Menu
▾
▴
kaldi-users — Kaldi Users
You can subscribe to this list here.
| 2011 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(2) |
Aug
(2) |
Sep
(1) |
Oct
(1) |
Nov
|
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2012 |
Jan
|
Feb
|
Mar
(8) |
Apr
(4) |
May
(2) |
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2013 |
Jan
|
Feb
(2) |
Mar
(2) |
Apr
(7) |
May
(31) |
Jun
(40) |
Jul
(65) |
Aug
(37) |
Sep
(12) |
Oct
(57) |
Nov
(15) |
Dec
(35) |
| 2014 |
Jan
(3) |
Feb
(30) |
Mar
(57) |
Apr
(26) |
May
(49) |
Jun
(26) |
Jul
(63) |
Aug
(33) |
Sep
(20) |
Oct
(153) |
Nov
(62) |
Dec
(20) |
| 2015 |
Jan
(6) |
Feb
(21) |
Mar
(42) |
Apr
(33) |
May
(76) |
Jun
(102) |
Jul
(39) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: David M. <da...@ca...> - 2014-02-05 15:19:02
|
Hi Karel, utils/int2sym.pl seems to expect -f option. This command-line produced sensible looking lattices lattice-align-words exp/tri3b/graph_bd_tgpr/phones/word_boundary.int exp/tri3b/final.mdl 'ark:gunzip -c exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | utils/int2sym.pl -f 3 exp/tri3b/graph_bd_tgpr/words.txt | utils/convert_slf.pl --word-to-node - ~/slf/ On the other hand, this one with the new tool produced lattices with word indexes: lattice-align-words-lexicon data/lang_bd/phones/align_lexicon.int exp/tri3b/final.mdl 'ark:gunzip -c exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | utils/convert_slf.pl --word-to-node - ~/slf/ The first pipe worked so I will go with that. Thanks, David On 05/02/2014 13:37, Vesely Karel wrote: > Hello David, > I just modified the example to include the int->word mapping, > the lattices will now contain words, it is commited. > > Eventually you can also try newer tool 'lattice-align-words-lexicon', > which uses lexicon to do the aligning. For this you need to have > data/lang/phones/align_lexicon.int > > K. > > > On 02/05/2014 01:32 PM, Korbinian Riedhammer wrote: >> Hi David, >> >> yes, the numbers associated with the words are indices, you can >> resolve them using the data/lang/words.txt file. The warnings you see >> relate to process of alignining the lattices to words (i.e. making >> sure that the word arcs (transitions) reflect the actual word >> boundaries in terms of time), and not to the SLF conversion. To get >> rid of the epsilon transitions you may need to determinize the >> lattices first (please correct me, if I'm wrong, Dan). >> >> Korbinian. >> >> On Wed, Feb 5, 2014 at 12:34 PM, David Mrva >> <da...@ca...> wrote: >>> Hi Karel, >>> >>> Thanks for you help. Based on you new example in the convert_slf.pl >>> script I >>> made up a command-line that now produces HTK lattices from wsj >>> example kaldi >>> lattices. It gives me a warning. Is it anything to worry about? The >>> resulting lattices have words labelled with numbers, see below. Are the >>> numbers some indexes? How can I get words to the HTK lattice? >>> >>> David >>> >>> |
|
From: Vesely K. <ive...@fi...> - 2014-02-05 13:38:02
|
Hello David, I just modified the example to include the int->word mapping, the lattices will now contain words, it is commited. Eventually you can also try newer tool 'lattice-align-words-lexicon', which uses lexicon to do the aligning. For this you need to have data/lang/phones/align_lexicon.int K. On 02/05/2014 01:32 PM, Korbinian Riedhammer wrote: > Hi David, > > yes, the numbers associated with the words are indices, you can > resolve them using the data/lang/words.txt file. The warnings you see > relate to process of alignining the lattices to words (i.e. making > sure that the word arcs (transitions) reflect the actual word > boundaries in terms of time), and not to the SLF conversion. To get > rid of the epsilon transitions you may need to determinize the > lattices first (please correct me, if I'm wrong, Dan). > > Korbinian. > > On Wed, Feb 5, 2014 at 12:34 PM, David Mrva <da...@ca...> wrote: >> Hi Karel, >> >> Thanks for you help. Based on you new example in the convert_slf.pl script I >> made up a command-line that now produces HTK lattices from wsj example kaldi >> lattices. It gives me a warning. Is it anything to worry about? The >> resulting lattices have words labelled with numbers, see below. Are the >> numbers some indexes? How can I get words to the HTK lattice? >> >> David >> >> >> [davidm@code4 s5]$ lattice-align-words >> exp/tri3b/graph_bd_tgpr/phones/word_boundary.int exp/tri3b/final.mdl >> 'ark:gunzip -c exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | >> utils/convert_slf.pl --word-to-node - ~/slf/ >> lattice-align-words exp/tri3b/graph_bd_tgpr/phones/word_boundary.int >> exp/tri3b/final.mdl 'ark:gunzip -c >> exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) >> [Lattice has input epsilons and/or is not input-deterministic (in Mohri >> sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow >> and-or blow up in memory. >> LOG (lattice-align-words:main():lattice-align-words.cc:122) Successfully >> aligned 43 lattices; 0 had errors. >> >> resulting lattice snippet: >> >> VERSION=1.1 >> UTTERANCE=440c0401 >> N=291 L=411 >> I=0 W=!NULL t=0.00 >> I=1 W=!NULL t=0.88 >> I=2 W=!NULL t=1.14 >> I=3 W=!NULL t=1.38 >> I=4 W=!NULL t=1.75 >> I=5 W=!NULL t=1.38 >> I=6 W=!NULL t=1.75 >> I=7 W=!NULL t=0.87 >> I=8 W=!NULL t=0.74 >> I=9 W=!NULL t=0.77 >> I=10 W=!NULL t=0.87 >> I=11 W=!NULL t=1.34 >> I=12 W=!NULL t=0.74 >> I=13 W=!NULL t=0.77 >> I=14 W=!NULL t=0.87 >> I=15 W=!NULL t=1.34 >> I=16 W=!NULL t=1.75 >> I=17 W=!NULL t=2.08 >> I=18 W=!NULL t=2.14 >> I=19 W=!NULL t=2.43 >> I=20 W=!NULL t=2.56 >> I=21 W=!NULL t=2.11 >> I=22 W=!NULL t=2.43 >> I=23 W=!NULL t=2.56 >> I=24 W=!NULL t=3.10 >> I=25 W=!NULL t=3.25 >> I=26 W=!NULL t=3.59 >> I=27 W=!NULL t=3.10 >> I=28 W=!NULL t=3.25 >> I=29 W=!NULL t=3.59 >> I=30 W=!NULL t=3.59 >> I=31 W=!NULL t=3.59 >> I=32 W=!NULL t=3.59 >> I=33 W=!NULL t=3.59 >> I=34 W=!NULL t=3.78 >> I=35 W=!NULL t=4.21 >> I=36 W=!NULL t=3.78 >> I=37 W=!NULL t=3.99 >> I=38 W=!NULL t=3.80 >> I=39 W=!NULL t=3.99 >> I=40 W=!NULL t=3.78 >> ....... >> I=245 W=73728 t=4.06 >> I=246 W=4230 t=4.21 >> I=247 W=93495 t=4.21 >> I=248 W=140327 t=4.52 >> I=249 W=140327 t=4.52 >> I=250 W=62898 t=3.99 >> I=251 W=72789 t=4.21 >> I=252 W=72799 t=4.21 >> I=253 W=73728 t=4.06 >> I=254 W=4230 t=4.21 >> I=255 W=93495 t=4.21 >> I=256 W=140327 t=4.52 >> I=257 W=132935 t=5.58 >> I=258 W=140327 t=4.52 >> I=259 W=132935 t=5.58 >> I=260 W=140327 t=4.52 >> I=261 W=140327 t=4.52 >> I=262 W=4230 t=3.99 >> I=263 W=72789 t=4.21 >> I=264 W=72799 t=4.21 >> I=265 W=74646 t=4.21 >> I=266 W=140327 t=4.52 >> I=267 W=132935 t=5.58 >> I=268 W=140327 t=4.52 >> I=269 W=132935 t=5.58 >> I=270 W=140327 t=4.52 >> I=271 W=132935 t=5.58 >> I=272 W=124859 t=6.46 >> I=273 W=4230 t=6.62 >> I=274 W=99831 t=6.92 >> I=275 W=43150 t=7.78 >> I=276 W=18179 t=8.47 >> I=277 W=46751 t=8.60 >> I=278 W=25 t=8.70 >> I=279 W=4086 t=8.70 >> I=280 W=93075 t=8.70 >> I=281 W=135497 t=9.45 >> I=282 W=135497 t=9.45 >> I=283 W=121952 t=9.96 >> I=284 W=126807 t=9.96 >> I=285 W=0 t=11.35 >> I=286 W=135497 t=9.45 >> I=287 W=121952 t=9.96 >> I=288 W=126807 t=9.96 >> I=289 W=0 t=11.35 >> I=290 W=0 t=11.35 >> J=0 S=0 E=86 v=0.000000 a=-6128.520000 l=-28.250300 >> J=1 S=86 E=12 v=0.000000 a=0.000000 l=0.000000 >> J=2 S=0 E=87 v=0.000000 a=-6128.520000 l=-14.239300 >> J=3 S=87 E=8 v=0.000000 a=0.000000 l=0.000000 >> J=4 S=0 E=88 v=0.000000 a=-8043.850000 l=-26.245500 >> J=5 S=88 E=7 v=0.000000 a=0.000000 l=0.000000 >> J=6 S=0 E=89 v=0.000000 a=-5821.820000 l=-18.684200 >> J=7 S=89 E=1 v=0.000000 a=0.000000 l=0.000000 >> J=8 S=1 E=90 v=0.000000 a=0.000000 l=0.000000 >> J=9 S=90 E=2 v=0.000000 a=0.000000 l=0.000000 >> J=10 S=2 E=91 v=0.000000 a=-2222.850000 l=-17.005600 >> J=11 S=91 E=5 v=0.000000 a=0.000000 l=0.000000 >> J=12 S=2 E=92 v=0.000000 a=-2222.850000 l=-12.768900 >> J=13 S=92 E=3 v=0.000000 a=0.000000 l=0.000000 >> J=14 S=3 E=93 v=0.000000 a=-952.576000 l=-7.415230 >> J=15 S=93 E=4 v=0.000000 a=0.000000 l=0.000000 >> J=16 S=4 E=94 v=0.000000 a=-4826.290000 l=-11.488200 >> J=17 S=94 E=21 v=0.000000 a=0.000000 l=0.000000 >> J=18 S=4 E=95 v=0.000000 a=-4888.020000 l=-9.040100 >> J=19 S=95 E=17 v=0.000000 a=0.000000 l=0.000000 >> J=20 S=5 E=96 v=0.000000 a=-952.576000 l=-9.439240 >> J=21 S=96 E=6 v=0.000000 a=0.000000 l=0.000000 >> J=22 S=6 E=97 v=0.000000 a=-4826.290000 l=-11.566900 >> J=23 S=97 E=21 v=0.000000 a=0.000000 l=0.000000 >> J=24 S=6 E=98 v=0.000000 a=-4888.020000 l=-9.118870 >> J=25 S=98 E=17 v=0.000000 a=0.000000 l=0.000000 >> J=26 S=7 E=99 v=0.000000 a=0.000000 l=0.000000 >> J=27 S=99 E=15 v=0.000000 a=0.000000 l=0.000000 >> J=28 S=8 E=100 v=0.000000 a=0.000000 l=0.000000 >> J=29 S=100 E=9 v=0.000000 a=0.000000 l=0.000000 >> J=30 S=9 E=101 v=0.000000 a=0.000000 l=0.000000 >> J=31 S=101 E=10 v=0.000000 a=0.000000 l=0.000000 >> J=32 S=10 E=102 v=0.000000 a=-1906.000000 l=-16.253300 >> J=33 S=102 E=11 v=0.000000 a=0.000000 l=0.000000 >> J=34 S=11 E=103 v=0.000000 a=-952.576000 l=-9.806090 >> J=35 S=103 E=16 v=0.000000 a=0.000000 l=0.000000 >> J=36 S=12 E=104 v=0.000000 a=0.000000 l=0.000000 >> J=37 S=104 E=13 v=0.000000 a=0.000000 l=0.000000 >> J=38 S=13 E=105 v=0.000000 a=0.000000 l=0.000000 >> J=39 S=105 E=14 v=0.000000 a=0.000000 l=0.000000 >> J=40 S=14 E=106 v=0.000000 a=-1906.000000 l=-16.458200 >> J=41 S=106 E=15 v=0.000000 a=0.000000 l=0.000000 >> J=42 S=15 E=107 v=0.000000 a=-952.576000 l=-9.806090 >> J=43 S=107 E=16 v=0.000000 a=0.000000 l=0.000000 >> J=44 S=16 E=108 v=0.000000 a=-4826.290000 l=-11.566900 >> J=45 S=108 E=21 v=0.000000 a=0.000000 l=0.000000 >> ......... >> >> >> >> >> >> On 04/02/2014 18:06, Vesely Karel wrote: >> >> Hello, >> the script 'utils/convert_slf.pl' is updated with a new example how to call >> it. >> >> Also I refactored the mode which puts links into nodes, to consider >> situation when a node has incoming arcs with different words. >> >> The default output format is back to generate lattices with words >> in links, as it is more compact. >> >> Karel. >> >> >> >> On 02/04/2014 05:15 PM, Karel Veselý wrote: >> >> Hi, >> sure, recently I rewrote that script recently in cooperation with Korbinian. >> I'll have a look at it, >> Karel. >> >> >> >> >> >> Dne 4.2.2014 17:05, Daniel Povey napsal(a): >> >> You may have seen the message because lattice-word-align expects the model >> file as its first argument; this is an error in the help message of >> utils/convert_slf.pl. Also, the help messagepl should probably be updated >> to use lattice-align-words as its suggested input. >> Karel, you made some recent changes to that program; do you think you could >> help with this? >> Dan >> >> >>> Hi, >>> After running the egs/wsj/s5/run.sh I call the convert_slf.pl script on >>> one of the lattices in tri3b step following an exampl in the scripts help >>> output and get a deprecation message: >>> >>> [davidm@code7 s5]$ utils/convert_slf.pl >>> Convert kaldi lattices to HTK SLF (v1.1) format. >>> Usage: convert_slf.pl [options] lat-file.txt [out-dir] >>> e.g. lattice-word-align 'ark:gunzip -c lat.gz |' ark,t:- | >>> utils/convert_slf.pl - slf/ >>> Options regarding the SLF output: >>> --lmscale x LM weight (default: lmscale=1) >>> --acscale x Acoustic weight (default: acscale=1) >>> --wdpenalty x Word insertion penalty (default: 0) >>> --framerate x Frame rate to compute timing information (default: 0.01) >>> [davidm@code7 s5]$ lattice-word-align 'ark:gunzip -c >>> exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | >>> utils/convert_slf.pl - ~/slf/ >>> lattice-word-align 'ark:gunzip -c >>> exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- >>> >>> (note: from the s5 scripts onward, this is deprecated, see >>> lattice-align-words) >>> Create word-aligned lattices (in which the arcs correspond with >>> word boundaries) >>> Usage: lattice-word-align [options] <model> <lattice-rspecifier> >>> <lattice-wspecifier> >>> e.g.: lattice-word-align --silence-phones=1:2 --wbegin-phones=2:6:10:14 \ >>> --wend-phones=3:7:11:15 --winternal-phones=4:8:12:16 >>> --wbegin-and-end-phones=5:9:13:17 \ >>> --silence-label=2 --partial-word-label=16342 \ >>> final.mdl ark:1.lats ark:aligned.lats >>> >>> Options: >>> --output-error-lats : If true, output aligned lattices even if >>> there was an error (e.g. caused by forced-out lattice) (bool, default = >>> true) >>> --partial-word-label : Numeric id of word symbol that is to be >>> used for arcs in the word-aligned lattice corresponding to partial words at >>> the end of "forced-out" utterances (zero is OK) (int, default = 0) >>> --reorder : True if the lattices were generated from >>> graphs that had the --reorder option true, relating to reordering self-loops >>> (typically true) (bool, default = true) >>> --silence-has-olabels : If true, silence phones have output labels >>> in the lattice, just >>> like regular words. [This means you can't have un-labeled silences] >>> (bool, default = false) >>> --silence-label : Numeric id of word symbol that is to be >>> used for silence arcs in the word-aligned lattice (zero is OK) (int, default >>> = 0) >>> --silence-may-be-word-internal : If true, silence may appear inside >>> words' prons (but not at begin/end!) >>> (bool, default = false) >>> --silence-phones : Colon-separated list of numeric ids of >>> phones that are used for silence (and other non-word events such as noise) >>> (string, default = "") >>> --test : If true, activate checks designed to test >>> the code. (bool, default = false) >>> --wbegin-and-end-phones : Colon-separated list of numeric ids of >>> phones that are used for single-phone words. (string, default = "") >>> --wbegin-phones : Colon-separated list of numeric ids of >>> phones that begin a word (string, default = "") >>> --wend-phones : Colon-separated list of numeric ids of >>> phones that end a word (string, default = "") >>> --winternal-phones : Colon-separated list of numeric ids of >>> phones that are internal to a word (string, default = "") >>> >>> Standard options: >>> --config : Configuration file with options (string, >>> default = "") >>> --help : Print out usage message (bool, default = >>> false) >>> --print-args : Print the command line arguments (to >>> stderr) (bool, default = true) >>> --verbose : Verbose level (higher->more logging) (int, >>> default = 0) >>> >>> What is the right command-line now for calling the convert_slf.pl script >>> on a binary lattice file(s)? >>> >>> Many thanks, >>> David >>> >>> >>> >>> >>> On 20/12/2013 10:44, Karel Veselý wrote: >>> >>> There is also a script directly in kaldi wsj/s5 from Korbinian Riedhammer: >>> >>> >>> http://sourceforge.net/p/kaldi/code/HEAD/tree/trunk/egs/wsj/s5/utils/convert_slf.pl >>> >>> Karel >>> >>> Dne 19. 12. 2013 21:05, Daniel Povey napsal(a): >>> >>> It looks like Paul Dixon has published something he wrote, here: >>> >>> https://gist.github.com/edobashira/5811963 >>> >>> Dan >>> >>> >>> >>> On Thu, Dec 19, 2013 at 3:20 AM, wei.shi-新 <we...@im...> wrote: >>>> Hi all, >>>> >>>> Though I know the purpose of kaldi is not to be compatible with HTK, I >>>> still want to ask, is there a tool that can convert kaldi lattices( e.g., >>>> generated by gmm-latgen-faster ) to HTK compatible format? >>>> >>>> >>>> Shi Wei >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>> organizations don't have a clear picture of how application performance >>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>> your >>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >>>> Pro! >>>> >>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>> _______________________________________________ >>>> Kaldi-users mailing list >>>> Kal...@li... >>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>> >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >>> Pro! >>> >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> >>> >>> >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >>> >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >>> Pro! >>> >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> >>> >>> >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Managing the Performance of Cloud-Based Applications >>> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >>> Read the Whitepaper. >>> >>> http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >> >> >> >> ------------------------------------------------------------------------------ >> Managing the Performance of Cloud-Based Applications >> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >> Read the Whitepaper. >> http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk >> >> >> >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> >> |
|
From: Korbinian R. <kor...@gm...> - 2014-02-05 12:33:33
|
Hi David, yes, the numbers associated with the words are indices, you can resolve them using the data/lang/words.txt file. The warnings you see relate to process of alignining the lattices to words (i.e. making sure that the word arcs (transitions) reflect the actual word boundaries in terms of time), and not to the SLF conversion. To get rid of the epsilon transitions you may need to determinize the lattices first (please correct me, if I'm wrong, Dan). Korbinian. On Wed, Feb 5, 2014 at 12:34 PM, David Mrva <da...@ca...> wrote: > Hi Karel, > > Thanks for you help. Based on you new example in the convert_slf.pl script I > made up a command-line that now produces HTK lattices from wsj example kaldi > lattices. It gives me a warning. Is it anything to worry about? The > resulting lattices have words labelled with numbers, see below. Are the > numbers some indexes? How can I get words to the HTK lattice? > > David > > > [davidm@code4 s5]$ lattice-align-words > exp/tri3b/graph_bd_tgpr/phones/word_boundary.int exp/tri3b/final.mdl > 'ark:gunzip -c exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | > utils/convert_slf.pl --word-to-node - ~/slf/ > lattice-align-words exp/tri3b/graph_bd_tgpr/phones/word_boundary.int > exp/tri3b/final.mdl 'ark:gunzip -c > exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) > [Lattice has input epsilons and/or is not input-deterministic (in Mohri > sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow > and-or blow up in memory. > LOG (lattice-align-words:main():lattice-align-words.cc:122) Successfully > aligned 43 lattices; 0 had errors. > > resulting lattice snippet: > > VERSION=1.1 > UTTERANCE=440c0401 > N=291 L=411 > I=0 W=!NULL t=0.00 > I=1 W=!NULL t=0.88 > I=2 W=!NULL t=1.14 > I=3 W=!NULL t=1.38 > I=4 W=!NULL t=1.75 > I=5 W=!NULL t=1.38 > I=6 W=!NULL t=1.75 > I=7 W=!NULL t=0.87 > I=8 W=!NULL t=0.74 > I=9 W=!NULL t=0.77 > I=10 W=!NULL t=0.87 > I=11 W=!NULL t=1.34 > I=12 W=!NULL t=0.74 > I=13 W=!NULL t=0.77 > I=14 W=!NULL t=0.87 > I=15 W=!NULL t=1.34 > I=16 W=!NULL t=1.75 > I=17 W=!NULL t=2.08 > I=18 W=!NULL t=2.14 > I=19 W=!NULL t=2.43 > I=20 W=!NULL t=2.56 > I=21 W=!NULL t=2.11 > I=22 W=!NULL t=2.43 > I=23 W=!NULL t=2.56 > I=24 W=!NULL t=3.10 > I=25 W=!NULL t=3.25 > I=26 W=!NULL t=3.59 > I=27 W=!NULL t=3.10 > I=28 W=!NULL t=3.25 > I=29 W=!NULL t=3.59 > I=30 W=!NULL t=3.59 > I=31 W=!NULL t=3.59 > I=32 W=!NULL t=3.59 > I=33 W=!NULL t=3.59 > I=34 W=!NULL t=3.78 > I=35 W=!NULL t=4.21 > I=36 W=!NULL t=3.78 > I=37 W=!NULL t=3.99 > I=38 W=!NULL t=3.80 > I=39 W=!NULL t=3.99 > I=40 W=!NULL t=3.78 > ....... > I=245 W=73728 t=4.06 > I=246 W=4230 t=4.21 > I=247 W=93495 t=4.21 > I=248 W=140327 t=4.52 > I=249 W=140327 t=4.52 > I=250 W=62898 t=3.99 > I=251 W=72789 t=4.21 > I=252 W=72799 t=4.21 > I=253 W=73728 t=4.06 > I=254 W=4230 t=4.21 > I=255 W=93495 t=4.21 > I=256 W=140327 t=4.52 > I=257 W=132935 t=5.58 > I=258 W=140327 t=4.52 > I=259 W=132935 t=5.58 > I=260 W=140327 t=4.52 > I=261 W=140327 t=4.52 > I=262 W=4230 t=3.99 > I=263 W=72789 t=4.21 > I=264 W=72799 t=4.21 > I=265 W=74646 t=4.21 > I=266 W=140327 t=4.52 > I=267 W=132935 t=5.58 > I=268 W=140327 t=4.52 > I=269 W=132935 t=5.58 > I=270 W=140327 t=4.52 > I=271 W=132935 t=5.58 > I=272 W=124859 t=6.46 > I=273 W=4230 t=6.62 > I=274 W=99831 t=6.92 > I=275 W=43150 t=7.78 > I=276 W=18179 t=8.47 > I=277 W=46751 t=8.60 > I=278 W=25 t=8.70 > I=279 W=4086 t=8.70 > I=280 W=93075 t=8.70 > I=281 W=135497 t=9.45 > I=282 W=135497 t=9.45 > I=283 W=121952 t=9.96 > I=284 W=126807 t=9.96 > I=285 W=0 t=11.35 > I=286 W=135497 t=9.45 > I=287 W=121952 t=9.96 > I=288 W=126807 t=9.96 > I=289 W=0 t=11.35 > I=290 W=0 t=11.35 > J=0 S=0 E=86 v=0.000000 a=-6128.520000 l=-28.250300 > J=1 S=86 E=12 v=0.000000 a=0.000000 l=0.000000 > J=2 S=0 E=87 v=0.000000 a=-6128.520000 l=-14.239300 > J=3 S=87 E=8 v=0.000000 a=0.000000 l=0.000000 > J=4 S=0 E=88 v=0.000000 a=-8043.850000 l=-26.245500 > J=5 S=88 E=7 v=0.000000 a=0.000000 l=0.000000 > J=6 S=0 E=89 v=0.000000 a=-5821.820000 l=-18.684200 > J=7 S=89 E=1 v=0.000000 a=0.000000 l=0.000000 > J=8 S=1 E=90 v=0.000000 a=0.000000 l=0.000000 > J=9 S=90 E=2 v=0.000000 a=0.000000 l=0.000000 > J=10 S=2 E=91 v=0.000000 a=-2222.850000 l=-17.005600 > J=11 S=91 E=5 v=0.000000 a=0.000000 l=0.000000 > J=12 S=2 E=92 v=0.000000 a=-2222.850000 l=-12.768900 > J=13 S=92 E=3 v=0.000000 a=0.000000 l=0.000000 > J=14 S=3 E=93 v=0.000000 a=-952.576000 l=-7.415230 > J=15 S=93 E=4 v=0.000000 a=0.000000 l=0.000000 > J=16 S=4 E=94 v=0.000000 a=-4826.290000 l=-11.488200 > J=17 S=94 E=21 v=0.000000 a=0.000000 l=0.000000 > J=18 S=4 E=95 v=0.000000 a=-4888.020000 l=-9.040100 > J=19 S=95 E=17 v=0.000000 a=0.000000 l=0.000000 > J=20 S=5 E=96 v=0.000000 a=-952.576000 l=-9.439240 > J=21 S=96 E=6 v=0.000000 a=0.000000 l=0.000000 > J=22 S=6 E=97 v=0.000000 a=-4826.290000 l=-11.566900 > J=23 S=97 E=21 v=0.000000 a=0.000000 l=0.000000 > J=24 S=6 E=98 v=0.000000 a=-4888.020000 l=-9.118870 > J=25 S=98 E=17 v=0.000000 a=0.000000 l=0.000000 > J=26 S=7 E=99 v=0.000000 a=0.000000 l=0.000000 > J=27 S=99 E=15 v=0.000000 a=0.000000 l=0.000000 > J=28 S=8 E=100 v=0.000000 a=0.000000 l=0.000000 > J=29 S=100 E=9 v=0.000000 a=0.000000 l=0.000000 > J=30 S=9 E=101 v=0.000000 a=0.000000 l=0.000000 > J=31 S=101 E=10 v=0.000000 a=0.000000 l=0.000000 > J=32 S=10 E=102 v=0.000000 a=-1906.000000 l=-16.253300 > J=33 S=102 E=11 v=0.000000 a=0.000000 l=0.000000 > J=34 S=11 E=103 v=0.000000 a=-952.576000 l=-9.806090 > J=35 S=103 E=16 v=0.000000 a=0.000000 l=0.000000 > J=36 S=12 E=104 v=0.000000 a=0.000000 l=0.000000 > J=37 S=104 E=13 v=0.000000 a=0.000000 l=0.000000 > J=38 S=13 E=105 v=0.000000 a=0.000000 l=0.000000 > J=39 S=105 E=14 v=0.000000 a=0.000000 l=0.000000 > J=40 S=14 E=106 v=0.000000 a=-1906.000000 l=-16.458200 > J=41 S=106 E=15 v=0.000000 a=0.000000 l=0.000000 > J=42 S=15 E=107 v=0.000000 a=-952.576000 l=-9.806090 > J=43 S=107 E=16 v=0.000000 a=0.000000 l=0.000000 > J=44 S=16 E=108 v=0.000000 a=-4826.290000 l=-11.566900 > J=45 S=108 E=21 v=0.000000 a=0.000000 l=0.000000 > ......... > > > > > > On 04/02/2014 18:06, Vesely Karel wrote: > > Hello, > the script 'utils/convert_slf.pl' is updated with a new example how to call > it. > > Also I refactored the mode which puts links into nodes, to consider > situation when a node has incoming arcs with different words. > > The default output format is back to generate lattices with words > in links, as it is more compact. > > Karel. > > > > On 02/04/2014 05:15 PM, Karel Veselý wrote: > > Hi, > sure, recently I rewrote that script recently in cooperation with Korbinian. > I'll have a look at it, > Karel. > > > > > > Dne 4.2.2014 17:05, Daniel Povey napsal(a): > > You may have seen the message because lattice-word-align expects the model > file as its first argument; this is an error in the help message of > utils/convert_slf.pl. Also, the help messagepl should probably be updated > to use lattice-align-words as its suggested input. > Karel, you made some recent changes to that program; do you think you could > help with this? > Dan > > >> Hi, >> After running the egs/wsj/s5/run.sh I call the convert_slf.pl script on >> one of the lattices in tri3b step following an exampl in the scripts help >> output and get a deprecation message: >> >> [davidm@code7 s5]$ utils/convert_slf.pl >> Convert kaldi lattices to HTK SLF (v1.1) format. >> Usage: convert_slf.pl [options] lat-file.txt [out-dir] >> e.g. lattice-word-align 'ark:gunzip -c lat.gz |' ark,t:- | >> utils/convert_slf.pl - slf/ >> Options regarding the SLF output: >> --lmscale x LM weight (default: lmscale=1) >> --acscale x Acoustic weight (default: acscale=1) >> --wdpenalty x Word insertion penalty (default: 0) >> --framerate x Frame rate to compute timing information (default: 0.01) >> [davidm@code7 s5]$ lattice-word-align 'ark:gunzip -c >> exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | >> utils/convert_slf.pl - ~/slf/ >> lattice-word-align 'ark:gunzip -c >> exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- >> >> (note: from the s5 scripts onward, this is deprecated, see >> lattice-align-words) >> Create word-aligned lattices (in which the arcs correspond with >> word boundaries) >> Usage: lattice-word-align [options] <model> <lattice-rspecifier> >> <lattice-wspecifier> >> e.g.: lattice-word-align --silence-phones=1:2 --wbegin-phones=2:6:10:14 \ >> --wend-phones=3:7:11:15 --winternal-phones=4:8:12:16 >> --wbegin-and-end-phones=5:9:13:17 \ >> --silence-label=2 --partial-word-label=16342 \ >> final.mdl ark:1.lats ark:aligned.lats >> >> Options: >> --output-error-lats : If true, output aligned lattices even if >> there was an error (e.g. caused by forced-out lattice) (bool, default = >> true) >> --partial-word-label : Numeric id of word symbol that is to be >> used for arcs in the word-aligned lattice corresponding to partial words at >> the end of "forced-out" utterances (zero is OK) (int, default = 0) >> --reorder : True if the lattices were generated from >> graphs that had the --reorder option true, relating to reordering self-loops >> (typically true) (bool, default = true) >> --silence-has-olabels : If true, silence phones have output labels >> in the lattice, just >> like regular words. [This means you can't have un-labeled silences] >> (bool, default = false) >> --silence-label : Numeric id of word symbol that is to be >> used for silence arcs in the word-aligned lattice (zero is OK) (int, default >> = 0) >> --silence-may-be-word-internal : If true, silence may appear inside >> words' prons (but not at begin/end!) >> (bool, default = false) >> --silence-phones : Colon-separated list of numeric ids of >> phones that are used for silence (and other non-word events such as noise) >> (string, default = "") >> --test : If true, activate checks designed to test >> the code. (bool, default = false) >> --wbegin-and-end-phones : Colon-separated list of numeric ids of >> phones that are used for single-phone words. (string, default = "") >> --wbegin-phones : Colon-separated list of numeric ids of >> phones that begin a word (string, default = "") >> --wend-phones : Colon-separated list of numeric ids of >> phones that end a word (string, default = "") >> --winternal-phones : Colon-separated list of numeric ids of >> phones that are internal to a word (string, default = "") >> >> Standard options: >> --config : Configuration file with options (string, >> default = "") >> --help : Print out usage message (bool, default = >> false) >> --print-args : Print the command line arguments (to >> stderr) (bool, default = true) >> --verbose : Verbose level (higher->more logging) (int, >> default = 0) >> >> What is the right command-line now for calling the convert_slf.pl script >> on a binary lattice file(s)? >> >> Many thanks, >> David >> >> >> >> >> On 20/12/2013 10:44, Karel Veselý wrote: >> >> There is also a script directly in kaldi wsj/s5 from Korbinian Riedhammer: >> >> >> http://sourceforge.net/p/kaldi/code/HEAD/tree/trunk/egs/wsj/s5/utils/convert_slf.pl >> >> Karel >> >> Dne 19. 12. 2013 21:05, Daniel Povey napsal(a): >> >> It looks like Paul Dixon has published something he wrote, here: >> >> https://gist.github.com/edobashira/5811963 >> >> Dan >> >> >> >> On Thu, Dec 19, 2013 at 3:20 AM, wei.shi-新 <we...@im...> wrote: >>> >>> Hi all, >>> >>> Though I know the purpose of kaldi is not to be compatible with HTK, I >>> still want to ask, is there a tool that can convert kaldi lattices( e.g., >>> generated by gmm-latgen-faster ) to HTK compatible format? >>> >>> >>> Shi Wei >>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into >>> your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >>> Pro! >>> >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >> >> >> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >> Pro! >> >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> >> >> >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> >> >> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >> Pro! >> >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> >> >> >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> >> >> >> ------------------------------------------------------------------------------ >> Managing the Performance of Cloud-Based Applications >> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >> Read the Whitepaper. >> >> http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> > > > > > ------------------------------------------------------------------------------ > Managing the Performance of Cloud-Based Applications > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. > Read the Whitepaper. > http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk > > > > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > > > |
|
From: David M. <da...@ca...> - 2014-02-05 11:35:08
|
Hi Karel, Thanks for you help. Based on you new example in the convert_slf.pl script I made up a command-line that now produces HTK lattices from wsj example kaldi lattices. It gives me a warning. Is it anything to worry about? The resulting lattices have words labelled with numbers, see below. Are the numbers some indexes? How can I get words to the HTK lattice? David [davidm@code4 s5]$ lattice-align-words exp/tri3b/graph_bd_tgpr/phones/word_boundary.int exp/tri3b/final.mdl 'ark:gunzip -c exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | utils/convert_slf.pl --word-to-node - ~/slf/ lattice-align-words exp/tri3b/graph_bd_tgpr/phones/word_boundary.int exp/tri3b/final.mdl 'ark:gunzip -c exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. WARNING (lattice-align-words:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory. LOG (lattice-align-words:main():lattice-align-words.cc:122) Successfully aligned 43 lattices; 0 had errors. resulting lattice snippet: VERSION=1.1 UTTERANCE=440c0401 N=291 L=411 I=0 W=!NULL t=0.00 I=1 W=!NULL t=0.88 I=2 W=!NULL t=1.14 I=3 W=!NULL t=1.38 I=4 W=!NULL t=1.75 I=5 W=!NULL t=1.38 I=6 W=!NULL t=1.75 I=7 W=!NULL t=0.87 I=8 W=!NULL t=0.74 I=9 W=!NULL t=0.77 I=10 W=!NULL t=0.87 I=11 W=!NULL t=1.34 I=12 W=!NULL t=0.74 I=13 W=!NULL t=0.77 I=14 W=!NULL t=0.87 I=15 W=!NULL t=1.34 I=16 W=!NULL t=1.75 I=17 W=!NULL t=2.08 I=18 W=!NULL t=2.14 I=19 W=!NULL t=2.43 I=20 W=!NULL t=2.56 I=21 W=!NULL t=2.11 I=22 W=!NULL t=2.43 I=23 W=!NULL t=2.56 I=24 W=!NULL t=3.10 I=25 W=!NULL t=3.25 I=26 W=!NULL t=3.59 I=27 W=!NULL t=3.10 I=28 W=!NULL t=3.25 I=29 W=!NULL t=3.59 I=30 W=!NULL t=3.59 I=31 W=!NULL t=3.59 I=32 W=!NULL t=3.59 I=33 W=!NULL t=3.59 I=34 W=!NULL t=3.78 I=35 W=!NULL t=4.21 I=36 W=!NULL t=3.78 I=37 W=!NULL t=3.99 I=38 W=!NULL t=3.80 I=39 W=!NULL t=3.99 I=40 W=!NULL t=3.78 ....... I=245 W=73728 t=4.06 I=246 W=4230 t=4.21 I=247 W=93495 t=4.21 I=248 W=140327 t=4.52 I=249 W=140327 t=4.52 I=250 W=62898 t=3.99 I=251 W=72789 t=4.21 I=252 W=72799 t=4.21 I=253 W=73728 t=4.06 I=254 W=4230 t=4.21 I=255 W=93495 t=4.21 I=256 W=140327 t=4.52 I=257 W=132935 t=5.58 I=258 W=140327 t=4.52 I=259 W=132935 t=5.58 I=260 W=140327 t=4.52 I=261 W=140327 t=4.52 I=262 W=4230 t=3.99 I=263 W=72789 t=4.21 I=264 W=72799 t=4.21 I=265 W=74646 t=4.21 I=266 W=140327 t=4.52 I=267 W=132935 t=5.58 I=268 W=140327 t=4.52 I=269 W=132935 t=5.58 I=270 W=140327 t=4.52 I=271 W=132935 t=5.58 I=272 W=124859 t=6.46 I=273 W=4230 t=6.62 I=274 W=99831 t=6.92 I=275 W=43150 t=7.78 I=276 W=18179 t=8.47 I=277 W=46751 t=8.60 I=278 W=25 t=8.70 I=279 W=4086 t=8.70 I=280 W=93075 t=8.70 I=281 W=135497 t=9.45 I=282 W=135497 t=9.45 I=283 W=121952 t=9.96 I=284 W=126807 t=9.96 I=285 W=0 t=11.35 I=286 W=135497 t=9.45 I=287 W=121952 t=9.96 I=288 W=126807 t=9.96 I=289 W=0 t=11.35 I=290 W=0 t=11.35 J=0 S=0 E=86 v=0.000000 a=-6128.520000 l=-28.250300 J=1 S=86 E=12 v=0.000000 a=0.000000 l=0.000000 J=2 S=0 E=87 v=0.000000 a=-6128.520000 l=-14.239300 J=3 S=87 E=8 v=0.000000 a=0.000000 l=0.000000 J=4 S=0 E=88 v=0.000000 a=-8043.850000 l=-26.245500 J=5 S=88 E=7 v=0.000000 a=0.000000 l=0.000000 J=6 S=0 E=89 v=0.000000 a=-5821.820000 l=-18.684200 J=7 S=89 E=1 v=0.000000 a=0.000000 l=0.000000 J=8 S=1 E=90 v=0.000000 a=0.000000 l=0.000000 J=9 S=90 E=2 v=0.000000 a=0.000000 l=0.000000 J=10 S=2 E=91 v=0.000000 a=-2222.850000 l=-17.005600 J=11 S=91 E=5 v=0.000000 a=0.000000 l=0.000000 J=12 S=2 E=92 v=0.000000 a=-2222.850000 l=-12.768900 J=13 S=92 E=3 v=0.000000 a=0.000000 l=0.000000 J=14 S=3 E=93 v=0.000000 a=-952.576000 l=-7.415230 J=15 S=93 E=4 v=0.000000 a=0.000000 l=0.000000 J=16 S=4 E=94 v=0.000000 a=-4826.290000 l=-11.488200 J=17 S=94 E=21 v=0.000000 a=0.000000 l=0.000000 J=18 S=4 E=95 v=0.000000 a=-4888.020000 l=-9.040100 J=19 S=95 E=17 v=0.000000 a=0.000000 l=0.000000 J=20 S=5 E=96 v=0.000000 a=-952.576000 l=-9.439240 J=21 S=96 E=6 v=0.000000 a=0.000000 l=0.000000 J=22 S=6 E=97 v=0.000000 a=-4826.290000 l=-11.566900 J=23 S=97 E=21 v=0.000000 a=0.000000 l=0.000000 J=24 S=6 E=98 v=0.000000 a=-4888.020000 l=-9.118870 J=25 S=98 E=17 v=0.000000 a=0.000000 l=0.000000 J=26 S=7 E=99 v=0.000000 a=0.000000 l=0.000000 J=27 S=99 E=15 v=0.000000 a=0.000000 l=0.000000 J=28 S=8 E=100 v=0.000000 a=0.000000 l=0.000000 J=29 S=100 E=9 v=0.000000 a=0.000000 l=0.000000 J=30 S=9 E=101 v=0.000000 a=0.000000 l=0.000000 J=31 S=101 E=10 v=0.000000 a=0.000000 l=0.000000 J=32 S=10 E=102 v=0.000000 a=-1906.000000 l=-16.253300 J=33 S=102 E=11 v=0.000000 a=0.000000 l=0.000000 J=34 S=11 E=103 v=0.000000 a=-952.576000 l=-9.806090 J=35 S=103 E=16 v=0.000000 a=0.000000 l=0.000000 J=36 S=12 E=104 v=0.000000 a=0.000000 l=0.000000 J=37 S=104 E=13 v=0.000000 a=0.000000 l=0.000000 J=38 S=13 E=105 v=0.000000 a=0.000000 l=0.000000 J=39 S=105 E=14 v=0.000000 a=0.000000 l=0.000000 J=40 S=14 E=106 v=0.000000 a=-1906.000000 l=-16.458200 J=41 S=106 E=15 v=0.000000 a=0.000000 l=0.000000 J=42 S=15 E=107 v=0.000000 a=-952.576000 l=-9.806090 J=43 S=107 E=16 v=0.000000 a=0.000000 l=0.000000 J=44 S=16 E=108 v=0.000000 a=-4826.290000 l=-11.566900 J=45 S=108 E=21 v=0.000000 a=0.000000 l=0.000000 ......... On 04/02/2014 18:06, Vesely Karel wrote: > Hello, > the script 'utils/convert_slf.pl' is updated with a new example how to > call it. > > Also I refactored the mode which puts links into nodes, to consider > situation when a node has incoming arcs with different words. > > The default output format is back to generate lattices with words > in links, as it is more compact. > > Karel. > > > > On 02/04/2014 05:15 PM, Karel Veselý wrote: >> Hi, >> sure, recently I rewrote that script recently in cooperation with >> Korbinian. I'll have a look at it, >> Karel. >> >> >> >> >> >> Dne 4.2.2014 17:05, Daniel Povey napsal(a): >>> You may have seen the message because lattice-word-align expects the >>> model file as its first argument; this is an error in the help >>> message of utils/convert_slf.pl <http://convert_slf.pl>. Also, the >>> help messagepl should probably be updated to use lattice-align-words >>> as its suggested input. >>> Karel, you made some recent changes to that program; do you think >>> you could help with this? >>> Dan >>> >>> >>> Hi, >>> After running the egs/wsj/s5/run.sh I call the convert_slf.pl >>> <http://convert_slf.pl> script on one of the lattices in tri3b >>> step following an exampl in the scripts help output and get a >>> deprecation message: >>> >>> [davidm@code7 s5]$ utils/convert_slf.pl <http://convert_slf.pl> >>> Convert kaldi lattices to HTK SLF (v1.1) format. >>> Usage: convert_slf.pl <http://convert_slf.pl> [options] >>> lat-file.txt [out-dir] >>> e.g. lattice-word-align 'ark:gunzip -c lat.gz |' ark,t:- | >>> utils/convert_slf.pl <http://convert_slf.pl> - slf/ >>> Options regarding the SLF output: >>> --lmscale x LM weight (default: lmscale=1) >>> --acscale x Acoustic weight (default: acscale=1) >>> --wdpenalty x Word insertion penalty (default: 0) >>> --framerate x Frame rate to compute timing information >>> (default: 0.01) >>> [davidm@code7 s5]$ lattice-word-align 'ark:gunzip -c >>> exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | >>> utils/convert_slf.pl <http://convert_slf.pl> - ~/slf/ >>> lattice-word-align 'ark:gunzip -c >>> exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- >>> >>> (note: from the s5 scripts onward, this is deprecated, see >>> lattice-align-words) >>> Create word-aligned lattices (in which the arcs correspond with >>> word boundaries) >>> Usage: lattice-word-align [options] <model> <lattice-rspecifier> >>> <lattice-wspecifier> >>> e.g.: lattice-word-align --silence-phones=1:2 >>> --wbegin-phones=2:6:10:14 \ >>> --wend-phones=3:7:11:15 --winternal-phones=4:8:12:16 >>> --wbegin-and-end-phones=5:9:13:17 \ >>> --silence-label=2 --partial-word-label=16342 \ >>> final.mdl ark:1.lats ark:aligned.lats >>> >>> Options: >>> --output-error-lats : If true, output aligned lattices >>> even if there was an error (e.g. caused by forced-out lattice) >>> (bool, default = true) >>> --partial-word-label : Numeric id of word symbol that >>> is to be used for arcs in the word-aligned lattice corresponding >>> to partial words at the end of "forced-out" utterances (zero is >>> OK) (int, default = 0) >>> --reorder : True if the lattices were >>> generated from graphs that had the --reorder option true, >>> relating to reordering self-loops (typically true) (bool, >>> default = true) >>> --silence-has-olabels : If true, silence phones have >>> output labels in the lattice, just >>> like regular words. [This means you can't have un-labeled >>> silences] (bool, default = false) >>> --silence-label : Numeric id of word symbol that >>> is to be used for silence arcs in the word-aligned lattice (zero >>> is OK) (int, default = 0) >>> --silence-may-be-word-internal : If true, silence may appear >>> inside words' prons (but not at begin/end!) >>> (bool, default = false) >>> --silence-phones : Colon-separated list of numeric >>> ids of phones that are used for silence (and other non-word >>> events such as noise) (string, default = "") >>> --test : If true, activate checks >>> designed to test the code. (bool, default = false) >>> --wbegin-and-end-phones : Colon-separated list of numeric >>> ids of phones that are used for single-phone words. (string, >>> default = "") >>> --wbegin-phones : Colon-separated list of numeric >>> ids of phones that begin a word (string, default = "") >>> --wend-phones : Colon-separated list of numeric >>> ids of phones that end a word (string, default = "") >>> --winternal-phones : Colon-separated list of numeric >>> ids of phones that are internal to a word (string, default = "") >>> >>> Standard options: >>> --config : Configuration file with options >>> (string, default = "") >>> --help : Print out usage message (bool, >>> default = false) >>> --print-args : Print the command line arguments >>> (to stderr) (bool, default = true) >>> --verbose : Verbose level (higher->more >>> logging) (int, default = 0) >>> >>> What is the right command-line now for calling the >>> convert_slf.pl <http://convert_slf.pl> script on a binary >>> lattice file(s)? >>> >>> Many thanks, >>> David >>> >>> >>> >>> >>> On 20/12/2013 10:44, Karel Veselý wrote: >>>> There is also a script directly in kaldi wsj/s5 from Korbinian >>>> Riedhammer: >>>> >>>> http://sourceforge.net/p/kaldi/code/HEAD/tree/trunk/egs/wsj/s5/utils/convert_slf.pl >>>> >>>> Karel >>>> >>>> Dne 19. 12. 2013 21:05, Daniel Povey napsal(a): >>>>> It looks like Paul Dixon has published something he wrote, here: >>>>> >>>>> https://gist.github.com/edobashira/5811963 >>>>> >>>>> Dan >>>>> >>>>> >>>>> >>>>> On Thu, Dec 19, 2013 at 3:20 AM, wei.shi-? >>>>> <we...@im... <mailto:we...@im...>> wrote: >>>>> >>>>> Hi all, >>>>> Though I know the purpose of kaldi is not to be compatible >>>>> with HTK, I still want to ask, is there a tool that can >>>>> convert kaldi lattices( e.g., generated by >>>>> gmm-latgen-faster ) to HTK compatible format? >>>>> Shi Wei >>>>> >>>>> ------------------------------------------------------------------------------ >>>>> Rapidly troubleshoot problems before they affect your >>>>> business. Most IT >>>>> organizations don't have a clear picture of how >>>>> application performance >>>>> affects their revenue. With AppDynamics, you get 100% >>>>> visibility into your >>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL >>>>> of AppDynamics Pro! >>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>> _______________________________________________ >>>>> Kaldi-users mailing list >>>>> Kal...@li... >>>>> <mailto:Kal...@li...> >>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>> >>>>> >>>>> >>>>> >>>>> ------------------------------------------------------------------------------ >>>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>>> organizations don't have a clear picture of how application performance >>>>> affects their revenue. With AppDynamics, you get 100% visibility into your >>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! >>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>> >>>>> >>>>> _______________________________________________ >>>>> Kaldi-users mailing list >>>>> Kal...@li... <mailto:Kal...@li...> >>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>> organizations don't have a clear picture of how application performance >>>> affects their revenue. With AppDynamics, you get 100% visibility into your >>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! >>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>> >>>> >>>> _______________________________________________ >>>> Kaldi-users mailing list >>>> Kal...@li... <mailto:Kal...@li...> >>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >>> >>> ------------------------------------------------------------------------------ >>> Managing the Performance of Cloud-Based Applications >>> Take advantage of what the Cloud has to offer - Avoid Common >>> Pitfalls. >>> Read the Whitepaper. >>> http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> <mailto:Kal...@li...> >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >>> >> >> >> >> ------------------------------------------------------------------------------ >> Managing the Performance of Cloud-Based Applications >> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >> Read the Whitepaper. >> http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk >> >> >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users > |
|
From: Vesely K. <ive...@fi...> - 2014-02-04 18:06:27
|
Hello, the script 'utils/convert_slf.pl' is updated with a new example how to call it. Also I refactored the mode which puts links into nodes, to consider situation when a node has incoming arcs with different words. The default output format is back to generate lattices with words in links, as it is more compact. Karel. On 02/04/2014 05:15 PM, Karel Veselý wrote: > Hi, > sure, recently I rewrote that script recently in cooperation with > Korbinian. I'll have a look at it, > Karel. > > > > > > Dne 4.2.2014 17:05, Daniel Povey napsal(a): >> You may have seen the message because lattice-word-align expects the >> model file as its first argument; this is an error in the help >> message of utils/convert_slf.pl <http://convert_slf.pl>. Also, the >> help messagepl should probably be updated to use lattice-align-words >> as its suggested input. >> Karel, you made some recent changes to that program; do you think you >> could help with this? >> Dan >> >> >> Hi, >> After running the egs/wsj/s5/run.sh I call the convert_slf.pl >> <http://convert_slf.pl> script on one of the lattices in tri3b >> step following an exampl in the scripts help output and get a >> deprecation message: >> >> [davidm@code7 s5]$ utils/convert_slf.pl <http://convert_slf.pl> >> Convert kaldi lattices to HTK SLF (v1.1) format. >> Usage: convert_slf.pl <http://convert_slf.pl> [options] >> lat-file.txt [out-dir] >> e.g. lattice-word-align 'ark:gunzip -c lat.gz |' ark,t:- | >> utils/convert_slf.pl <http://convert_slf.pl> - slf/ >> Options regarding the SLF output: >> --lmscale x LM weight (default: lmscale=1) >> --acscale x Acoustic weight (default: acscale=1) >> --wdpenalty x Word insertion penalty (default: 0) >> --framerate x Frame rate to compute timing information >> (default: 0.01) >> [davidm@code7 s5]$ lattice-word-align 'ark:gunzip -c >> exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | >> utils/convert_slf.pl <http://convert_slf.pl> - ~/slf/ >> lattice-word-align 'ark:gunzip -c >> exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- >> >> (note: from the s5 scripts onward, this is deprecated, see >> lattice-align-words) >> Create word-aligned lattices (in which the arcs correspond with >> word boundaries) >> Usage: lattice-word-align [options] <model> <lattice-rspecifier> >> <lattice-wspecifier> >> e.g.: lattice-word-align --silence-phones=1:2 >> --wbegin-phones=2:6:10:14 \ >> --wend-phones=3:7:11:15 --winternal-phones=4:8:12:16 >> --wbegin-and-end-phones=5:9:13:17 \ >> --silence-label=2 --partial-word-label=16342 \ >> final.mdl ark:1.lats ark:aligned.lats >> >> Options: >> --output-error-lats : If true, output aligned lattices >> even if there was an error (e.g. caused by forced-out lattice) >> (bool, default = true) >> --partial-word-label : Numeric id of word symbol that is >> to be used for arcs in the word-aligned lattice corresponding to >> partial words at the end of "forced-out" utterances (zero is OK) >> (int, default = 0) >> --reorder : True if the lattices were >> generated from graphs that had the --reorder option true, >> relating to reordering self-loops (typically true) (bool, default >> = true) >> --silence-has-olabels : If true, silence phones have >> output labels in the lattice, just >> like regular words. [This means you can't have un-labeled >> silences] (bool, default = false) >> --silence-label : Numeric id of word symbol that is >> to be used for silence arcs in the word-aligned lattice (zero is >> OK) (int, default = 0) >> --silence-may-be-word-internal : If true, silence may appear >> inside words' prons (but not at begin/end!) >> (bool, default = false) >> --silence-phones : Colon-separated list of numeric >> ids of phones that are used for silence (and other non-word >> events such as noise) (string, default = "") >> --test : If true, activate checks designed >> to test the code. (bool, default = false) >> --wbegin-and-end-phones : Colon-separated list of numeric >> ids of phones that are used for single-phone words. (string, >> default = "") >> --wbegin-phones : Colon-separated list of numeric >> ids of phones that begin a word (string, default = "") >> --wend-phones : Colon-separated list of numeric >> ids of phones that end a word (string, default = "") >> --winternal-phones : Colon-separated list of numeric >> ids of phones that are internal to a word (string, default = "") >> >> Standard options: >> --config : Configuration file with options >> (string, default = "") >> --help : Print out usage message (bool, >> default = false) >> --print-args : Print the command line arguments >> (to stderr) (bool, default = true) >> --verbose : Verbose level (higher->more >> logging) (int, default = 0) >> >> What is the right command-line now for calling the convert_slf.pl >> <http://convert_slf.pl> script on a binary lattice file(s)? >> >> Many thanks, >> David >> >> >> >> >> On 20/12/2013 10:44, Karel Veselý wrote: >>> There is also a script directly in kaldi wsj/s5 from Korbinian >>> Riedhammer: >>> >>> http://sourceforge.net/p/kaldi/code/HEAD/tree/trunk/egs/wsj/s5/utils/convert_slf.pl >>> >>> Karel >>> >>> Dne 19. 12. 2013 21:05, Daniel Povey napsal(a): >>>> It looks like Paul Dixon has published something he wrote, here: >>>> >>>> https://gist.github.com/edobashira/5811963 >>>> >>>> Dan >>>> >>>> >>>> >>>> On Thu, Dec 19, 2013 at 3:20 AM, wei.shi-? <we...@im... >>>> <mailto:we...@im...>> wrote: >>>> >>>> Hi all, >>>> Though I know the purpose of kaldi is not to be compatible >>>> with HTK, I still want to ask, is there a tool that can >>>> convert kaldi lattices( e.g., generated by >>>> gmm-latgen-faster ) to HTK compatible format? >>>> Shi Wei >>>> >>>> ------------------------------------------------------------------------------ >>>> Rapidly troubleshoot problems before they affect your >>>> business. Most IT >>>> organizations don't have a clear picture of how application >>>> performance >>>> affects their revenue. With AppDynamics, you get 100% >>>> visibility into your >>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL >>>> of AppDynamics Pro! >>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>> _______________________________________________ >>>> Kaldi-users mailing list >>>> Kal...@li... >>>> <mailto:Kal...@li...> >>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>> >>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>> organizations don't have a clear picture of how application performance >>>> affects their revenue. With AppDynamics, you get 100% visibility into your >>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! >>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>> >>>> >>>> _______________________________________________ >>>> Kaldi-users mailing list >>>> Kal...@li... <mailto:Kal...@li...> >>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> >>> >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... <mailto:Kal...@li...> >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> >> ------------------------------------------------------------------------------ >> Managing the Performance of Cloud-Based Applications >> Take advantage of what the Cloud has to offer - Avoid Common >> Pitfalls. >> Read the Whitepaper. >> http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> <mailto:Kal...@li...> >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> > > > > ------------------------------------------------------------------------------ > Managing the Performance of Cloud-Based Applications > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. > Read the Whitepaper. > http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk > > > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users |
|
From: Karel V. <ive...@fi...> - 2014-02-04 16:22:38
|
Hi, sure, recently I rewrote that script recently in cooperation with Korbinian. I'll have a look at it, Karel. Dne 4.2.2014 17:05, Daniel Povey napsal(a): > You may have seen the message because lattice-word-align expects the > model file as its first argument; this is an error in the help message > of utils/convert_slf.pl <http://convert_slf.pl>. Also, the help > messagepl should probably be updated to use lattice-align-words as its > suggested input. > Karel, you made some recent changes to that program; do you think you > could help with this? > Dan > > > Hi, > After running the egs/wsj/s5/run.sh I call the convert_slf.pl > <http://convert_slf.pl> script on one of the lattices in tri3b > step following an exampl in the scripts help output and get a > deprecation message: > > [davidm@code7 s5]$ utils/convert_slf.pl <http://convert_slf.pl> > Convert kaldi lattices to HTK SLF (v1.1) format. > Usage: convert_slf.pl <http://convert_slf.pl> [options] > lat-file.txt [out-dir] > e.g. lattice-word-align 'ark:gunzip -c lat.gz |' ark,t:- | > utils/convert_slf.pl <http://convert_slf.pl> - slf/ > Options regarding the SLF output: > --lmscale x LM weight (default: lmscale=1) > --acscale x Acoustic weight (default: acscale=1) > --wdpenalty x Word insertion penalty (default: 0) > --framerate x Frame rate to compute timing information > (default: 0.01) > [davidm@code7 s5]$ lattice-word-align 'ark:gunzip -c > exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | > utils/convert_slf.pl <http://convert_slf.pl> - ~/slf/ > lattice-word-align 'ark:gunzip -c > exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- > > (note: from the s5 scripts onward, this is deprecated, see > lattice-align-words) > Create word-aligned lattices (in which the arcs correspond with > word boundaries) > Usage: lattice-word-align [options] <model> <lattice-rspecifier> > <lattice-wspecifier> > e.g.: lattice-word-align --silence-phones=1:2 > --wbegin-phones=2:6:10:14 \ > --wend-phones=3:7:11:15 --winternal-phones=4:8:12:16 > --wbegin-and-end-phones=5:9:13:17 \ > --silence-label=2 --partial-word-label=16342 \ > final.mdl ark:1.lats ark:aligned.lats > > Options: > --output-error-lats : If true, output aligned lattices > even if there was an error (e.g. caused by forced-out lattice) > (bool, default = true) > --partial-word-label : Numeric id of word symbol that is > to be used for arcs in the word-aligned lattice corresponding to > partial words at the end of "forced-out" utterances (zero is OK) > (int, default = 0) > --reorder : True if the lattices were > generated from graphs that had the --reorder option true, relating > to reordering self-loops (typically true) (bool, default = true) > --silence-has-olabels : If true, silence phones have > output labels in the lattice, just > like regular words. [This means you can't have un-labeled > silences] (bool, default = false) > --silence-label : Numeric id of word symbol that is > to be used for silence arcs in the word-aligned lattice (zero is > OK) (int, default = 0) > --silence-may-be-word-internal : If true, silence may appear > inside words' prons (but not at begin/end!) > (bool, default = false) > --silence-phones : Colon-separated list of numeric > ids of phones that are used for silence (and other non-word events > such as noise) (string, default = "") > --test : If true, activate checks designed > to test the code. (bool, default = false) > --wbegin-and-end-phones : Colon-separated list of numeric > ids of phones that are used for single-phone words. (string, > default = "") > --wbegin-phones : Colon-separated list of numeric > ids of phones that begin a word (string, default = "") > --wend-phones : Colon-separated list of numeric > ids of phones that end a word (string, default = "") > --winternal-phones : Colon-separated list of numeric > ids of phones that are internal to a word (string, default = "") > > Standard options: > --config : Configuration file with options > (string, default = "") > --help : Print out usage message (bool, > default = false) > --print-args : Print the command line arguments > (to stderr) (bool, default = true) > --verbose : Verbose level (higher->more > logging) (int, default = 0) > > What is the right command-line now for calling the convert_slf.pl > <http://convert_slf.pl> script on a binary lattice file(s)? > > Many thanks, > David > > > > > On 20/12/2013 10:44, Karel Veselý wrote: >> There is also a script directly in kaldi wsj/s5 from Korbinian >> Riedhammer: >> >> http://sourceforge.net/p/kaldi/code/HEAD/tree/trunk/egs/wsj/s5/utils/convert_slf.pl >> >> Karel >> >> Dne 19. 12. 2013 21:05, Daniel Povey napsal(a): >>> It looks like Paul Dixon has published something he wrote, here: >>> >>> https://gist.github.com/edobashira/5811963 >>> >>> Dan >>> >>> >>> >>> On Thu, Dec 19, 2013 at 3:20 AM, wei.shi-新 <we...@im... >>> <mailto:we...@im...>> wrote: >>> >>> Hi all, >>> Though I know the purpose of kaldi is not to be compatible >>> with HTK, I still want to ask, is there a tool that can >>> convert kaldi lattices( e.g., generated by >>> gmm-latgen-faster ) to HTK compatible format? >>> Shi Wei >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your >>> business. Most IT >>> organizations don't have a clear picture of how application >>> performance >>> affects their revenue. With AppDynamics, you get 100% >>> visibility into your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL >>> of AppDynamics Pro! >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> <mailto:Kal...@li...> >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> >>> >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... <mailto:Kal...@li...> >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> >> >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... <mailto:Kal...@li...> >> https://lists.sourceforge.net/lists/listinfo/kaldi-users > > > ------------------------------------------------------------------------------ > Managing the Performance of Cloud-Based Applications > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. > Read the Whitepaper. > http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > <mailto:Kal...@li...> > https://lists.sourceforge.net/lists/listinfo/kaldi-users > > |
|
From: Daniel P. <dp...@gm...> - 2014-02-04 16:05:55
|
You may have seen the message because lattice-word-align expects the model file as its first argument; this is an error in the help message of utils/ convert_slf.pl. Also, the help messagepl should probably be updated to use lattice-align-words as its suggested input. Karel, you made some recent changes to that program; do you think you could help with this? Dan Hi, > After running the egs/wsj/s5/run.sh I call the convert_slf.pl script on > one of the lattices in tri3b step following an exampl in the scripts help > output and get a deprecation message: > > [davidm@code7 s5]$ utils/convert_slf.pl > Convert kaldi lattices to HTK SLF (v1.1) format. > Usage: convert_slf.pl [options] lat-file.txt [out-dir] > e.g. lattice-word-align 'ark:gunzip -c lat.gz |' ark,t:- | utils/ > convert_slf.pl - slf/ > Options regarding the SLF output: > --lmscale x LM weight (default: lmscale=1) > --acscale x Acoustic weight (default: acscale=1) > --wdpenalty x Word insertion penalty (default: 0) > --framerate x Frame rate to compute timing information (default: 0.01) > [davidm@code7 s5]$ lattice-word-align 'ark:gunzip -c > exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- | utils/ > convert_slf.pl - ~/slf/ > lattice-word-align 'ark:gunzip -c > exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- > > (note: from the s5 scripts onward, this is deprecated, see > lattice-align-words) > Create word-aligned lattices (in which the arcs correspond with > word boundaries) > Usage: lattice-word-align [options] <model> <lattice-rspecifier> > <lattice-wspecifier> > e.g.: lattice-word-align --silence-phones=1:2 --wbegin-phones=2:6:10:14 \ > --wend-phones=3:7:11:15 --winternal-phones=4:8:12:16 > --wbegin-and-end-phones=5:9:13:17 \ > --silence-label=2 --partial-word-label=16342 \ > final.mdl ark:1.lats ark:aligned.lats > > Options: > --output-error-lats : If true, output aligned lattices even if > there was an error (e.g. caused by forced-out lattice) (bool, default = > true) > --partial-word-label : Numeric id of word symbol that is to be > used for arcs in the word-aligned lattice corresponding to partial words at > the end of "forced-out" utterances (zero is OK) (int, default = 0) > --reorder : True if the lattices were generated from > graphs that had the --reorder option true, relating to reordering > self-loops (typically true) (bool, default = true) > --silence-has-olabels : If true, silence phones have output labels > in the lattice, just > like regular words. [This means you can't have un-labeled silences] > (bool, default = false) > --silence-label : Numeric id of word symbol that is to be > used for silence arcs in the word-aligned lattice (zero is OK) (int, > default = 0) > --silence-may-be-word-internal : If true, silence may appear inside > words' prons (but not at begin/end!) > (bool, default = false) > --silence-phones : Colon-separated list of numeric ids of > phones that are used for silence (and other non-word events such as noise) > (string, default = "") > --test : If true, activate checks designed to test > the code. (bool, default = false) > --wbegin-and-end-phones : Colon-separated list of numeric ids of > phones that are used for single-phone words. (string, default = "") > --wbegin-phones : Colon-separated list of numeric ids of > phones that begin a word (string, default = "") > --wend-phones : Colon-separated list of numeric ids of > phones that end a word (string, default = "") > --winternal-phones : Colon-separated list of numeric ids of > phones that are internal to a word (string, default = "") > > Standard options: > --config : Configuration file with options (string, > default = "") > --help : Print out usage message (bool, default = > false) > --print-args : Print the command line arguments (to > stderr) (bool, default = true) > --verbose : Verbose level (higher->more logging) (int, > default = 0) > > What is the right command-line now for calling the convert_slf.pl script > on a binary lattice file(s)? > > Many thanks, > David > > > > > On 20/12/2013 10:44, Karel Veselý wrote: > > There is also a script directly in kaldi wsj/s5 from Korbinian Riedhammer: > > > http://sourceforge.net/p/kaldi/code/HEAD/tree/trunk/egs/wsj/s5/utils/convert_slf.pl > > Karel > > Dne 19. 12. 2013 21:05, Daniel Povey napsal(a): > > It looks like Paul Dixon has published something he wrote, here: > > https://gist.github.com/edobashira/5811963 > > Dan > > > > On Thu, Dec 19, 2013 at 3:20 AM, wei.shi-新 <we...@im...> wrote: > >> Hi all, >> >> Though I know the purpose of kaldi is not to be compatible with HTK, I >> still want to ask, is there a tool that can convert kaldi lattices( e.g., >> generated by gmm-latgen-faster ) to HTK compatible format? >> >> >> Shi Wei >> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >> Pro! >> >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> > > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > > > > _______________________________________________ > Kaldi-users mailing lis...@li...://lists.sourceforge.net/lists/listinfo/kaldi-users > > > > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > > > > _______________________________________________ > Kaldi-users mailing lis...@li...://lists.sourceforge.net/lists/listinfo/kaldi-users > > > > > ------------------------------------------------------------------------------ > Managing the Performance of Cloud-Based Applications > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. > Read the Whitepaper. > > http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > > |
|
From: David M. <da...@ca...> - 2014-02-04 13:20:16
|
Hi,
After running the egs/wsj/s5/run.sh I call the convert_slf.pl script on
one of the lattices in tri3b step following an exampl in the scripts
help output and get a deprecation message:
[davidm@code7 s5]$ utils/convert_slf.pl
Convert kaldi lattices to HTK SLF (v1.1) format.
Usage: convert_slf.pl [options] lat-file.txt [out-dir]
e.g. lattice-word-align 'ark:gunzip -c lat.gz |' ark,t:- |
utils/convert_slf.pl - slf/
Options regarding the SLF output:
--lmscale x LM weight (default: lmscale=1)
--acscale x Acoustic weight (default: acscale=1)
--wdpenalty x Word insertion penalty (default: 0)
--framerate x Frame rate to compute timing information (default: 0.01)
[davidm@code7 s5]$ lattice-word-align 'ark:gunzip -c
exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:- |
utils/convert_slf.pl - ~/slf/
lattice-word-align 'ark:gunzip -c
exp/tri3b/decode_bd_tgpr_eval92_fg/lat.1.gz |' ark,t:-
(note: from the s5 scripts onward, this is deprecated, see
lattice-align-words)
Create word-aligned lattices (in which the arcs correspond with
word boundaries)
Usage: lattice-word-align [options] <model> <lattice-rspecifier>
<lattice-wspecifier>
e.g.: lattice-word-align --silence-phones=1:2 --wbegin-phones=2:6:10:14 \
--wend-phones=3:7:11:15 --winternal-phones=4:8:12:16
--wbegin-and-end-phones=5:9:13:17 \
--silence-label=2 --partial-word-label=16342 \
final.mdl ark:1.lats ark:aligned.lats
Options:
--output-error-lats : If true, output aligned lattices even
if there was an error (e.g. caused by forced-out lattice) (bool, default
= true)
--partial-word-label : Numeric id of word symbol that is to be
used for arcs in the word-aligned lattice corresponding to partial words
at the end of "forced-out" utterances (zero is OK) (int, default = 0)
--reorder : True if the lattices were generated
from graphs that had the --reorder option true, relating to reordering
self-loops (typically true) (bool, default = true)
--silence-has-olabels : If true, silence phones have output
labels in the lattice, just
like regular words. [This means you can't have un-labeled silences]
(bool, default = false)
--silence-label : Numeric id of word symbol that is to be
used for silence arcs in the word-aligned lattice (zero is OK) (int,
default = 0)
--silence-may-be-word-internal : If true, silence may appear inside
words' prons (but not at begin/end!)
(bool, default = false)
--silence-phones : Colon-separated list of numeric ids of
phones that are used for silence (and other non-word events such as
noise) (string, default = "")
--test : If true, activate checks designed to
test the code. (bool, default = false)
--wbegin-and-end-phones : Colon-separated list of numeric ids of
phones that are used for single-phone words. (string, default = "")
--wbegin-phones : Colon-separated list of numeric ids of
phones that begin a word (string, default = "")
--wend-phones : Colon-separated list of numeric ids of
phones that end a word (string, default = "")
--winternal-phones : Colon-separated list of numeric ids of
phones that are internal to a word (string, default = "")
Standard options:
--config : Configuration file with options
(string, default = "")
--help : Print out usage message (bool, default
= false)
--print-args : Print the command line arguments (to
stderr) (bool, default = true)
--verbose : Verbose level (higher->more logging)
(int, default = 0)
What is the right command-line now for calling the convert_slf.pl script
on a binary lattice file(s)?
Many thanks,
David
On 20/12/2013 10:44, Karel Veselý wrote:
> There is also a script directly in kaldi wsj/s5 from Korbinian Riedhammer:
>
> http://sourceforge.net/p/kaldi/code/HEAD/tree/trunk/egs/wsj/s5/utils/convert_slf.pl
>
> Karel
>
> Dne 19. 12. 2013 21:05, Daniel Povey napsal(a):
>> It looks like Paul Dixon has published something he wrote, here:
>>
>> https://gist.github.com/edobashira/5811963
>>
>> Dan
>>
>>
>>
>> On Thu, Dec 19, 2013 at 3:20 AM, wei.shi-? <we...@im...
>> <mailto:we...@im...>> wrote:
>>
>> Hi all,
>> Though I know the purpose of kaldi is not to be compatible with
>> HTK, I still want to ask, is there a tool that can convert kaldi
>> lattices( e.g., generated by gmm-latgen-faster ) to HTK
>> compatible format?
>> Shi Wei
>>
>> ------------------------------------------------------------------------------
>> Rapidly troubleshoot problems before they affect your business.
>> Most IT
>> organizations don't have a clear picture of how application
>> performance
>> affects their revenue. With AppDynamics, you get 100% visibility
>> into your
>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of
>> AppDynamics Pro!
>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Kaldi-users mailing list
>> Kal...@li...
>> <mailto:Kal...@li...>
>> https://lists.sourceforge.net/lists/listinfo/kaldi-users
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Rapidly troubleshoot problems before they affect your business. Most IT
>> organizations don't have a clear picture of how application performance
>> affects their revenue. With AppDynamics, you get 100% visibility into your
>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
>>
>>
>> _______________________________________________
>> Kaldi-users mailing list
>> Kal...@li...
>> https://lists.sourceforge.net/lists/listinfo/kaldi-users
>
>
>
> ------------------------------------------------------------------------------
> Rapidly troubleshoot problems before they affect your business. Most IT
> organizations don't have a clear picture of how application performance
> affects their revenue. With AppDynamics, you get 100% visibility into your
> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
>
>
> _______________________________________________
> Kaldi-users mailing list
> Kal...@li...
> https://lists.sourceforge.net/lists/listinfo/kaldi-users
|
|
From: dophist <do...@gm...> - 2014-01-20 03:03:51
|
Hi all, I understand kaldi uses viterbi instead of Baum-welch, and ali-to-post only produces "dummy" state-posterior with unity 1. so I'm wondering if it is possible to generate state-level posterior during force alignment? best regards, Jiayu |
|
From: Josef N. <jos...@gm...> - 2014-01-07 11:20:41
|
Hi, (duplicate since I did not reply to the list the first time) This tool was not designed for and is almost certainly not compatible Kaldi. Kaldi employs some additional symbol management for the backoffs, which are not supported in my toolchain (I made it before Kaldi existed as a public project). You will have similar problems if you try to use the OpenGRM ngram tools without modification (but the google tool does the best hole-plugging). In general, default optimization commands (det, min, etc.) do not guarantee anything regarding symbol synchronization. I believe Kaldi does some extra work in this regard. Using positional phonemes can make it easier to retrieve the boundaries (I think there are some tools in Kaldi to do this after the fact). Probably it is not appropriate for this list since it is not compatible with Kaldi. -Joe On Tue, Jan 7, 2014 at 10:22 AM, E <oth...@ao...> wrote: > Hello, > > This is probably off-topic but I hope someone here knows answer. I am > converting ARPA LM to FST using "transducersaurus". > The command that I use is C*min(det(L*(G))) > > My problem is, the *.fst.txt file that is created has output symbols at > unexpected locations. > > I expect the output symbols to be <eps> at all phonemes in the word except > last phoneme and output symbol = "word" for only the last phoneme. > > But what I'm seeing is that sometimes output symbol = "word" even for > phonemes that are not last phonemes in the word. > > I wonder what the issue is and how to properly identify word end boundary in > an FST. > > Let me know if sharing my files will help in diagnosis. > > Thanks, > Ethan > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics > Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > |
|
From: E <oth...@ao...> - 2014-01-07 09:22:11
|
Hello, This is probably off-topic but I hope someone here knows answer. I am converting ARPA LM to FST using "transducersaurus". The command that I use is C*min(det(L*(G))) My problem is, the *.fst.txt file that is created has output symbols at unexpected locations. I expect the output symbols to be <eps> at all phonemes in the word except last phoneme and output symbol = "word" for only the last phoneme. But what I'm seeing is that sometimes output symbol = "word" even for phonemes that are not last phonemes in the word. I wonder what the issue is and how to properly identify word end boundary in an FST. Let me know if sharing my files will help in diagnosis. Thanks, Ethan |
|
From: Karel V. <ive...@fi...> - 2013-12-20 10:44:33
|
There is also a script directly in kaldi wsj/s5 from Korbinian Riedhammer: http://sourceforge.net/p/kaldi/code/HEAD/tree/trunk/egs/wsj/s5/utils/convert_slf.pl Karel Dne 19. 12. 2013 21:05, Daniel Povey napsal(a): > It looks like Paul Dixon has published something he wrote, here: > > https://gist.github.com/edobashira/5811963 > > Dan > > > > On Thu, Dec 19, 2013 at 3:20 AM, wei.shi-? <we...@im... > <mailto:we...@im...>> wrote: > > Hi all, > Though I know the purpose of kaldi is not to be compatible with > HTK, I still want to ask, is there a tool that can convert kaldi > lattices( e.g., generated by gmm-latgen-faster ) to HTK > compatible format? > Shi Wei > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. > Most IT > organizations don't have a clear picture of how application > performance > affects their revenue. With AppDynamics, you get 100% visibility > into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of > AppDynamics Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > <mailto:Kal...@li...> > https://lists.sourceforge.net/lists/listinfo/kaldi-users > > > > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > > > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users |
|
From: Daniel P. <dp...@gm...> - 2013-12-20 07:57:10
|
No, they can still be equivalent in the FST sense but be different in that way. Dan On Thu, Dec 19, 2013 at 10:57 PM, Lahiru Samarakoon <lah...@gm...>wrote: > Please help me with this one last query. > If the kaldi program "fstisstochastic" gives significantly different > values for two fsts, is it safe to say those are not equivalent? > > For example, > fstisstochastic A.fst > *0.693747* -0.877091 > > fstisstochastic B.fst > *24.1318* -0.877091 > > Thanks in advance, > Lahiru > > > On Fri, Dec 20, 2013 at 1:08 PM, Lahiru Samarakoon <lah...@gm...>wrote: > >> Thanks a lot for the clarification Dan. >> >> >> On Fri, Dec 20, 2013 at 1:04 PM, Daniel Povey <dp...@gm...> wrote: >> >>> No, because for each path there will be multiple alternatives in B and >>> the sum of the path weights will be more than one. I recommend to address >>> further questions on this topic to www.openfst.org. >>> Dan >>> >>> >>> >>> On Thu, Dec 19, 2013 at 8:53 PM, Lahiru Samarakoon <lah...@gm...>wrote: >>> >>>> Is this also applies to the log semiring? >>>> Thanks >>>> Lahiru >>>> >>>> >>>> On Fri, Dec 20, 2013 at 12:46 PM, Daniel Povey <dp...@gm...>wrote: >>>> >>>>> >>>>> I am trying to decompose weighted acceptors (Language models) . I >>>>>> observe this behavior when I compose them to build a new fst. I think I am >>>>>> doing something wrong here. >>>>>> >>>>>> Please be kind enough to clarify me the following. >>>>>> >>>>>> Given a *weighted acceptor *"*A*", lets say acceptor "*B*" is >>>>>> created by setting all the weights of "A" to *zero *including the >>>>>> final states weights. >>>>>> is >>>>>> *"A compose B" equivalent to "A"?* >>>>>> >>>>> >>>>> If you are in the tropical semiring, then yes, because any path in A >>>>> is also in B (with zero cost). But "A compose B" is a wasteful way of >>>>> representing the acceptor because there are multiple alternative paths >>>>> through B that will end up multiplying the size, potentially by the n-gram >>>>> order minus one. So it's equivalent in the FST sense, but it's much >>>>> larger. >>>>> >>>>> Dan >>>>> >>>>> >>>>> >>>>>> >>>>>> if the answer is not equivalent, could you state the reason briefly >>>>>> >>>>>> Thank you, >>>>>> >>>>>> Best Regards, >>>>>> Lahiru >>>>>> >>>>>> >>>>>> On Fri, Dec 20, 2013 at 3:41 AM, Al Zatv <al...@gm...> wrote: >>>>>> >>>>>>> Hello! Actually, I don't have such a problem. I think Lahiru >>>>>>> Samarakoon, the topic starter, have it: >>>>>>> Quote: >>>>>>> ===== >>>>>>> >>>>>>> I have my own language model. When I minimize it, actually it >>>>>>> increases the number of states and arcs. >>>>>>> >>>>>>> How can this be possible? >>>>>>> I have created the LM in arpa format and used scripts from Kaldi to >>>>>>> create the G.fst >>>>>>> ===== >>>>>>> >>>>>>> Usually ARPA LMs are acceptors and they are determinized by default, >>>>>>> so if Lahiru really observe that min(G.fst) is bigger than G.fst, it is >>>>>>> suspicious. >>>>>>> 19.12.2013 20:09 пользователь "Michael Riley" <ri...@go...> >>>>>>> написал: >>>>>>> >>>>>>> Not sure what exactly you are doing or what code you are calling but >>>>>>>> if you are using the algorithms in OpenFst then >>>>>>>> |det(A)| can be > that |A| but |min(det(A))| <= |det(A)| if A is an >>>>>>>> acceptor. >>>>>>>> >>>>>>>> I'd be surprised if there is a bug here (given this condition I >>>>>>>> believe is tested on millions of randomly generated machines) but If you >>>>>>>> have a counter-example when using unmodified OpenFst as above, then send it >>>>>>>> to us along with the precise commands you ran that demonstrate the problem. >>>>>>>> >>>>>>>> -m >>>>>>>> >>>>>>>> -m >>>>>>>> >>>>>>>> >>>>>>>> On Thu, Dec 19, 2013 at 2:55 AM, Al Zatv <al...@gm...> wrote: >>>>>>>> >>>>>>>>> According to this document, size of transducer can be larger than >>>>>>>>> original. But ARPA LM is an acceptor, so it must have optimal number of >>>>>>>>> states, or I don't misunderstood something? >>>>>>>>> 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> >>>>>>>>> написал: >>>>>>>>> >>>>>>>>> http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon < >>>>>>>>>> lah...@gm...> wrote: >>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> HI All, >>>>>>>>>>> >>>>>>>>>>> I have my own language model. When I minimize it, actually it >>>>>>>>>>> increases the number of states and arcs. >>>>>>>>>>> >>>>>>>>>>> How can this be possible? >>>>>>>>>>> I have created the LM in arpa format and used scripts from Kaldi >>>>>>>>>>> to create the G.fst. >>>>>>>>>>> >>>>>>>>>>> Thank you, >>>>>>>>>>> >>>>>>>>>>> Best Regards, >>>>>>>>>>> Lahiru >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> ------------------------------------------------------------------------------ >>>>>>>>>>> Rapidly troubleshoot problems before they affect your business. >>>>>>>>>>> Most IT >>>>>>>>>>> organizations don't have a clear picture of how application >>>>>>>>>>> performance >>>>>>>>>>> affects their revenue. With AppDynamics, you get 100% visibility >>>>>>>>>>> into your >>>>>>>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>>>>>>> AppDynamics Pro! >>>>>>>>>>> >>>>>>>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>>>>>>> _______________________________________________ >>>>>>>>>>> Kaldi-users mailing list >>>>>>>>>>> Kal...@li... >>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> ------------------------------------------------------------------------------ >>>>>>>>>> Rapidly troubleshoot problems before they affect your business. >>>>>>>>>> Most IT >>>>>>>>>> organizations don't have a clear picture of how application >>>>>>>>>> performance >>>>>>>>>> affects their revenue. With AppDynamics, you get 100% visibility >>>>>>>>>> into your >>>>>>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>>>>>> AppDynamics Pro! >>>>>>>>>> >>>>>>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>>>>>> _______________________________________________ >>>>>>>>>> Kaldi-users mailing list >>>>>>>>>> Kal...@li... >>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>>>>>> >>>>>>>>>> >>>>>>>> >>>>>> >>>>>> >>>>>> ------------------------------------------------------------------------------ >>>>>> Rapidly troubleshoot problems before they affect your business. Most >>>>>> IT >>>>>> organizations don't have a clear picture of how application >>>>>> performance >>>>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>>>> your >>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>> AppDynamics Pro! >>>>>> >>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>> _______________________________________________ >>>>>> Kaldi-users mailing list >>>>>> Kal...@li... >>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>> >>>>>> >>>>> >>>> >>> >> > |
|
From: Lahiru S. <lah...@gm...> - 2013-12-20 06:57:47
|
Please help me with this one last query. If the kaldi program "fstisstochastic" gives significantly different values for two fsts, is it safe to say those are not equivalent? For example, fstisstochastic A.fst *0.693747* -0.877091 fstisstochastic B.fst *24.1318* -0.877091 Thanks in advance, Lahiru On Fri, Dec 20, 2013 at 1:08 PM, Lahiru Samarakoon <lah...@gm...>wrote: > Thanks a lot for the clarification Dan. > > > On Fri, Dec 20, 2013 at 1:04 PM, Daniel Povey <dp...@gm...> wrote: > >> No, because for each path there will be multiple alternatives in B and >> the sum of the path weights will be more than one. I recommend to address >> further questions on this topic to www.openfst.org. >> Dan >> >> >> >> On Thu, Dec 19, 2013 at 8:53 PM, Lahiru Samarakoon <lah...@gm...>wrote: >> >>> Is this also applies to the log semiring? >>> Thanks >>> Lahiru >>> >>> >>> On Fri, Dec 20, 2013 at 12:46 PM, Daniel Povey <dp...@gm...> wrote: >>> >>>> >>>> I am trying to decompose weighted acceptors (Language models) . I >>>>> observe this behavior when I compose them to build a new fst. I think I am >>>>> doing something wrong here. >>>>> >>>>> Please be kind enough to clarify me the following. >>>>> >>>>> Given a *weighted acceptor *"*A*", lets say acceptor "*B*" is >>>>> created by setting all the weights of "A" to *zero *including the >>>>> final states weights. >>>>> is >>>>> *"A compose B" equivalent to "A"?* >>>>> >>>> >>>> If you are in the tropical semiring, then yes, because any path in A is >>>> also in B (with zero cost). But "A compose B" is a wasteful way of >>>> representing the acceptor because there are multiple alternative paths >>>> through B that will end up multiplying the size, potentially by the n-gram >>>> order minus one. So it's equivalent in the FST sense, but it's much >>>> larger. >>>> >>>> Dan >>>> >>>> >>>> >>>>> >>>>> if the answer is not equivalent, could you state the reason briefly >>>>> >>>>> Thank you, >>>>> >>>>> Best Regards, >>>>> Lahiru >>>>> >>>>> >>>>> On Fri, Dec 20, 2013 at 3:41 AM, Al Zatv <al...@gm...> wrote: >>>>> >>>>>> Hello! Actually, I don't have such a problem. I think Lahiru >>>>>> Samarakoon, the topic starter, have it: >>>>>> Quote: >>>>>> ===== >>>>>> >>>>>> I have my own language model. When I minimize it, actually it >>>>>> increases the number of states and arcs. >>>>>> >>>>>> How can this be possible? >>>>>> I have created the LM in arpa format and used scripts from Kaldi to >>>>>> create the G.fst >>>>>> ===== >>>>>> >>>>>> Usually ARPA LMs are acceptors and they are determinized by default, >>>>>> so if Lahiru really observe that min(G.fst) is bigger than G.fst, it is >>>>>> suspicious. >>>>>> 19.12.2013 20:09 пользователь "Michael Riley" <ri...@go...> >>>>>> написал: >>>>>> >>>>>> Not sure what exactly you are doing or what code you are calling but >>>>>>> if you are using the algorithms in OpenFst then >>>>>>> |det(A)| can be > that |A| but |min(det(A))| <= |det(A)| if A is an >>>>>>> acceptor. >>>>>>> >>>>>>> I'd be surprised if there is a bug here (given this condition I >>>>>>> believe is tested on millions of randomly generated machines) but If you >>>>>>> have a counter-example when using unmodified OpenFst as above, then send it >>>>>>> to us along with the precise commands you ran that demonstrate the problem. >>>>>>> >>>>>>> -m >>>>>>> >>>>>>> -m >>>>>>> >>>>>>> >>>>>>> On Thu, Dec 19, 2013 at 2:55 AM, Al Zatv <al...@gm...> wrote: >>>>>>> >>>>>>>> According to this document, size of transducer can be larger than >>>>>>>> original. But ARPA LM is an acceptor, so it must have optimal number of >>>>>>>> states, or I don't misunderstood something? >>>>>>>> 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> >>>>>>>> написал: >>>>>>>> >>>>>>>> http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc >>>>>>>>> >>>>>>>>> >>>>>>>>> On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon < >>>>>>>>> lah...@gm...> wrote: >>>>>>>>> >>>>>>>>>> >>>>>>>>>> HI All, >>>>>>>>>> >>>>>>>>>> I have my own language model. When I minimize it, actually it >>>>>>>>>> increases the number of states and arcs. >>>>>>>>>> >>>>>>>>>> How can this be possible? >>>>>>>>>> I have created the LM in arpa format and used scripts from Kaldi >>>>>>>>>> to create the G.fst. >>>>>>>>>> >>>>>>>>>> Thank you, >>>>>>>>>> >>>>>>>>>> Best Regards, >>>>>>>>>> Lahiru >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> ------------------------------------------------------------------------------ >>>>>>>>>> Rapidly troubleshoot problems before they affect your business. >>>>>>>>>> Most IT >>>>>>>>>> organizations don't have a clear picture of how application >>>>>>>>>> performance >>>>>>>>>> affects their revenue. With AppDynamics, you get 100% visibility >>>>>>>>>> into your >>>>>>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>>>>>> AppDynamics Pro! >>>>>>>>>> >>>>>>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>>>>>> _______________________________________________ >>>>>>>>>> Kaldi-users mailing list >>>>>>>>>> Kal...@li... >>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> ------------------------------------------------------------------------------ >>>>>>>>> Rapidly troubleshoot problems before they affect your business. >>>>>>>>> Most IT >>>>>>>>> organizations don't have a clear picture of how application >>>>>>>>> performance >>>>>>>>> affects their revenue. With AppDynamics, you get 100% visibility >>>>>>>>> into your >>>>>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>>>>> AppDynamics Pro! >>>>>>>>> >>>>>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>>>>> _______________________________________________ >>>>>>>>> Kaldi-users mailing list >>>>>>>>> Kal...@li... >>>>>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>>>>> >>>>>>>>> >>>>>>> >>>>> >>>>> >>>>> ------------------------------------------------------------------------------ >>>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>>> organizations don't have a clear picture of how application performance >>>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>>> your >>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>> AppDynamics Pro! >>>>> >>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>> _______________________________________________ >>>>> Kaldi-users mailing list >>>>> Kal...@li... >>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>> >>>>> >>>> >>> >> > |
|
From: Lahiru S. <lah...@gm...> - 2013-12-20 05:08:26
|
Thanks a lot for the clarification Dan. On Fri, Dec 20, 2013 at 1:04 PM, Daniel Povey <dp...@gm...> wrote: > No, because for each path there will be multiple alternatives in B and the > sum of the path weights will be more than one. I recommend to address > further questions on this topic to www.openfst.org. > Dan > > > > On Thu, Dec 19, 2013 at 8:53 PM, Lahiru Samarakoon <lah...@gm...>wrote: > >> Is this also applies to the log semiring? >> Thanks >> Lahiru >> >> >> On Fri, Dec 20, 2013 at 12:46 PM, Daniel Povey <dp...@gm...> wrote: >> >>> >>> I am trying to decompose weighted acceptors (Language models) . I >>>> observe this behavior when I compose them to build a new fst. I think I am >>>> doing something wrong here. >>>> >>>> Please be kind enough to clarify me the following. >>>> >>>> Given a *weighted acceptor *"*A*", lets say acceptor "*B*" is created >>>> by setting all the weights of "A" to *zero *including the final states >>>> weights. >>>> is >>>> *"A compose B" equivalent to "A"?* >>>> >>> >>> If you are in the tropical semiring, then yes, because any path in A is >>> also in B (with zero cost). But "A compose B" is a wasteful way of >>> representing the acceptor because there are multiple alternative paths >>> through B that will end up multiplying the size, potentially by the n-gram >>> order minus one. So it's equivalent in the FST sense, but it's much >>> larger. >>> >>> Dan >>> >>> >>> >>>> >>>> if the answer is not equivalent, could you state the reason briefly >>>> >>>> Thank you, >>>> >>>> Best Regards, >>>> Lahiru >>>> >>>> >>>> On Fri, Dec 20, 2013 at 3:41 AM, Al Zatv <al...@gm...> wrote: >>>> >>>>> Hello! Actually, I don't have such a problem. I think Lahiru >>>>> Samarakoon, the topic starter, have it: >>>>> Quote: >>>>> ===== >>>>> >>>>> I have my own language model. When I minimize it, actually it >>>>> increases the number of states and arcs. >>>>> >>>>> How can this be possible? >>>>> I have created the LM in arpa format and used scripts from Kaldi to >>>>> create the G.fst >>>>> ===== >>>>> >>>>> Usually ARPA LMs are acceptors and they are determinized by default, >>>>> so if Lahiru really observe that min(G.fst) is bigger than G.fst, it is >>>>> suspicious. >>>>> 19.12.2013 20:09 пользователь "Michael Riley" <ri...@go...> >>>>> написал: >>>>> >>>>> Not sure what exactly you are doing or what code you are calling but >>>>>> if you are using the algorithms in OpenFst then >>>>>> |det(A)| can be > that |A| but |min(det(A))| <= |det(A)| if A is an >>>>>> acceptor. >>>>>> >>>>>> I'd be surprised if there is a bug here (given this condition I >>>>>> believe is tested on millions of randomly generated machines) but If you >>>>>> have a counter-example when using unmodified OpenFst as above, then send it >>>>>> to us along with the precise commands you ran that demonstrate the problem. >>>>>> >>>>>> -m >>>>>> >>>>>> -m >>>>>> >>>>>> >>>>>> On Thu, Dec 19, 2013 at 2:55 AM, Al Zatv <al...@gm...> wrote: >>>>>> >>>>>>> According to this document, size of transducer can be larger than >>>>>>> original. But ARPA LM is an acceptor, so it must have optimal number of >>>>>>> states, or I don't misunderstood something? >>>>>>> 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> >>>>>>> написал: >>>>>>> >>>>>>> http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc >>>>>>>> >>>>>>>> >>>>>>>> On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon < >>>>>>>> lah...@gm...> wrote: >>>>>>>> >>>>>>>>> >>>>>>>>> HI All, >>>>>>>>> >>>>>>>>> I have my own language model. When I minimize it, actually it >>>>>>>>> increases the number of states and arcs. >>>>>>>>> >>>>>>>>> How can this be possible? >>>>>>>>> I have created the LM in arpa format and used scripts from Kaldi >>>>>>>>> to create the G.fst. >>>>>>>>> >>>>>>>>> Thank you, >>>>>>>>> >>>>>>>>> Best Regards, >>>>>>>>> Lahiru >>>>>>>>> >>>>>>>>> >>>>>>>>> ------------------------------------------------------------------------------ >>>>>>>>> Rapidly troubleshoot problems before they affect your business. >>>>>>>>> Most IT >>>>>>>>> organizations don't have a clear picture of how application >>>>>>>>> performance >>>>>>>>> affects their revenue. With AppDynamics, you get 100% visibility >>>>>>>>> into your >>>>>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>>>>> AppDynamics Pro! >>>>>>>>> >>>>>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>>>>> _______________________________________________ >>>>>>>>> Kaldi-users mailing list >>>>>>>>> Kal...@li... >>>>>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> ------------------------------------------------------------------------------ >>>>>>>> Rapidly troubleshoot problems before they affect your business. >>>>>>>> Most IT >>>>>>>> organizations don't have a clear picture of how application >>>>>>>> performance >>>>>>>> affects their revenue. With AppDynamics, you get 100% visibility >>>>>>>> into your >>>>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>>>> AppDynamics Pro! >>>>>>>> >>>>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>>>> _______________________________________________ >>>>>>>> Kaldi-users mailing list >>>>>>>> Kal...@li... >>>>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>>>> >>>>>>>> >>>>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>> organizations don't have a clear picture of how application performance >>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>> your >>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>> AppDynamics Pro! >>>> >>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>> _______________________________________________ >>>> Kaldi-users mailing list >>>> Kal...@li... >>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>> >>>> >>> >> > |
|
From: Daniel P. <dp...@gm...> - 2013-12-20 05:04:53
|
No, because for each path there will be multiple alternatives in B and the sum of the path weights will be more than one. I recommend to address further questions on this topic to www.openfst.org. Dan On Thu, Dec 19, 2013 at 8:53 PM, Lahiru Samarakoon <lah...@gm...>wrote: > Is this also applies to the log semiring? > Thanks > Lahiru > > > On Fri, Dec 20, 2013 at 12:46 PM, Daniel Povey <dp...@gm...> wrote: > >> >> I am trying to decompose weighted acceptors (Language models) . I observe >>> this behavior when I compose them to build a new fst. I think I am doing >>> something wrong here. >>> >>> Please be kind enough to clarify me the following. >>> >>> Given a *weighted acceptor *"*A*", lets say acceptor "*B*" is created >>> by setting all the weights of "A" to *zero *including the final states >>> weights. >>> is >>> *"A compose B" equivalent to "A"?* >>> >> >> If you are in the tropical semiring, then yes, because any path in A is >> also in B (with zero cost). But "A compose B" is a wasteful way of >> representing the acceptor because there are multiple alternative paths >> through B that will end up multiplying the size, potentially by the n-gram >> order minus one. So it's equivalent in the FST sense, but it's much >> larger. >> >> Dan >> >> >> >>> >>> if the answer is not equivalent, could you state the reason briefly >>> >>> Thank you, >>> >>> Best Regards, >>> Lahiru >>> >>> >>> On Fri, Dec 20, 2013 at 3:41 AM, Al Zatv <al...@gm...> wrote: >>> >>>> Hello! Actually, I don't have such a problem. I think Lahiru >>>> Samarakoon, the topic starter, have it: >>>> Quote: >>>> ===== >>>> >>>> I have my own language model. When I minimize it, actually it increases >>>> the number of states and arcs. >>>> >>>> How can this be possible? >>>> I have created the LM in arpa format and used scripts from Kaldi to >>>> create the G.fst >>>> ===== >>>> >>>> Usually ARPA LMs are acceptors and they are determinized by default, so >>>> if Lahiru really observe that min(G.fst) is bigger than G.fst, it is >>>> suspicious. >>>> 19.12.2013 20:09 пользователь "Michael Riley" <ri...@go...> >>>> написал: >>>> >>>> Not sure what exactly you are doing or what code you are calling but if >>>>> you are using the algorithms in OpenFst then >>>>> |det(A)| can be > that |A| but |min(det(A))| <= |det(A)| if A is an >>>>> acceptor. >>>>> >>>>> I'd be surprised if there is a bug here (given this condition I >>>>> believe is tested on millions of randomly generated machines) but If you >>>>> have a counter-example when using unmodified OpenFst as above, then send it >>>>> to us along with the precise commands you ran that demonstrate the problem. >>>>> >>>>> -m >>>>> >>>>> -m >>>>> >>>>> >>>>> On Thu, Dec 19, 2013 at 2:55 AM, Al Zatv <al...@gm...> wrote: >>>>> >>>>>> According to this document, size of transducer can be larger than >>>>>> original. But ARPA LM is an acceptor, so it must have optimal number of >>>>>> states, or I don't misunderstood something? >>>>>> 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> >>>>>> написал: >>>>>> >>>>>> http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc >>>>>>> >>>>>>> >>>>>>> On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon < >>>>>>> lah...@gm...> wrote: >>>>>>> >>>>>>>> >>>>>>>> HI All, >>>>>>>> >>>>>>>> I have my own language model. When I minimize it, actually it >>>>>>>> increases the number of states and arcs. >>>>>>>> >>>>>>>> How can this be possible? >>>>>>>> I have created the LM in arpa format and used scripts from Kaldi to >>>>>>>> create the G.fst. >>>>>>>> >>>>>>>> Thank you, >>>>>>>> >>>>>>>> Best Regards, >>>>>>>> Lahiru >>>>>>>> >>>>>>>> >>>>>>>> ------------------------------------------------------------------------------ >>>>>>>> Rapidly troubleshoot problems before they affect your business. >>>>>>>> Most IT >>>>>>>> organizations don't have a clear picture of how application >>>>>>>> performance >>>>>>>> affects their revenue. With AppDynamics, you get 100% visibility >>>>>>>> into your >>>>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>>>> AppDynamics Pro! >>>>>>>> >>>>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>>>> _______________________________________________ >>>>>>>> Kaldi-users mailing list >>>>>>>> Kal...@li... >>>>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> >>>>>>> ------------------------------------------------------------------------------ >>>>>>> Rapidly troubleshoot problems before they affect your business. Most >>>>>>> IT >>>>>>> organizations don't have a clear picture of how application >>>>>>> performance >>>>>>> affects their revenue. With AppDynamics, you get 100% visibility >>>>>>> into your >>>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>>> AppDynamics Pro! >>>>>>> >>>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>>> _______________________________________________ >>>>>>> Kaldi-users mailing list >>>>>>> Kal...@li... >>>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>>> >>>>>>> >>>>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into >>> your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>> AppDynamics Pro! >>> >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >>> >> > |
|
From: Lahiru S. <lah...@gm...> - 2013-12-20 04:53:33
|
Is this also applies to the log semiring? Thanks Lahiru On Fri, Dec 20, 2013 at 12:46 PM, Daniel Povey <dp...@gm...> wrote: > > I am trying to decompose weighted acceptors (Language models) . I observe >> this behavior when I compose them to build a new fst. I think I am doing >> something wrong here. >> >> Please be kind enough to clarify me the following. >> >> Given a *weighted acceptor *"*A*", lets say acceptor "*B*" is created >> by setting all the weights of "A" to *zero *including the final states >> weights. >> is >> *"A compose B" equivalent to "A"?* >> > > If you are in the tropical semiring, then yes, because any path in A is > also in B (with zero cost). But "A compose B" is a wasteful way of > representing the acceptor because there are multiple alternative paths > through B that will end up multiplying the size, potentially by the n-gram > order minus one. So it's equivalent in the FST sense, but it's much > larger. > > Dan > > > >> >> if the answer is not equivalent, could you state the reason briefly >> >> Thank you, >> >> Best Regards, >> Lahiru >> >> >> On Fri, Dec 20, 2013 at 3:41 AM, Al Zatv <al...@gm...> wrote: >> >>> Hello! Actually, I don't have such a problem. I think Lahiru Samarakoon, >>> the topic starter, have it: >>> Quote: >>> ===== >>> >>> I have my own language model. When I minimize it, actually it increases >>> the number of states and arcs. >>> >>> How can this be possible? >>> I have created the LM in arpa format and used scripts from Kaldi to >>> create the G.fst >>> ===== >>> >>> Usually ARPA LMs are acceptors and they are determinized by default, so >>> if Lahiru really observe that min(G.fst) is bigger than G.fst, it is >>> suspicious. >>> 19.12.2013 20:09 пользователь "Michael Riley" <ri...@go...> >>> написал: >>> >>> Not sure what exactly you are doing or what code you are calling but if >>>> you are using the algorithms in OpenFst then >>>> |det(A)| can be > that |A| but |min(det(A))| <= |det(A)| if A is an >>>> acceptor. >>>> >>>> I'd be surprised if there is a bug here (given this condition I believe >>>> is tested on millions of randomly generated machines) but If you have a >>>> counter-example when using unmodified OpenFst as above, then send it to us >>>> along with the precise commands you ran that demonstrate the problem. >>>> >>>> -m >>>> >>>> -m >>>> >>>> >>>> On Thu, Dec 19, 2013 at 2:55 AM, Al Zatv <al...@gm...> wrote: >>>> >>>>> According to this document, size of transducer can be larger than >>>>> original. But ARPA LM is an acceptor, so it must have optimal number of >>>>> states, or I don't misunderstood something? >>>>> 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> >>>>> написал: >>>>> >>>>> http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc >>>>>> >>>>>> >>>>>> On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon < >>>>>> lah...@gm...> wrote: >>>>>> >>>>>>> >>>>>>> HI All, >>>>>>> >>>>>>> I have my own language model. When I minimize it, actually it >>>>>>> increases the number of states and arcs. >>>>>>> >>>>>>> How can this be possible? >>>>>>> I have created the LM in arpa format and used scripts from Kaldi to >>>>>>> create the G.fst. >>>>>>> >>>>>>> Thank you, >>>>>>> >>>>>>> Best Regards, >>>>>>> Lahiru >>>>>>> >>>>>>> >>>>>>> ------------------------------------------------------------------------------ >>>>>>> Rapidly troubleshoot problems before they affect your business. Most >>>>>>> IT >>>>>>> organizations don't have a clear picture of how application >>>>>>> performance >>>>>>> affects their revenue. With AppDynamics, you get 100% visibility >>>>>>> into your >>>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>>> AppDynamics Pro! >>>>>>> >>>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>>> _______________________________________________ >>>>>>> Kaldi-users mailing list >>>>>>> Kal...@li... >>>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>> ------------------------------------------------------------------------------ >>>>>> Rapidly troubleshoot problems before they affect your business. Most >>>>>> IT >>>>>> organizations don't have a clear picture of how application >>>>>> performance >>>>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>>>> your >>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>> AppDynamics Pro! >>>>>> >>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>> _______________________________________________ >>>>>> Kaldi-users mailing list >>>>>> Kal...@li... >>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>> >>>>>> >>>> >> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >> Pro! >> >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> > |
|
From: Daniel P. <dp...@gm...> - 2013-12-20 04:46:43
|
> I am trying to decompose weighted acceptors (Language models) . I observe > this behavior when I compose them to build a new fst. I think I am doing > something wrong here. > > Please be kind enough to clarify me the following. > > Given a *weighted acceptor *"*A*", lets say acceptor "*B*" is created by > setting all the weights of "A" to *zero *including the final states > weights. > is > *"A compose B" equivalent to "A"?* > If you are in the tropical semiring, then yes, because any path in A is also in B (with zero cost). But "A compose B" is a wasteful way of representing the acceptor because there are multiple alternative paths through B that will end up multiplying the size, potentially by the n-gram order minus one. So it's equivalent in the FST sense, but it's much larger. Dan > > if the answer is not equivalent, could you state the reason briefly > > Thank you, > > Best Regards, > Lahiru > > > On Fri, Dec 20, 2013 at 3:41 AM, Al Zatv <al...@gm...> wrote: > >> Hello! Actually, I don't have such a problem. I think Lahiru Samarakoon, >> the topic starter, have it: >> Quote: >> ===== >> >> I have my own language model. When I minimize it, actually it increases >> the number of states and arcs. >> >> How can this be possible? >> I have created the LM in arpa format and used scripts from Kaldi to >> create the G.fst >> ===== >> >> Usually ARPA LMs are acceptors and they are determinized by default, so >> if Lahiru really observe that min(G.fst) is bigger than G.fst, it is >> suspicious. >> 19.12.2013 20:09 пользователь "Michael Riley" <ri...@go...> написал: >> >> Not sure what exactly you are doing or what code you are calling but if >>> you are using the algorithms in OpenFst then >>> |det(A)| can be > that |A| but |min(det(A))| <= |det(A)| if A is an >>> acceptor. >>> >>> I'd be surprised if there is a bug here (given this condition I believe >>> is tested on millions of randomly generated machines) but If you have a >>> counter-example when using unmodified OpenFst as above, then send it to us >>> along with the precise commands you ran that demonstrate the problem. >>> >>> -m >>> >>> -m >>> >>> >>> On Thu, Dec 19, 2013 at 2:55 AM, Al Zatv <al...@gm...> wrote: >>> >>>> According to this document, size of transducer can be larger than >>>> original. But ARPA LM is an acceptor, so it must have optimal number of >>>> states, or I don't misunderstood something? >>>> 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> >>>> написал: >>>> >>>> http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc >>>>> >>>>> >>>>> On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon <lah...@gm... >>>>> > wrote: >>>>> >>>>>> >>>>>> HI All, >>>>>> >>>>>> I have my own language model. When I minimize it, actually it >>>>>> increases the number of states and arcs. >>>>>> >>>>>> How can this be possible? >>>>>> I have created the LM in arpa format and used scripts from Kaldi to >>>>>> create the G.fst. >>>>>> >>>>>> Thank you, >>>>>> >>>>>> Best Regards, >>>>>> Lahiru >>>>>> >>>>>> >>>>>> ------------------------------------------------------------------------------ >>>>>> Rapidly troubleshoot problems before they affect your business. Most >>>>>> IT >>>>>> organizations don't have a clear picture of how application >>>>>> performance >>>>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>>>> your >>>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>>> AppDynamics Pro! >>>>>> >>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>>> _______________________________________________ >>>>>> Kaldi-users mailing list >>>>>> Kal...@li... >>>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>>> >>>>>> >>>>> >>>>> >>>>> ------------------------------------------------------------------------------ >>>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>>> organizations don't have a clear picture of how application performance >>>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>>> your >>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>> AppDynamics Pro! >>>>> >>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>> _______________________________________________ >>>>> Kaldi-users mailing list >>>>> Kal...@li... >>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>> >>>>> >>> > > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics > Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > > |
|
From: Lahiru S. <lah...@gm...> - 2013-12-20 02:32:43
|
I am trying to decompose weighted acceptors (Language models) . I observe this behavior when I compose them to build a new fst. I think I am doing something wrong here. Please be kind enough to clarify me the following. Given a *weighted acceptor *"*A*", lets say acceptor "*B*" is created by setting all the weights of "A" to *zero *including the final states weights. is *"A compose B" equivalent to "A"?* if the answer is not equivalent, could you state the reason briefly Thank you, Best Regards, Lahiru On Fri, Dec 20, 2013 at 3:41 AM, Al Zatv <al...@gm...> wrote: > Hello! Actually, I don't have such a problem. I think Lahiru Samarakoon, > the topic starter, have it: > Quote: > ===== > > I have my own language model. When I minimize it, actually it increases > the number of states and arcs. > > How can this be possible? > I have created the LM in arpa format and used scripts from Kaldi to create > the G.fst > ===== > > Usually ARPA LMs are acceptors and they are determinized by default, so if > Lahiru really observe that min(G.fst) is bigger than G.fst, it is > suspicious. > 19.12.2013 20:09 пользователь "Michael Riley" <ri...@go...> написал: > > Not sure what exactly you are doing or what code you are calling but if >> you are using the algorithms in OpenFst then >> |det(A)| can be > that |A| but |min(det(A))| <= |det(A)| if A is an >> acceptor. >> >> I'd be surprised if there is a bug here (given this condition I believe >> is tested on millions of randomly generated machines) but If you have a >> counter-example when using unmodified OpenFst as above, then send it to us >> along with the precise commands you ran that demonstrate the problem. >> >> -m >> >> -m >> >> >> On Thu, Dec 19, 2013 at 2:55 AM, Al Zatv <al...@gm...> wrote: >> >>> According to this document, size of transducer can be larger than >>> original. But ARPA LM is an acceptor, so it must have optimal number of >>> states, or I don't misunderstood something? >>> 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> >>> написал: >>> >>> http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc >>>> >>>> >>>> On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon <lah...@gm...>wrote: >>>> >>>>> >>>>> HI All, >>>>> >>>>> I have my own language model. When I minimize it, actually it >>>>> increases the number of states and arcs. >>>>> >>>>> How can this be possible? >>>>> I have created the LM in arpa format and used scripts from Kaldi to >>>>> create the G.fst. >>>>> >>>>> Thank you, >>>>> >>>>> Best Regards, >>>>> Lahiru >>>>> >>>>> >>>>> ------------------------------------------------------------------------------ >>>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>>> organizations don't have a clear picture of how application performance >>>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>>> your >>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>>> AppDynamics Pro! >>>>> >>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>>> _______________________________________________ >>>>> Kaldi-users mailing list >>>>> Kal...@li... >>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>>> >>>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>> organizations don't have a clear picture of how application performance >>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>> your >>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>> AppDynamics Pro! >>>> >>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>> _______________________________________________ >>>> Kaldi-users mailing list >>>> Kal...@li... >>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>> >>>> >> |
|
From: Daniel P. <dp...@gm...> - 2013-12-19 20:05:45
|
It looks like Paul Dixon has published something he wrote, here: https://gist.github.com/edobashira/5811963 Dan On Thu, Dec 19, 2013 at 3:20 AM, wei.shi-新 <we...@im...> wrote: > Hi all, > > Though I know the purpose of kaldi is not to be compatible with HTK, I > still want to ask, is there a tool that can convert kaldi lattices( e.g., > generated by gmm-latgen-faster ) to HTK compatible format? > > > Shi Wei > > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics > Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > > |
|
From: Al Z. <al...@gm...> - 2013-12-19 19:42:01
|
Hello! Actually, I don't have such a problem. I think Lahiru Samarakoon, the topic starter, have it: Quote: ===== I have my own language model. When I minimize it, actually it increases the number of states and arcs. How can this be possible? I have created the LM in arpa format and used scripts from Kaldi to create the G.fst ===== Usually ARPA LMs are acceptors and they are determinized by default, so if Lahiru really observe that min(G.fst) is bigger than G.fst, it is suspicious. 19.12.2013 20:09 пользователь "Michael Riley" <ri...@go...> написал: > Not sure what exactly you are doing or what code you are calling but if > you are using the algorithms in OpenFst then > |det(A)| can be > that |A| but |min(det(A))| <= |det(A)| if A is an > acceptor. > > I'd be surprised if there is a bug here (given this condition I believe is > tested on millions of randomly generated machines) but If you have a > counter-example when using unmodified OpenFst as above, then send it to us > along with the precise commands you ran that demonstrate the problem. > > -m > > -m > > > On Thu, Dec 19, 2013 at 2:55 AM, Al Zatv <al...@gm...> wrote: > >> According to this document, size of transducer can be larger than >> original. But ARPA LM is an acceptor, so it must have optimal number of >> states, or I don't misunderstood something? >> 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> написал: >> >> http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc >>> >>> >>> On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon <lah...@gm...>wrote: >>> >>>> >>>> HI All, >>>> >>>> I have my own language model. When I minimize it, actually it increases >>>> the number of states and arcs. >>>> >>>> How can this be possible? >>>> I have created the LM in arpa format and used scripts from Kaldi to >>>> create the G.fst. >>>> >>>> Thank you, >>>> >>>> Best Regards, >>>> Lahiru >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>> organizations don't have a clear picture of how application performance >>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>> your >>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>> AppDynamics Pro! >>>> >>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>> _______________________________________________ >>>> Kaldi-users mailing list >>>> Kal...@li... >>>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>>> >>>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into >>> your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>> AppDynamics Pro! >>> >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >>> > |
|
From: Daniel P. <dp...@gm...> - 2013-12-19 19:15:46
|
If it's an acceptor (i.e. the input and output symbols are matched one for one) then minimization should not increase the size. But it could be that some of the operations you previously did on it, changed it so it was no longer an acceptor. Dan On Wed, Dec 18, 2013 at 11:55 PM, Al Zatv <al...@gm...> wrote: > According to this document, size of transducer can be larger than > original. But ARPA LM is an acceptor, so it must have optimal number of > states, or I don't misunderstood something? > 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> написал: > > http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc >> >> >> On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon <lah...@gm...>wrote: >> >>> >>> HI All, >>> >>> I have my own language model. When I minimize it, actually it increases >>> the number of states and arcs. >>> >>> How can this be possible? >>> I have created the LM in arpa format and used scripts from Kaldi to >>> create the G.fst. >>> >>> Thank you, >>> >>> Best Regards, >>> Lahiru >>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into >>> your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>> AppDynamics Pro! >>> >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >>> >> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >> Pro! >> >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics > Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > > |
|
From: Michael R. <ri...@go...> - 2013-12-19 16:09:23
|
Not sure what exactly you are doing or what code you are calling but if you are using the algorithms in OpenFst then |det(A)| can be > that |A| but |min(det(A))| <= |det(A)| if A is an acceptor. I'd be surprised if there is a bug here (given this condition I believe is tested on millions of randomly generated machines) but If you have a counter-example when using unmodified OpenFst as above, then send it to us along with the precise commands you ran that demonstrate the problem. -m -m On Thu, Dec 19, 2013 at 2:55 AM, Al Zatv <al...@gm...> wrote: > According to this document, size of transducer can be larger than > original. But ARPA LM is an acceptor, so it must have optimal number of > states, or I don't misunderstood something? > 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> написал: > > http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc >> >> >> On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon <lah...@gm...>wrote: >> >>> >>> HI All, >>> >>> I have my own language model. When I minimize it, actually it increases >>> the number of states and arcs. >>> >>> How can this be possible? >>> I have created the LM in arpa format and used scripts from Kaldi to >>> create the G.fst. >>> >>> Thank you, >>> >>> Best Regards, >>> Lahiru >>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into >>> your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>> AppDynamics Pro! >>> >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >>> >> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >> Pro! >> >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> |
|
From: wei.shi-新 <we...@im...> - 2013-12-19 11:57:00
|
Hi all, Though I know the purpose of kaldi is not to be compatible with HTK, I still want to ask, is there a tool that can convert kaldi lattices( e.g., generated by gmm-latgen-faster ) to HTK compatible format? Shi Wei |
|
From: Al Z. <al...@gm...> - 2013-12-19 07:55:10
|
According to this document, size of transducer can be larger than original. But ARPA LM is an acceptor, so it must have optimal number of states, or I don't misunderstood something? 19.12.2013 10:23 пользователь "Michael Riley" <ri...@go...> написал: > http://www.openfst.org/twiki/bin/view/FST/MinimizeDoc > > > On Thu, Dec 19, 2013 at 1:13 AM, Lahiru Samarakoon <lah...@gm...>wrote: > >> >> HI All, >> >> I have my own language model. When I minimize it, actually it increases >> the number of states and arcs. >> >> How can this be possible? >> I have created the LM in arpa format and used scripts from Kaldi to >> create the G.fst. >> >> Thank you, >> >> Best Regards, >> Lahiru >> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >> Pro! >> >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> >> > > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics > Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > > |
1 message has been excluded from this view by a project administrator.
