconfs/hrest.conf -D -M hmms/hmm.1 -T 1 hmms/hmm.0/C data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Reestimating HMM hmms/hmm.0/C . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : C
MaxIter : 10
Epsilon : 0.000100
Updating : Transitions Means Variances
- system is PLAIN
90 Examples loaded, Max length = 25, Min length = 3
Ave LogProb at iter 1 = -525.97860 using 90 examples
Ave LogProb at iter 2 = -530.95642 using 90 examples change = -4.97785
Ave LogProb at iter 3 = -530.95356 using 90 examples change = 0.00286
Ave LogProb at iter 4 = -530.95312 using 90 examples change = 0.00043
Ave LogProb at iter 5 = -530.95304 using 90 examples change = 0.00009
Estimation converged at iteration 5
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HRest for HMM V
HRest -A -u tmvw -w 3 -v 0.05 -i 10 -L labels/bcplabs/mon -l V -C toolconfs/hrest.conf -D -M hmms/hmm.1 -T 1 hmms/hmm.0/V data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Reestimating HMM hmms/hmm.0/V . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : V
MaxIter : 10
Epsilon : 0.000100
Updating : Transitions Means Variances
- system is PLAIN
68 Examples loaded, Max length = 28, Min length = 3
Ave LogProb at iter 1 = -590.79446 using 68 examples
Ave LogProb at iter 2 = -600.09358 using 68 examples change = -9.29914
Ave LogProb at iter 3 = -600.08226 using 68 examples change = 0.01131
Ave LogProb at iter 4 = -600.07560 using 68 examples change = 0.00668
Ave LogProb at iter 5 = -600.07071 using 68 examples change = 0.00491
Ave LogProb at iter 6 = -600.06859 using 68 examples change = 0.00215
Ave LogProb at iter 7 = -600.06830 using 68 examples change = 0.00030
Ave LogProb at iter 8 = -600.06801 using 68 examples change = 0.00028
Ave LogProb at iter 9 = -600.06801 using 68 examples change = -0.00002
Estimation converged at iteration 9
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HRest for HMM N
HRest -A -u tmvw -w 3 -v 0.05 -i 10 -L labels/bcplabs/mon -l N -C toolconfs/hrest.conf -D -M hmms/hmm.1 -T 1 hmms/hmm.0/N data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Reestimating HMM hmms/hmm.0/N . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : N
MaxIter : 10
Epsilon : 0.000100
Updating : Transitions Means Variances
- system is PLAIN
16 Examples loaded, Max length = 13, Min length = 4
Ave LogProb at iter 1 = -395.06061 using 16 examples
Ave LogProb at iter 2 = -403.25287 using 16 examples change = -8.19226
Ave LogProb at iter 3 = -403.23053 using 16 examples change = 0.02234
Ave LogProb at iter 4 = -403.20822 using 16 examples change = 0.02231
Ave LogProb at iter 5 = -403.19427 using 16 examples change = 0.01395
Ave LogProb at iter 6 = -403.19070 using 16 examples change = 0.00357
Ave LogProb at iter 7 = -403.19000 using 16 examples change = 0.00070
Ave LogProb at iter 8 = -403.18976 using 16 examples change = 0.00024
Ave LogProb at iter 9 = -403.18970 using 16 examples change = 0.00006
Estimation converged at iteration 9
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HRest for HMM L
HRest -A -u tmvw -w 3 -v 0.05 -i 10 -L labels/bcplabs/mon -l L -C toolconfs/hrest.conf -D -M hmms/hmm.1 -T 1 hmms/hmm.0/L data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Reestimating HMM hmms/hmm.0/L . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : L
MaxIter : 10
Epsilon : 0.000100
Updating : Transitions Means Variances
- system is PLAIN
25 Examples loaded, Max length = 14, Min length = 4
Ave LogProb at iter 1 = -419.20465 using 25 examples
Ave LogProb at iter 2 = -425.78465 using 25 examples change = -6.58000
Ave LogProb at iter 3 = -425.76195 using 25 examples change = 0.02268
Ave LogProb at iter 4 = -425.74531 using 25 examples change = 0.01665
Ave LogProb at iter 5 = -425.73094 using 25 examples change = 0.01436
Ave LogProb at iter 6 = -425.72137 using 25 examples change = 0.00956
Ave LogProb at iter 7 = -425.71652 using 25 examples change = 0.00485
Ave LogProb at iter 8 = -425.71375 using 25 examples change = 0.00277
Ave LogProb at iter 9 = -425.71164 using 25 examples change = 0.00210
Ave LogProb at iter 10 = -425.71000 using 25 examples change = 0.00164
Estimation aborted at iteration 10
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Iteration 1 of Embedded Re-estimation
HERest -A -w 3 -v 0.05 -C toolconfs/herest.conf -u tmvw -d hmms/hmm.1 -D -M hmms/hmm.2 -L labels/bcplabs/mon -t 2000.0 -T 1 lists/bcplist data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[4]
Module/Tool Parameter Value
# BINARYACCFORMAT TRUE
# KEEPDISTINCT TRUE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
HERest ML Updating: Transitions Means Variances
System is PLAIN
5 Logical/5 Physical Models Loaded, VecSize=26
Pruning-On[2000.0]
Processing Data: tr1.mfc; Label tr1.lab
Utterance prob per frame = -5.980291e+01
Processing Data: tr2.mfc; Label tr2.lab
Utterance prob per frame = -5.937575e+01
Processing Data: tr3.mfc; Label tr3.lab
Utterance prob per frame = -5.935562e+01
Processing Data: tr4.mfc; Label tr4.lab
Utterance prob per frame = -5.919727e+01
Processing Data: tr5.mfc; Label tr5.lab
Utterance prob per frame = -5.789522e+01
Processing Data: tr6.mfc; Label tr6.lab
Utterance prob per frame = -5.910780e+01
Processing Data: tr7.mfc; Label tr7.lab
Utterance prob per frame = -5.739422e+01
Total 27 floored variance elements in 15 different mixes
Saving hmm's to dir hmms/hmm.2
Reestimation complete - average log prob per frame = -5.900193e+01
- total frames seen = 1.811000e+03
HTK Configuration Parameters[4]
Module/Tool Parameter Value
BINARYACCFORMAT TRUE
KEEPDISTINCT TRUE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
cp: la cible `hmms/tmp' n'est pas un répertoire
Iteration 2 of Embedded Re-estimation
HERest -A -w 3 -v 0.05 -C toolconfs/herest.conf -u tmvw -d hmms/tmp -D -M hmms/hmm.2 -L labels/bcplabs/mon -t 2000.0 -T 1 lists/bcplist data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[4]
Module/Tool Parameter Value
# BINARYACCFORMAT TRUE
# KEEPDISTINCT TRUE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
ERROR [+5010] InitSource: Cannot open source file hmms/tmp/C
ERROR [+7010] LoadHMMSet: Can't find file
ERROR [+2321] Initialise: LoadHMMSet failed
FATAL ERROR - Terminating program HERest
cp: la cible `hmms/tmp' n'est pas un répertoire
Iteration 3 of Embedded Re-estimation
HERest -A -w 3 -v 0.05 -C toolconfs/herest.conf -u tmvw -d hmms/tmp -D -M hmms/hmm.2 -L labels/bcplabs/mon -t 2000.0 -T 1 lists/bcplist data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[4]
Module/Tool Parameter Value
# BINARYACCFORMAT TRUE
# KEEPDISTINCT TRUE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
ERROR [+5010] InitSource: Cannot open source file hmms/tmp/C
ERROR [+7010] LoadHMMSet: Can't find file
ERROR [+2321] Initialise: LoadHMMSet failed
FATAL ERROR - Terminating program HERest
cp: la cible `hmms/tmp' n'est pas un répertoire
Removed 0 files from hmms/tmp
Testing on the training set
Can't open test
root@beji-mohamed-desktop:~/Téléchargements/samples/HTKDemo#
I run succesfly ./configure , make all , make install but I had a problem testing the HTKDemo like showed below :
Number of owners = 1
SegLab : S
maxIter : 10
epsilon : 0.000100
minSeg : 3
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
16 Observation Sequences Loaded
Starting Estimation Process
Iteration 1: Average LogP = -804.22485
Iteration 2: Average LogP = -786.82153 Change = 17.40332
Iteration 3: Average LogP = -786.21497 Change = 0.60657
Iteration 4: Average LogP = -786.21503 Change = -0.00006
Estimation converged at iteration 5
Output written to directory hmms/hmm.0
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HInit for HMM C
HInit -A -i 10 -L labels/bcplabs/mon -l C -o C -C toolconfs/hinit.conf -D -M hmms/hmm.0 -T 1 proto/C data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Initialising HMM proto/C . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : C
maxIter : 10
epsilon : 0.000100
minSeg : 3
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
90 Observation Sequences Loaded
Starting Estimation Process
Iteration 1: Average LogP = -531.87134
Iteration 2: Average LogP = -527.31677 Change = 4.55458
Iteration 3: Average LogP = -526.60242 Change = 0.71439
Iteration 4: Average LogP = -526.32343 Change = 0.27898
Iteration 5: Average LogP = -526.26605 Change = 0.05737
Iteration 6: Average LogP = -526.24451 Change = 0.02152
Iteration 7: Average LogP = -526.23560 Change = 0.00892
Iteration 8: Average LogP = -526.23364 Change = 0.00196
Iteration 9: Average LogP = -526.22852 Change = 0.00513
Iteration 10: Average LogP = -526.22852 Change = 0.00000
Estimation converged at iteration 11
Output written to directory hmms/hmm.0
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HInit for HMM V
HInit -A -i 10 -L labels/bcplabs/mon -l V -o V -C toolconfs/hinit.conf -D -M hmms/hmm.0 -T 1 proto/V data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Initialising HMM proto/V . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : V
maxIter : 10
epsilon : 0.000100
minSeg : 3
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
68 Observation Sequences Loaded
Starting Estimation Process
Iteration 1: Average LogP = -599.68671
Iteration 2: Average LogP = -592.98328 Change = 6.70342
Iteration 3: Average LogP = -592.21442 Change = 0.76883
Iteration 4: Average LogP = -591.48163 Change = 0.73280
Iteration 5: Average LogP = -591.14313 Change = 0.33848
Iteration 6: Average LogP = -591.08777 Change = 0.05535
Iteration 7: Average LogP = -591.05005 Change = 0.03773
Iteration 8: Average LogP = -590.98999 Change = 0.06004
Iteration 9: Average LogP = -590.98560 Change = 0.00441
Iteration 10: Average LogP = -590.98558 Change = 0.00001
Estimation converged at iteration 11
Output written to directory hmms/hmm.0
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HInit for HMM N
HInit -A -i 10 -L labels/bcplabs/mon -l N -o N -C toolconfs/hinit.conf -D -M hmms/hmm.0 -T 1 proto/N data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Initialising HMM proto/N . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : N
maxIter : 10
epsilon : 0.000100
minSeg : 3
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
16 Observation Sequences Loaded
Starting Estimation Process
Iteration 1: Average LogP = -398.18735
Iteration 2: Average LogP = -395.48718 Change = 2.70016
Iteration 3: Average LogP = -395.26947 Change = 0.21771
Iteration 4: Average LogP = -395.17661 Change = 0.09286
Iteration 5: Average LogP = -395.17661 Change = 0.00000
Estimation converged at iteration 6
Output written to directory hmms/hmm.0
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HInit for HMM L
HInit -A -i 10 -L labels/bcplabs/mon -l L -o L -C toolconfs/hinit.conf -D -M hmms/hmm.0 -T 1 proto/L data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Initialising HMM proto/L . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : L
maxIter : 10
epsilon : 0.000100
minSeg : 3
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
25 Observation Sequences Loaded
Starting Estimation Process
Iteration 1: Average LogP = -424.39636
Iteration 2: Average LogP = -420.31281 Change = 4.08355
Iteration 3: Average LogP = -419.43475 Change = 0.87804
Iteration 4: Average LogP = -419.35437 Change = 0.08038
Iteration 5: Average LogP = -419.32385 Change = 0.03050
Iteration 6: Average LogP = -419.32387 Change = -0.00001
Estimation converged at iteration 7
Output written to directory hmms/hmm.0
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HRest for HMM S
HRest -A -u tmvw -w 3 -v 0.05 -i 10 -L labels/bcplabs/mon -l S -C toolconfs/hrest.conf -D -M hmms/hmm.1 -T 1 hmms/hmm.0/S data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Reestimating HMM hmms/hmm.0/S . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : S
MaxIter : 10
Epsilon : 0.000100
Updating : Transitions Means Variances
- system is PLAIN
16 Examples loaded, Max length = 20, Min length = 13
Ave LogProb at iter 1 = -785.97754 using 16 examples
Ave LogProb at iter 2 = -799.75970 using 16 examples change = -13.78217
Ave LogProb at iter 3 = -799.59235 using 16 examples change = 0.16736
Ave LogProb at iter 4 = -799.49408 using 16 examples change = 0.09827
Ave LogProb at iter 5 = -799.43396 using 16 examples change = 0.06012
Ave LogProb at iter 6 = -799.36908 using 16 examples change = 0.06488
Ave LogProb at iter 7 = -799.31201 using 16 examples change = 0.05707
Ave LogProb at iter 8 = -799.28625 using 16 examples change = 0.02576
Ave LogProb at iter 9 = -799.26813 using 16 examples change = 0.01813
Ave LogProb at iter 10 = -798.93494 using 16 examples change = 0.33319
Estimation aborted at iteration 10
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HRest for HMM C
HRest -A -u tmvw -w 3 -v 0.05 -i 10 -L labels/bcplabs/mon -l C -C toolconfs/hrest.conf -D -M hmms/hmm.1 -T 1 hmms/hmm.0/C data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Reestimating HMM hmms/hmm.0/C . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : C
MaxIter : 10
Epsilon : 0.000100
Updating : Transitions Means Variances
- system is PLAIN
90 Examples loaded, Max length = 25, Min length = 3
Ave LogProb at iter 1 = -525.97860 using 90 examples
Ave LogProb at iter 2 = -530.95642 using 90 examples change = -4.97785
Ave LogProb at iter 3 = -530.95356 using 90 examples change = 0.00286
Ave LogProb at iter 4 = -530.95312 using 90 examples change = 0.00043
Ave LogProb at iter 5 = -530.95304 using 90 examples change = 0.00009
Estimation converged at iteration 5
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HRest for HMM V
HRest -A -u tmvw -w 3 -v 0.05 -i 10 -L labels/bcplabs/mon -l V -C toolconfs/hrest.conf -D -M hmms/hmm.1 -T 1 hmms/hmm.0/V data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Reestimating HMM hmms/hmm.0/V . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : V
MaxIter : 10
Epsilon : 0.000100
Updating : Transitions Means Variances
- system is PLAIN
68 Examples loaded, Max length = 28, Min length = 3
Ave LogProb at iter 1 = -590.79446 using 68 examples
Ave LogProb at iter 2 = -600.09358 using 68 examples change = -9.29914
Ave LogProb at iter 3 = -600.08226 using 68 examples change = 0.01131
Ave LogProb at iter 4 = -600.07560 using 68 examples change = 0.00668
Ave LogProb at iter 5 = -600.07071 using 68 examples change = 0.00491
Ave LogProb at iter 6 = -600.06859 using 68 examples change = 0.00215
Ave LogProb at iter 7 = -600.06830 using 68 examples change = 0.00030
Ave LogProb at iter 8 = -600.06801 using 68 examples change = 0.00028
Ave LogProb at iter 9 = -600.06801 using 68 examples change = -0.00002
Estimation converged at iteration 9
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HRest for HMM N
HRest -A -u tmvw -w 3 -v 0.05 -i 10 -L labels/bcplabs/mon -l N -C toolconfs/hrest.conf -D -M hmms/hmm.1 -T 1 hmms/hmm.0/N data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Reestimating HMM hmms/hmm.0/N . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : N
MaxIter : 10
Epsilon : 0.000100
Updating : Transitions Means Variances
- system is PLAIN
16 Examples loaded, Max length = 13, Min length = 4
Ave LogProb at iter 1 = -395.06061 using 16 examples
Ave LogProb at iter 2 = -403.25287 using 16 examples change = -8.19226
Ave LogProb at iter 3 = -403.23053 using 16 examples change = 0.02234
Ave LogProb at iter 4 = -403.20822 using 16 examples change = 0.02231
Ave LogProb at iter 5 = -403.19427 using 16 examples change = 0.01395
Ave LogProb at iter 6 = -403.19070 using 16 examples change = 0.00357
Ave LogProb at iter 7 = -403.19000 using 16 examples change = 0.00070
Ave LogProb at iter 8 = -403.18976 using 16 examples change = 0.00024
Ave LogProb at iter 9 = -403.18970 using 16 examples change = 0.00006
Estimation converged at iteration 9
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Calling HRest for HMM L
HRest -A -u tmvw -w 3 -v 0.05 -i 10 -L labels/bcplabs/mon -l L -C toolconfs/hrest.conf -D -M hmms/hmm.1 -T 1 hmms/hmm.0/L data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[3]
Module/Tool Parameter Value
# KEEPDISTINCT FALSE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
Reestimating HMM hmms/hmm.0/L . . .
States : 2 3 4 (width)
Mixes s1: 1 1 1 ( 26 )
Num Using: 0 0 0
Parm Kind: MFCC_E_D
Number of owners = 1
SegLab : L
MaxIter : 10
Epsilon : 0.000100
Updating : Transitions Means Variances
- system is PLAIN
25 Examples loaded, Max length = 14, Min length = 4
Ave LogProb at iter 1 = -419.20465 using 25 examples
Ave LogProb at iter 2 = -425.78465 using 25 examples change = -6.58000
Ave LogProb at iter 3 = -425.76195 using 25 examples change = 0.02268
Ave LogProb at iter 4 = -425.74531 using 25 examples change = 0.01665
Ave LogProb at iter 5 = -425.73094 using 25 examples change = 0.01436
Ave LogProb at iter 6 = -425.72137 using 25 examples change = 0.00956
Ave LogProb at iter 7 = -425.71652 using 25 examples change = 0.00485
Ave LogProb at iter 8 = -425.71375 using 25 examples change = 0.00277
Ave LogProb at iter 9 = -425.71164 using 25 examples change = 0.00210
Ave LogProb at iter 10 = -425.71000 using 25 examples change = 0.00164
Estimation aborted at iteration 10
HTK Configuration Parameters[3]
Module/Tool Parameter Value
KEEPDISTINCT FALSE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
Iteration 1 of Embedded Re-estimation
HERest -A -w 3 -v 0.05 -C toolconfs/herest.conf -u tmvw -d hmms/hmm.1 -D -M hmms/hmm.2 -L labels/bcplabs/mon -t 2000.0 -T 1 lists/bcplist data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[4]
Module/Tool Parameter Value
# BINARYACCFORMAT TRUE
# KEEPDISTINCT TRUE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
HERest ML Updating: Transitions Means Variances
System is PLAIN
5 Logical/5 Physical Models Loaded, VecSize=26
Pruning-On[2000.0]
Processing Data: tr1.mfc; Label tr1.lab
Utterance prob per frame = -5.980291e+01
Processing Data: tr2.mfc; Label tr2.lab
Utterance prob per frame = -5.937575e+01
Processing Data: tr3.mfc; Label tr3.lab
Utterance prob per frame = -5.935562e+01
Processing Data: tr4.mfc; Label tr4.lab
Utterance prob per frame = -5.919727e+01
Processing Data: tr5.mfc; Label tr5.lab
Utterance prob per frame = -5.789522e+01
Processing Data: tr6.mfc; Label tr6.lab
Utterance prob per frame = -5.910780e+01
Processing Data: tr7.mfc; Label tr7.lab
Utterance prob per frame = -5.739422e+01
Total 27 floored variance elements in 15 different mixes
Saving hmm's to dir hmms/hmm.2
Reestimation complete - average log prob per frame = -5.900193e+01
- total frames seen = 1.811000e+03
HTK Configuration Parameters[4]
Module/Tool Parameter Value
BINARYACCFORMAT TRUE
KEEPDISTINCT TRUE
SAVEGLOBOPTS TRUE
TARGETKIND MFCC_E_D
cp: la cible `hmms/tmp' n'est pas un répertoire
Iteration 2 of Embedded Re-estimation
HERest -A -w 3 -v 0.05 -C toolconfs/herest.conf -u tmvw -d hmms/tmp -D -M hmms/hmm.2 -L labels/bcplabs/mon -t 2000.0 -T 1 lists/bcplist data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[4]
Module/Tool Parameter Value
# BINARYACCFORMAT TRUE
# KEEPDISTINCT TRUE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
ERROR [+5010] InitSource: Cannot open source file hmms/tmp/C
ERROR [+7010] LoadHMMSet: Can't find file
ERROR [+2321] Initialise: LoadHMMSet failed
FATAL ERROR - Terminating program HERest
cp: la cible `hmms/tmp' n'est pas un répertoire
Iteration 3 of Embedded Re-estimation
HERest -A -w 3 -v 0.05 -C toolconfs/herest.conf -u tmvw -d hmms/tmp -D -M hmms/hmm.2 -L labels/bcplabs/mon -t 2000.0 -T 1 lists/bcplist data/train/tr1.mfc data/train/tr2.mfc data/train/tr3.mfc data/train/tr4.mfc data/train/tr5.mfc data/train/tr6.mfc data/train/tr7.mfc
HTK Configuration Parameters[4]
Module/Tool Parameter Value
# BINARYACCFORMAT TRUE
# KEEPDISTINCT TRUE
# SAVEGLOBOPTS TRUE
# TARGETKIND MFCC_E_D
ERROR [+5010] InitSource: Cannot open source file hmms/tmp/C
ERROR [+7010] LoadHMMSet: Can't find file
ERROR [+2321] Initialise: LoadHMMSet failed
FATAL ERROR - Terminating program HERest
cp: la cible `hmms/tmp' n'est pas un répertoire
Removed 0 files from hmms/tmp
Testing on the training set
Can't open test
root@beji-mohamed-desktop:~/Téléchargements/samples/HTKDemo#