[Subject Prev][Subject Next][Thread Prev][Thread Next][Date Index][Thread Index]

[hts-users:02910] Re: Error in "regression-class trees for adaptation"


    I suggest that you can check the line endings of your labels
files. I've met a similar problem, and found that it is caused by the
mixed line endings used in the labels. They must all the the unix
version of line ending (\n) for HTS to work properly.

    I think that you'd better make sure all the text file used have
unix line endings.

Yongxin Wang

On Wed, Jul 6, 2011 at 14:16, 김일환 <cutekbj@xxxxxxxxx> wrote:
> I'm building a voice using the adaptation example provided in the hts web.
>
> Everyting runs fine until the "regression-class trees for adaptation" step I
> go the error below and I didn't manage to find what is the problem with the
> data.
>
> Is this question file problem?, or label symbol set??
>
> My full-label is,
>
> iy^m-ow+th=ww/S:2@2_3/U:1_10|0-0
> m^ow-th+ww=kh/S:3@1_3/U:1_10|0-0
> ow^th-ww+kh=ow/S:3@1_3/U:1_10|0-0
> ...
>
> I used 2.1.1 version.
>
> Would someone by so kind to give a hint about how to continue debugging this
> issue?
>
>
>
> =======================================================================================
> Start building regression-class trees for adaptation at Thu May 12 09:52:45
> CEST 2011
> =======================================================================================
>
> /veu4/usuaris20/tts/tools/hts/bin/HHEd -A -B -C
> /veu4/usuaris20/jadell/work/hp/work//configs/trn.cnf -D -T 1 -p -i -H
> /veu4/usuaris20/jadell/work/hp/work/models/qst001/ver1/cmp/re_clustered.mmf
> -M /veu4/usuaris20/jadell/work/hp/work//models/qst001/ver1/cmp/regTrees
> /veu4/usuaris20/jadell/work/hp/work//edfiles/qst001/ver1/cmp/reg.hed
> /veu4/usuaris20/jadell/work/hp/work//data/lists/full.list
>
> HTK Configuration Parameters[10]
>   Module/Tool     Parameter                  Value
> #                 MINDUR                         5
> #                 MAXSTDDEVCOEF                 10
> #                 APPLYDURVARFLOOR              TRUE
> #                 DURVARFLOORPERCENTILE          1.000000
> #                 SHRINKOCCTHRESH  Vector 4 500.0 100.0 100.0 100.0
> #                 VFLOORSCALESTR  Vector 4 0.01 0.01 0.01 0.01
> #                 MINLEAFOCC                     0
> #                 NATURALWRITEORDER              TRUE
> #                 NATURALREADORDER              TRUE
> #                 APPLYVFLOOR                 TRUE
>
> HHEd
>  161701/161701 Models Loaded [7 states max, 2 mixes max]
>
> // load stats file
> LS
> /veu4/usuaris20/jadell/work/hp/work/HP2Hagrid/stats/qst001/ver1/cmp.stats.re-clustered
>  Loading state occupation stats
>   Stats loaded for 161701 models
>   Mean Occupation Count = 3.604810
>
> // construct regression class tree
> RC 32 reg
>  Building regression tree with 32 terminals (4 streams)
> Creating regression class tree with ident reg.tree and baseclass reg.base
> Splitting Node 1, score 1.000000e+10
> (Stream splitting)
> Splitting Node 3, score 1.000000e+10
> (Stream splitting)
>
> .......
>
> (Stream=4, vSize=1)
> Splitting Node 174, score 1.653266e+01
> (Stream=4, vSize=1)
> Splitting Node 175, score 1.628670e+01
> (Stream=4, vSize=1)
> Splitting Node 172, score 1.590286e+01
> (Stream=4, vSize=1)
> Splitting Node 179, score 1.589252e+01
> (Stream=4, vSize=1)
> No more nodes to split..
> stream 1: #terminals=6
> stream 2: #terminals=32
> stream 3: #terminals=32
> stream 4: #terminals=32
>
>  ERROR [+999]  CheckBaseClass: Component specified multiple times
>  ERROR [+999]  BaseClass check failed
>  FATAL ERROR - Terminating program /veu4/usuaris20/tts/tools/hts/bin/HHEd
>

Follow-Ups
[hts-users:02915] Re: Error in "regression-class trees for adaptation", 김일환
[hts-users:02921] Re: Error in "regression-class trees for adaptation", 那兴宇
References
[hts-users:02909] Error in "regression-class trees for adaptation", 김일환