[hts-users:04512] Error On Phoneme #
- Subject: [hts-users:04512] Error On Phoneme #
- From: Franco Negri <franconegri2004@xxxxxxxxx>
- Date: Thu, 23 Mar 2017 11:24:41 -0300
- Authentication-results: mailgw.mains.nitech.ac.jp; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b=BHpX93V4
- Delivered-to: hts-users@xxxxxxxxxxxxxxx
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=mime-version:from:date:message-id:subject:to; bh=1X1ynj3IicACtVMe8m3zK1Wbi7uGsfIupcq4t/OcnvI=; b=BHpX93V4PtYxAkKg8ocxsr5GzvAgVUS5+m370TYMJZPCzfihLLR4OAxqzPLGtf7Sxx qxdPbto3//mqDrYqwPV9TSkRIrN9yG+ntuhaw9Ulmi3uKOHQT1ucftCEU16zwNRhCb5d D2nTLiumzw8a84/MsWDG7S9sW0wRJsZJu0iSXEHDgLltagITE2v6j6HorCzzRhS7W6Y6 k5yDyFJUAUv/Wmfqh1eo7MxdeufzR9491XW3wHW2isMFWW0247oWGyOJAdt0uPU6EPv0 2Z5fwfSLOgtw/YioaFb5gCIa7iM/5kvRGxN2I9uxGL7o5y3+vhPWLD4h9h5AuCjDd35c OXZw==
Hi, i'm experiencing a confusing error:
=================================================================
Start computing variance floors at Thu Mar 23 11:00:49 ART 2017
=================================================================
/usr/local/HTS-2.3/bin/HCompV -A -C
/home/franckn/htk/habla_utils/main/configs/qst001/ver1/trn.cnf -D -T 1
-S /home/franckn/htk/habla_utils/main/data/scp/train.scp -m -M
/home/franckn/htk/habla_utils/main/models/qst001/ver1/cmp -o
/home/franckn/htk/habla_utils/main/models/qst001/ver1/cmp/average.mmf
/home/franckn/htk/habla_utils/main/proto/qst001/ver1/state-5_stream-4_mgc-105_lf0-3.prt
HTK Configuration Parameters[8]
Module/Tool Parameter Value
# MINDUR 5
# MAXSTDDEVCOEF 10
# APPLYDURVARFLOOR TRUE
# DURVARFLOORPERCENTILE 1.000000
# VFLOORSCALESTR Vector 4 0.01 0.01 0.01 0.01
# NATURALWRITEORDER TRUE
# NATURALREADORDER TRUE
# APPLYVFLOOR TRUE
Calculating Fixed Variance
HMM Prototype:
/home/franckn/htk/habla_utils/main/proto/qst001/ver1/state-5_stream-4_mgc-105_lf0-3.prt
Segment Label: None
Num Streams : 4
UpdatingMeans: Yes
Target Direct: /home/franckn/htk/habla_utils/main/models/qst001/ver1/cmp
567162 speech frames accumulated for stream 1
329871 speech frames accumulated for stream 2
316010 speech frames accumulated for stream 3
316010 speech frames accumulated for stream 4
Updating HMM Means and Covariances
Output written to directory
/home/franckn/htk/habla_utils/main/models/qst001/ver1/cmp
Var floor macros output to file
/home/franckn/htk/habla_utils/main/models/qst001/ver1/cmp/vFloors
HTK Configuration Parameters[8]
Module/Tool Parameter Value
# MINDUR 5
# MAXSTDDEVCOEF 10
# APPLYDURVARFLOOR TRUE
# DURVARFLOORPERCENTILE 1.000000
VFLOORSCALESTR Vector 4 0.01 0.01 0.01 0.01
NATURALWRITEORDER TRUE
NATURALREADORDER TRUE
# APPLYVFLOOR TRUE
=====================================================================
Start initialization & reestimation at Thu Mar 23 11:00:53 ART 2017
=====================================================================
=============== # ================
/usr/local/HTS-2.3/bin/HInit -A -C
/home/franckn/htk/habla_utils/main/configs/qst001/ver1/trn.cnf -D -T 1
-S /home/franckn/htk/habla_utils/main/data/scp/train.scp -m 1 -u tmvw
-w 5000 -H /home/franckn/htk/habla_utils/main/models/qst001/ver1/cmp/init.mmf
-M /home/franckn/htk/habla_utils/main/models/qst001/ver1/cmp/HInit -I
/home/franckn/htk/habla_utils/main/data/labels/mono.mlf -l # -o #
/home/franckn/htk/habla_utils/main/proto/qst001/ver1/state-5_stream-4_mgc-105_lf0-3.prt
HTK Configuration Parameters[8]
Module/Tool Parameter Value
# MINDUR 5
# MAXSTDDEVCOEF 10
# APPLYDURVARFLOOR TRUE
# DURVARFLOORPERCENTILE 1.000000
# VFLOORSCALESTR Vector 4 0.01 0.01 0.01 0.01
# NATURALWRITEORDER TRUE
# NATURALREADORDER TRUE
# APPLYVFLOOR TRUE
Initialising HMM
/home/franckn/htk/habla_utils/main/proto/qst001/ver1/state-5_stream-4_mgc-105_lf0-3.prt
. . .
States : 2 3 4 5 6 (width)
Mixes s1: 1 1 1 1 1 ( 105 )
( 105 )
Num Using: 0 0 0 0 0
Mixes s2: 2 2 2 2 2 ( 1 )
( MSD <= 1 )
Num Using: 0 0 0 0 0
Mixes s3: 2 2 2 2 2 ( 1 )
( MSD <= 1 )
Num Using: 0 0 0 0 0
Mixes s4: 2 2 2 2 2 ( 1 )
( MSD <= 1 )
Num Using: 0 0 0 0 0
Parm Kind: USER
Number of owners = 1
SegLab : #
maxIter : 20
epsilon : 0.000100
minSeg : 1
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
ERROR [+2121] HInit: Too Few Observation Sequences [0]
FATAL ERROR - Terminating program /usr/local/HTS-2.3/bin/HInit
Error in /usr/local/HTS-2.3/bin/HInit -A -C
/home/franckn/htk/habla_utils/main/configs/qst001/ver1/trn.cnf -D -T 1
-S /home/franckn/htk/habla_utils/main/data/scp/train.scp
-m 1 -u tmvw -w 5000 -H
/home/franckn/htk/habla_utils/main/models/qst001/ver1/cmp/init.mmf -M
/home/franckn/htk/habla_utils/main/models/qst001/ver1/cmp/HInit -I
/home/franckn/htk/habla_utils/main/data/labels/mono.mlf -l # -o #
/home/franckn/htk/habla_utils/main/proto/qst001/ver1/state-5_stream-4_mgc-105_lf0-3.prt
According to what I understand this means hts is finding too few
examples of the label "#", witch is very rare given that my labs all
look like this:
0 0.326719 #
0.326719 0.414777 l
0.414777 0.480529 a
0.480529 0.551305 s
[...]
2.200350 2.290139 i1
2.290139 2.444838 l
2.444838 2.488431 #
2.488431 2.602645 d
2.602645 2.694007 i
2.694007 2.874066 f
2.874066 2.953478 i0
[...]
3.774066 3.863322 i1
3.863322 3.971125 ny
3.971125 4.120223 o
4.120223 4.312 #
- Follow-Ups
-
- [hts-users:04513] Re: Error On Phoneme #, Sébastien Le Maguer