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: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-05-08 14:10:47
|
On Wed, May 8, 2013 at 2:52 PM, Mailing list used for User Communication and Updates <kal...@li...> wrote: > Using Opensuse 12.3 64 bit, kernel 3.9. > > 1. I issued svn co svn://svn.code.sf.net/p/kaldi/code/trunk kaldi-trunk > 2. download appeared to finish correctly > 3. examined INSTALL, went into dir /tools and did make, finished without > error as far as I can tell > 4. then ../src and ./configure, which stops at > > $ ./configure > Configuring ... > Checking OpenFST library in /home/colin/kaldi-trunk/tools/openfst ... > Checking OpenFst library was patched. > Doing OS specific configurations ... > ***configure failed: Static OpenFST library not found: See > ../tools/INSTALL *** Could you check to see if you have the following file kaldi-trunk/tools/openfst/lib/libfst.a? Normally typing make in the tools directory should create it, unless there was some error that didn't get reported. > 5. Checked tools install, it just tells me to do make. I did a search on > "openfst*" to find a library or executable on the system, but outside > the download dirs nothing available. No, it's not going to be in the systems directories, but just in the tools directory. > 6. attempted to compile as root (bad, but I thought maybe compile needed > to be able to write somewhere) no difference. You don't need to do anything as root. > 7. attempted to define --fst-root but no change to error. That's only helpful if you already have a OpenFST installation that you want to use. > Any suggestions? > > -- > --- > Colin Beckingham > http://www.it4gh.com > > ------------------------------------------------------------------------------ > Learn Graph Databases - Download FREE O'Reilly Book > "Graph Databases" is the definitive new guide to graph databases and > their applications. This 200-page book is written by three acclaimed > leaders in the field. The early access version is available now. > Download your free book today! http://p.sf.net/sfu/neotech_d2d_may > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-05-08 13:52:03
|
Using Opensuse 12.3 64 bit, kernel 3.9. 1. I issued svn co svn://svn.code.sf.net/p/kaldi/code/trunk kaldi-trunk 2. download appeared to finish correctly 3. examined INSTALL, went into dir /tools and did make, finished without error as far as I can tell 4. then ../src and ./configure, which stops at $ ./configure Configuring ... Checking OpenFST library in /home/colin/kaldi-trunk/tools/openfst ... Checking OpenFst library was patched. Doing OS specific configurations ... ***configure failed: Static OpenFST library not found: See ../tools/INSTALL *** 5. Checked tools install, it just tells me to do make. I did a search on "openfst*" to find a library or executable on the system, but outside the download dirs nothing available. 6. attempted to compile as root (bad, but I thought maybe compile needed to be able to write somewhere) no difference. 7. attempted to define --fst-root but no change to error. Any suggestions? -- --- Colin Beckingham http://www.it4gh.com |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-04-26 13:35:59
|
Hi, as for taking input from a microphone you might want to have a look at the (optional) online decoder. There is a simple demo script in egs/voxforge/online-demo. There is also a version that can decode .wav files without a separate feature extraction step. The code is not extensively tested and lacks some polish and flexibility, but perhaps you can use it as an example. Vassil On Fri, Apr 26, 2013 at 3:55 PM, Mailing list used for User Communication and Updates <kal...@li...> wrote: > Dear all, > I'm working on the development of an R&D voice-to-voice module and I would > like to use Kaldi for the transcription part (we made some tests with > PocketSphinx but we found that Kaldi is more adapted for our task). > > The problem is that I'm not really expert in speech2text (Kaldi models will > be provided by another team) and I did not found scripts (or plugins) > allowing an easy integration of Kaldi : How to manage streams (with > gstreammer for example)? how we can decode a single wav file? etc. > > Is someone else has already handle these problems? are there any scripts to > facilitate this task? > > thanks in advance for your reply, > > Best regards > > > ------------------------------------------------------------------------------ > Try New Relic Now & We'll Send You this Cool Shirt > New Relic is the only SaaS-based application performance monitoring service > that delivers powerful full stack analytics. Optimize and monitor your > browser, app, & servers with just a few lines of code. Try New Relic > and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-04-26 12:55:56
|
Dear all, I'm working on the development of an R&D voice-to-voice module and I would like to use Kaldi for the transcription part (we made some tests with PocketSphinx but we found that Kaldi is more adapted for our task). The problem is that I'm not really expert in speech2text (Kaldi models will be provided by another team) and I did not found scripts (or plugins) allowing an easy integration of Kaldi : How to manage streams (with gstreammer for example)? how we can decode a single wav file? etc. Is someone else has already handle these problems? are there any scripts to facilitate this task? thanks in advance for your reply, Best regards |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-04-16 12:27:43
|
Thanks all of you, I guess that will suffice ;) I'll post here again if we encounter any other difficulties. - Thomas On Tue, Apr 16, 2013 at 2:14 PM, Mailing list used for User Communication and Updates <kal...@li...> wrote: > I think you can get it by numerous ways in KALDI, one of them is : > > lattice-align-words $lang/phones/word_boundary.int $model ark:- ark:- \| > \ > nbest-to-ctm ark:- - \| \ > utils/int2sym.pl -f 5 $lang/words.txt \| \ > $filter_cmd '>' $dir/ctm > > Note that file $lang/phones/word_boundary.int is a phone id to > position-information > map file, you should have it if you follow the standard training procedure > > Please check the following file: > > egs/tidigits/s5/steps/get_train_ctm.sh > > Haihua > > > On Tue, Apr 16, 2013 at 6:14 PM, Mailing list used for User Communication > and Updates <kal...@li...> wrote: > >> Hi, >> >> you can have a look at the commands toward the end of >> egs/rm/s1/steps/decode_tri1.sh (ali-to-phones, phones-to-prons, >> prons-to-wordali). This AFAIK is the "old method" to obtain this >> information. Dan mentioned some time ago, that there are new tools >> that can be used to achieve the same, using the identity of the word >> position dependent phones, but I never had the need to use this new >> method. >> >> Vassil >> >> On Tue, Apr 16, 2013 at 1:00 PM, Mailing list used for User >> Communication and Updates <kal...@li...> wrote: >> > Hi, >> > >> > we recently started trying out kaldi here. Right now we do have a >> system up >> > and running based on the switchboard recipes, using our own (German) >> data. >> > >> > The problem we currently have is: How can you get timing information for >> > every recognized word from the decoding step? To illustrate, what we >> have >> > are those transcriptions, output by the decoder: >> > >> > [...] >> > SD0041_06 die letzten drei Spiele >> > [...] >> > >> > what we want is actually, for that specific example, something like >> > 003 025 die >> > 034 063 letzten >> > 070 082 drei >> > 088 104 Spiele >> > so that we have start and end frame number or time stamp for every word >> that >> > was recognized (numbers here are made up) >> > >> > How can we get that? >> > >> > Thanks in advance, >> > Thomas >> > >> > >> ------------------------------------------------------------------------------ >> > Precog is a next-generation analytics platform capable of advanced >> > analytics on semi-structured data. The platform includes APIs for >> building >> > apps and a phenomenal toolset for data science. Developers can use >> > our toolset for easy data analysis & visualization. Get a free account! >> > http://www2.precog.com/precogplatform/slashdotnewsletter >> > _______________________________________________ >> > Kaldi-users mailing list >> > Kal...@li... >> > https://lists.sourceforge.net/lists/listinfo/kaldi-users >> > >> >> >> ------------------------------------------------------------------------------ >> Precog is a next-generation analytics platform capable of advanced >> analytics on semi-structured data. The platform includes APIs for building >> apps and a phenomenal toolset for data science. Developers can use >> our toolset for easy data analysis & visualization. Get a free account! >> http://www2.precog.com/precogplatform/slashdotnewsletter >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users >> > > > > ------------------------------------------------------------------------------ > Precog is a next-generation analytics platform capable of advanced > analytics on semi-structured data. The platform includes APIs for building > apps and a phenomenal toolset for data science. Developers can use > our toolset for easy data analysis & visualization. Get a free account! > http://www2.precog.com/precogplatform/slashdotnewsletter > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-04-16 12:14:45
|
I think you can get it by numerous ways in KALDI, one of them is :
lattice-align-words $lang/phones/word_boundary.int $model ark:- ark:- \| \
nbest-to-ctm ark:- - \| \
utils/int2sym.pl -f 5 $lang/words.txt \| \
$filter_cmd '>' $dir/ctm
Note that file $lang/phones/word_boundary.int is a phone id to
position-information
map file, you should have it if you follow the standard training procedure
Please check the following file:
egs/tidigits/s5/steps/get_train_ctm.sh
Haihua
On Tue, Apr 16, 2013 at 6:14 PM, Mailing list used for User Communication
and Updates <kal...@li...> wrote:
> Hi,
>
> you can have a look at the commands toward the end of
> egs/rm/s1/steps/decode_tri1.sh (ali-to-phones, phones-to-prons,
> prons-to-wordali). This AFAIK is the "old method" to obtain this
> information. Dan mentioned some time ago, that there are new tools
> that can be used to achieve the same, using the identity of the word
> position dependent phones, but I never had the need to use this new
> method.
>
> Vassil
>
> On Tue, Apr 16, 2013 at 1:00 PM, Mailing list used for User
> Communication and Updates <kal...@li...> wrote:
> > Hi,
> >
> > we recently started trying out kaldi here. Right now we do have a system
> up
> > and running based on the switchboard recipes, using our own (German)
> data.
> >
> > The problem we currently have is: How can you get timing information for
> > every recognized word from the decoding step? To illustrate, what we have
> > are those transcriptions, output by the decoder:
> >
> > [...]
> > SD0041_06 die letzten drei Spiele
> > [...]
> >
> > what we want is actually, for that specific example, something like
> > 003 025 die
> > 034 063 letzten
> > 070 082 drei
> > 088 104 Spiele
> > so that we have start and end frame number or time stamp for every word
> that
> > was recognized (numbers here are made up)
> >
> > How can we get that?
> >
> > Thanks in advance,
> > Thomas
> >
> >
> ------------------------------------------------------------------------------
> > Precog is a next-generation analytics platform capable of advanced
> > analytics on semi-structured data. The platform includes APIs for
> building
> > apps and a phenomenal toolset for data science. Developers can use
> > our toolset for easy data analysis & visualization. Get a free account!
> > http://www2.precog.com/precogplatform/slashdotnewsletter
> > _______________________________________________
> > Kaldi-users mailing list
> > Kal...@li...
> > https://lists.sourceforge.net/lists/listinfo/kaldi-users
> >
>
>
> ------------------------------------------------------------------------------
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
> _______________________________________________
> Kaldi-users mailing list
> Kal...@li...
> https://lists.sourceforge.net/lists/listinfo/kaldi-users
>
|
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-04-16 11:42:25
|
Thomas- take a look at egs/swbd/s5/local/score_sclite.sh within the "if [ $stage -le 0 ]; " statement where CTMs are generated. What you need is basically the CTM format used in NIST scoring. -Arnab On Tue, Apr 16, 2013 at 11:00 AM, Mailing list used for User Communication and Updates <kal...@li...> wrote: > Hi, > > we recently started trying out kaldi here. Right now we do have a system up > and running based on the switchboard recipes, using our own (German) data. > > The problem we currently have is: How can you get timing information for > every recognized word from the decoding step? To illustrate, what we have > are those transcriptions, output by the decoder: > > [...] > SD0041_06 die letzten drei Spiele > [...] > > what we want is actually, for that specific example, something like > 003 025 die > 034 063 letzten > 070 082 drei > 088 104 Spiele > so that we have start and end frame number or time stamp for every word that > was recognized (numbers here are made up) > > How can we get that? > > Thanks in advance, > Thomas > > ------------------------------------------------------------------------------ > Precog is a next-generation analytics platform capable of advanced > analytics on semi-structured data. The platform includes APIs for building > apps and a phenomenal toolset for data science. Developers can use > our toolset for easy data analysis & visualization. Get a free account! > http://www2.precog.com/precogplatform/slashdotnewsletter > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-04-16 10:14:51
|
Hi, you can have a look at the commands toward the end of egs/rm/s1/steps/decode_tri1.sh (ali-to-phones, phones-to-prons, prons-to-wordali). This AFAIK is the "old method" to obtain this information. Dan mentioned some time ago, that there are new tools that can be used to achieve the same, using the identity of the word position dependent phones, but I never had the need to use this new method. Vassil On Tue, Apr 16, 2013 at 1:00 PM, Mailing list used for User Communication and Updates <kal...@li...> wrote: > Hi, > > we recently started trying out kaldi here. Right now we do have a system up > and running based on the switchboard recipes, using our own (German) data. > > The problem we currently have is: How can you get timing information for > every recognized word from the decoding step? To illustrate, what we have > are those transcriptions, output by the decoder: > > [...] > SD0041_06 die letzten drei Spiele > [...] > > what we want is actually, for that specific example, something like > 003 025 die > 034 063 letzten > 070 082 drei > 088 104 Spiele > so that we have start and end frame number or time stamp for every word that > was recognized (numbers here are made up) > > How can we get that? > > Thanks in advance, > Thomas > > ------------------------------------------------------------------------------ > Precog is a next-generation analytics platform capable of advanced > analytics on semi-structured data. The platform includes APIs for building > apps and a phenomenal toolset for data science. Developers can use > our toolset for easy data analysis & visualization. Get a free account! > http://www2.precog.com/precogplatform/slashdotnewsletter > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-04-16 10:00:33
|
Hi, we recently started trying out kaldi here. Right now we do have a system up and running based on the switchboard recipes, using our own (German) data. The problem we currently have is: How can you get timing information for every recognized word from the decoding step? To illustrate, what we have are those transcriptions, output by the decoder: [...] SD0041_06 die letzten drei Spiele [...] what we want is actually, for that specific example, something like 003 025 die 034 063 letzten 070 082 drei 088 104 Spiele so that we have start and end frame number or time stamp for every word that was recognized (numbers here are made up) How can we get that? Thanks in advance, Thomas |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-03-21 15:21:17
|
You may notice that the svn commands now require a password. You can get password-less access by posting your ssh keys to the link in the account services page: https://sourceforge.net/account/services Apparently the keys are synced every 10mins, but it may take longer. So you won't get password-less access immediately after posting the keys. If what I just said made no sense, you can look at the following on how to generate ssh keys, etc. https://sourceforge.net/apps/trac/sourceforge/wiki/SSH%20keys On Thu, Mar 21, 2013 at 12:51 AM, Daniel Povey <dp...@gm...> wrote: > > > > Hi all, > > This evening something broke on Sourceforge's svn server (error message: > svn: Can't find a temporary directory: Internal error) and I decided that > now would be a good time to move Kaldi over to the "new" version of > Sourceforge's software. I have now done this, and it works, but because of > the move, the old repositories are no longer valid. (I don't know if you > will still be be able to access them after they fix the server issue, but > regardless, they would be out of date). > > In order to switch your repositories over to the "new" location, please do > as follows. > > If you DO NOT need to commit changes to the repository, change directory to > the top level of where you checked Kaldi out, and do: > > svn switch --relocate https://kaldi.svn.sourceforge.net/svnroot/kaldi > svn://svn.code.sf.net/p/kaldi/code > > If you DO need to commit changes at some point, then do as follows, with > USERNAME replaced with your Sourceforge username: > > svn switch --relocate https://kaldi.svn.sourceforge.net/svnroot/kaldi > svn+ssh://USE...@sv.../p/kaldi/code/ > > Note: if at some point in the future you need to change your repository from > a non-commit to commit-is-possible kind of repository, you can always do as > follows: > svn switch --relocate svn://svn.code.sf.net/p/kaldi/code > svn+ssh://USE...@sv.../p/kaldi/code/ > > BTW, I will hopefully have time to give a proper update in a few days. I > have quite a lot to report but no time right now to prepare something. > > Dan > > > > > ------------------------------------------------------------------------------ > Everyone hates slow websites. So do we. > Make your web apps faster with AppDynamics > Download AppDynamics Lite for free today: > http://p.sf.net/sfu/appdyn_d2d_mar > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-03-21 00:51:19
|
Hi all, This evening something broke on Sourceforge's svn server (error message: svn: Can't find a temporary directory: Internal error) and I decided that now would be a good time to move Kaldi over to the "new" version of Sourceforge's software. I have now done this, and it works, but because of the move, the old repositories are no longer valid. (I don't know if you will still be be able to access them after they fix the server issue, but regardless, they would be out of date). In order to switch your repositories over to the "new" location, please do as follows. If you DO NOT need to commit changes to the repository, change directory to the top level of where you checked Kaldi out, and do: svn switch --relocate https://kaldi.svn.sourceforge.net/svnroot/kaldisvn:// svn.code.sf.net/p/kaldi/code If you DO need to commit changes at some point, then do as follows, with USERNAME replaced with your Sourceforge username: svn switch --relocate https://kaldi.svn.sourceforge.net/svnroot/kaldisvn+ssh:// USE...@sv.../p/kaldi/code/ Note: if at some point in the future you need to change your repository from a non-commit to commit-is-possible kind of repository, you can always do as follows: svn switch --relocate svn://svn.code.sf.net/p/kaldi/code svn+ssh:// USE...@sv.../p/kaldi/code/ BTW, I will hopefully have time to give a proper update in a few days. I have quite a lot to report but no time right now to prepare something. Dan |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-02-11 22:40:09
|
Another approach is to remove the optional silences phones from the lexicon transducer and to either add silence loops to each state of the language model or compose the language model with the stochastic silence transducer as described in "A generalized construction of integrated speech recognition transducers". For both of these case silence will need to appears as word in your lexicon. Paul On 12 February 2013 01:53, Mailing list used for User Communication and Updates <kal...@li...> wrote: > > [forwarding the answer to a question in case it's useful for others; this > is just an archiving exercise.] > > On Mon, Feb 11, 2013 at 11:52 AM, Daniel Povey <dp...@gm...> wrote: > >> It is normal not to have a symbol for silence. >> The way the FST composition works is, the LM has no silence symbols in >> it, and we compose with a lexicon that adds optional silence phones (but >> does not put silence symbols in). >> Probably the easiest way to get the silence info, if you are using the s5 >> recipe, is to use the "lattice-align-words" program. This program is >> basically for lining the input and output symbols up so that the input >> symbols (pdfs) for a word are on the same arc as that word; it puts the >> optional silences on separate arcs. You can specify what symbol you want it >> to use for silence; by default it is epsilon (0). You would have to make >> sure there is a symbol in your words.txt for silence. >> >> Dan >> >> >> >> On Mon, Feb 11, 2013 at 11:13 AM <someone> wrote: >> >>> Hi Dan & Arnab,**** >>> >>> ** ** >>> >>> I do not see !SIL in the output of WSJ0 decoding.**** >>> >>> ** ** >>> >>> Is it OK?**** >>> >>> ** ** >>> >>> What should I change in order to get !SIL information in the decoding?** >>> ** >>> >>> ** ** >>> >>> Thanks,**** >>> >>> [name removed] >>> >> > > > ------------------------------------------------------------------------------ > Free Next-Gen Firewall Hardware Offer > Buy your Sophos next-gen firewall before the end March 2013 > and get the hardware for free! Learn more. > http://p.sf.net/sfu/sophos-d2d-feb > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2013-02-11 16:53:56
|
[forwarding the answer to a question in case it's useful for others; this is just an archiving exercise.] On Mon, Feb 11, 2013 at 11:52 AM, Daniel Povey <dp...@gm...> wrote: > It is normal not to have a symbol for silence. > The way the FST composition works is, the LM has no silence symbols in it, > and we compose with a lexicon that adds optional silence phones (but does > not put silence symbols in). > Probably the easiest way to get the silence info, if you are using the s5 > recipe, is to use the "lattice-align-words" program. This program is > basically for lining the input and output symbols up so that the input > symbols (pdfs) for a word are on the same arc as that word; it puts the > optional silences on separate arcs. You can specify what symbol you want it > to use for silence; by default it is epsilon (0). You would have to make > sure there is a symbol in your words.txt for silence. > > Dan > > > > On Mon, Feb 11, 2013 at 11:13 AM <someone> wrote: > >> Hi Dan & Arnab,**** >> >> ** ** >> >> I do not see !SIL in the output of WSJ0 decoding.**** >> >> ** ** >> >> Is it OK?**** >> >> ** ** >> >> What should I change in order to get !SIL information in the decoding?*** >> * >> >> ** ** >> >> Thanks,**** >> >> [name removed] >> > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-06-01 00:52:49
|
Thank you for your very quick reply. I can see I have used version 1 of Kaldi by mistake and am now setting up the trunk version at Rev: 981. Neil On 05/31/2012 05:38 PM, Mailing list used for User Communication and Updates wrote: > When did you get this? I don't see anything like that line you > replaced in configure. Do "svn info" at the top level and tell me > what you get. We're right now on openfst-1.2.10. > You may have to do "svn update". > Or maybe you're using the "v1"? We recommend to use the trunk. > Dan > > On Thu, May 31, 2012 at 7:23 PM, Mailing list used for User > Communication and Updates<kal...@li...> wrote: >> Having difficulty getting openfst with Kaldi installed. See right after >> the **************** lines. >> >> ubuntu 12.04, gcc 4.6.3 >> >> sph2pipe - Use instructions in tools/INSTALL >> dot - apt-get install graphviz >> irstlm - Use instructions in tools/INSTALL >> ATLAS - apt-get install libatlas-base-dev >> cp /usr/include/atlas/clapack.h /usr/include/ >> >> sclite - Use instructions in tools/INSTALL except do >> tar -xovjf sctk-2.4.0-20091110-0958.tar.bz2 >> cd sctk-2.4.0 ... >> >> Make the following change noted in this post. >> Re: [Kaldi-developers] Error while compiling sclite >> From: Daniel Povey<dpovey@gm...> - 2011-12-25 21:26 >> >> Replace Filter::Filter with ::Filter >> There are also two instances in recording.cpp >> >> asclite/core/recording.h >> asclite/core/recording.cpp >> >> openfst >> - Using the tools/INSTALL instructions obtains: >> **************** >> >> kaldi-1/tools/openfst-1.2.7/src/script/./../include/fst/generic-register.h:77: undefined reference to `dlopen' >> >> kaldi-1/tools/openfst-1.2.7/src/script/./../include/fst/generic-register.h:79: undefined reference to `dlerror' >> >> dlopen and dlerror are prototyped in /usr/include/dlfcn.h that >> configure finds. >> >> - Install openfst-1.3.2.tar.gz from >> http://www.openfst.org/twiki/bin/view/FST/FstDownload. 'make install' >> did not copy the bin files to /usr/local/bin. As root: >> cd openfst-1.3.2/bin >> cp fst* /usr/local/bin/ >> >> change the symlink in kaldi-1/tools to point to the new >> openfst-1.3.2 folder. >> cd kaldi-1/tools >> rm openfst >> ln -s myPath/openfst-1.3.2 openfst >> >> configure kaldi >> cd kaldi-1/src >> Change the configure file according to the following diff. >> diff configure configure_original >> 28c28 >> < #ATLASROOT=../tools/ATLAS/ >> --- >> > ATLASROOT=../tools/ATLAS/ >> 30,31c30 >> < ATLASROOT=/usr >> < >> --- >> 173,174d171 >> < elif [ -f /usr/lib/atlas-base/libatlas.a ]; then >> < ATLASLIBDIR=/usr/lib/atlas-base >> >> ./configure >> Gives the following exception output: >> **************** >> Checking for OpenFST library ... >> Using OpenFST library in ../tools/openfst >> ../tools/minimize.h ../tools/openfst/include/fst/minimize.h differ: byte >> 15, line 2 >> **Warning: ../tools/minimize.h and >> ../tools/openfst/include/fst/minimize.h differ: >> patch not applied? FST tools may not work. See ../tools/INSTALL >> >> I am thinking that making changes to use the latest openfst may be a >> good direction. Or it may be that the install is OK, except that the >> 'path' seems ominous. Or it may be easier to get openfst-1.2.7 going in >> the short run. >> >> I see that some of you are are using Kaldi on Ubuntu and may be able to >> provide a solution. >> >> Neil >> >> >> ------------------------------------------------------------------------------ >> Live Security Virtual Conference >> Exclusive live event will cover all the ways today's security and >> threat landscape has changed and how IT managers can respond. Discussions >> will include endpoint security, mobile security and the latest in malware >> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ >> _______________________________________________ >> Kaldi-users mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-users > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-05-31 23:38:46
|
When did you get this? I don't see anything like that line you replaced in configure. Do "svn info" at the top level and tell me what you get. We're right now on openfst-1.2.10. You may have to do "svn update". Or maybe you're using the "v1"? We recommend to use the trunk. Dan On Thu, May 31, 2012 at 7:23 PM, Mailing list used for User Communication and Updates <kal...@li...> wrote: > Having difficulty getting openfst with Kaldi installed. See right after > the **************** lines. > > ubuntu 12.04, gcc 4.6.3 > > sph2pipe - Use instructions in tools/INSTALL > dot - apt-get install graphviz > irstlm - Use instructions in tools/INSTALL > ATLAS - apt-get install libatlas-base-dev > cp /usr/include/atlas/clapack.h /usr/include/ > > sclite - Use instructions in tools/INSTALL except do > tar -xovjf sctk-2.4.0-20091110-0958.tar.bz2 > cd sctk-2.4.0 ... > > Make the following change noted in this post. > Re: [Kaldi-developers] Error while compiling sclite > From: Daniel Povey <dpovey@gm...> - 2011-12-25 21:26 > > Replace Filter::Filter with ::Filter > There are also two instances in recording.cpp > > asclite/core/recording.h > asclite/core/recording.cpp > > openfst > - Using the tools/INSTALL instructions obtains: > **************** > > kaldi-1/tools/openfst-1.2.7/src/script/./../include/fst/generic-register.h:77: undefined reference to `dlopen' > > kaldi-1/tools/openfst-1.2.7/src/script/./../include/fst/generic-register.h:79: undefined reference to `dlerror' > > dlopen and dlerror are prototyped in /usr/include/dlfcn.h that > configure finds. > > - Install openfst-1.3.2.tar.gz from > http://www.openfst.org/twiki/bin/view/FST/FstDownload. 'make install' > did not copy the bin files to /usr/local/bin. As root: > cd openfst-1.3.2/bin > cp fst* /usr/local/bin/ > > change the symlink in kaldi-1/tools to point to the new > openfst-1.3.2 folder. > cd kaldi-1/tools > rm openfst > ln -s myPath/openfst-1.3.2 openfst > > configure kaldi > cd kaldi-1/src > Change the configure file according to the following diff. > diff configure configure_original > 28c28 > < #ATLASROOT=../tools/ATLAS/ > --- > > ATLASROOT=../tools/ATLAS/ > 30,31c30 > < ATLASROOT=/usr > < > --- > 173,174d171 > < elif [ -f /usr/lib/atlas-base/libatlas.a ]; then > < ATLASLIBDIR=/usr/lib/atlas-base > > ./configure > Gives the following exception output: > **************** > Checking for OpenFST library ... > Using OpenFST library in ../tools/openfst > ../tools/minimize.h ../tools/openfst/include/fst/minimize.h differ: byte > 15, line 2 > **Warning: ../tools/minimize.h and > ../tools/openfst/include/fst/minimize.h differ: > patch not applied? FST tools may not work. See ../tools/INSTALL > > I am thinking that making changes to use the latest openfst may be a > good direction. Or it may be that the install is OK, except that the > 'path' seems ominous. Or it may be easier to get openfst-1.2.7 going in > the short run. > > I see that some of you are are using Kaldi on Ubuntu and may be able to > provide a solution. > > Neil > > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-05-31 23:23:53
|
Having difficulty getting openfst with Kaldi installed. See right after
the **************** lines.
ubuntu 12.04, gcc 4.6.3
sph2pipe - Use instructions in tools/INSTALL
dot - apt-get install graphviz
irstlm - Use instructions in tools/INSTALL
ATLAS - apt-get install libatlas-base-dev
cp /usr/include/atlas/clapack.h /usr/include/
sclite - Use instructions in tools/INSTALL except do
tar -xovjf sctk-2.4.0-20091110-0958.tar.bz2
cd sctk-2.4.0 ...
Make the following change noted in this post.
Re: [Kaldi-developers] Error while compiling sclite
From: Daniel Povey <dpovey@gm...> - 2011-12-25 21:26
Replace Filter::Filter with ::Filter
There are also two instances in recording.cpp
asclite/core/recording.h
asclite/core/recording.cpp
openfst
- Using the tools/INSTALL instructions obtains:
****************
kaldi-1/tools/openfst-1.2.7/src/script/./../include/fst/generic-register.h:77: undefined reference to `dlopen'
kaldi-1/tools/openfst-1.2.7/src/script/./../include/fst/generic-register.h:79: undefined reference to `dlerror'
dlopen and dlerror are prototyped in /usr/include/dlfcn.h that
configure finds.
- Install openfst-1.3.2.tar.gz from
http://www.openfst.org/twiki/bin/view/FST/FstDownload. 'make install'
did not copy the bin files to /usr/local/bin. As root:
cd openfst-1.3.2/bin
cp fst* /usr/local/bin/
change the symlink in kaldi-1/tools to point to the new
openfst-1.3.2 folder.
cd kaldi-1/tools
rm openfst
ln -s myPath/openfst-1.3.2 openfst
configure kaldi
cd kaldi-1/src
Change the configure file according to the following diff.
diff configure configure_original
28c28
< #ATLASROOT=../tools/ATLAS/
---
> ATLASROOT=../tools/ATLAS/
30,31c30
< ATLASROOT=/usr
<
---
173,174d171
< elif [ -f /usr/lib/atlas-base/libatlas.a ]; then
< ATLASLIBDIR=/usr/lib/atlas-base
./configure
Gives the following exception output:
****************
Checking for OpenFST library ...
Using OpenFST library in ../tools/openfst
../tools/minimize.h ../tools/openfst/include/fst/minimize.h differ: byte
15, line 2
**Warning: ../tools/minimize.h and
../tools/openfst/include/fst/minimize.h differ:
patch not applied? FST tools may not work. See ../tools/INSTALL
I am thinking that making changes to use the latest openfst may be a
good direction. Or it may be that the install is OK, except that the
'path' seems ominous. Or it may be easier to get openfst-1.2.7 going in
the short run.
I see that some of you are are using Kaldi on Ubuntu and may be able to
provide a solution.
Neil
|
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-04-23 01:30:16
|
I don't like having word insertion penalties, as it's one more parameter to tune and as far as I know it doesn't buy you much-- this is why it's not directly supported. An easy way to do it would be to use fstprint and fstcompile to convert HCLG.fst to text and back, and use an awk script to, whenever you see that there is a word label (output symbol is nonzero), add a constant to the cost (the last field on the line). Dan > Hi, Dan > I do appreciate your direction. > Could you tell me how to setup word insertion penalty in > gmm-decode-faster? > > > Best regards, > > hlwang > > > At 2012-04-20 10:46:19,"Mailing list used for User Communication and Updates" <kal...@li...> wrote: >>> I am going to compare the decoding performance using difference >>> parameters by Kaldi. >> >>That's great! >> >>> I have got some results, but I can't obtain an exact conclusion. >>> 1) What is the biggest difference between gmm-decode-faster and >>> gmm-decode-kaldi? gmm-decode-kaldi can generate lattice, whilst >>> gmm-decode-faster can't. I found the performances of gmm-decode-faster and >>> gmm-decode-kaldi are almost same. >> >>I think you mean the other way round, no? Anyway the normal method to >>do lattice >>generation is gmm-latgen-faster, which is derived from >>gmm-decode-faster. gmm-decode-kaldi was >>intended to be a highly optimized decoder, but in the end some of >>those optimization techniques did >>not work as well as we had hoped. >> >>> 2) How do the decoding parameters of gmm-decode-faster effect the >>> performance? I found the acoustic-scale can obviously effect the >>> performance, but other parameters almost have no effect. In general, >>> max-active=16000, beam=16.0, is the best one. >> >>I would be cautious about saying that something is the best one. >>Surely it depends >>what point in the speed/accuracy tradeoff you want. >> >>> 3) The beam in gmm-decode-faster is difference with that in Hvite in >>> the scale. How to adjust the beam to get an optimal result. >> >>The difference is the acoustic scale. The beam divided by the >>acoustic scale should be >>comparable to a HVite beam. >> >>> 4) How to optimize the generated WFST to improve decoding >>> performance? What are the important matters needing to pay attention in >>> composition, determination and minimization operations in Kaldi's WFST >>> generation? >> >>This is a very big subject. There are many methods. >>One is simply using a smaller WFST by using a smaller >>language model, and then rescoring with a larger one. There are two paths here: >>either do lattice rescoring, or use the "biglm" decoder. I would >>appreciate it if you could >>find out which approach is faster, as I have not had time. In >>general, I think such >>experiments should take the form of generating speed/accuracy curves >>and seeing which >>curve lies above or below the other. >> >>Search for "hbka.pdf" and read it, to learn the general principles and >>theory of WFSTs-- although >>that is a rather technical presentation. >> >>Dan >> >> >>> hlwang >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> For Developers, A Lot Can Happen In A Second. >>> Boundary is the first to Know...and Tell You. >>> Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! >>> http://p.sf.net/sfu/Boundary-d2dvs2 >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >> >>------------------------------------------------------------------------------ >>For Developers, A Lot Can Happen In A Second. >>Boundary is the first to Know...and Tell You. >>Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! >>http://p.sf.net/sfu/Boundary-d2dvs2 >>_______________________________________________ >>Kaldi-users mailing list >>Kal...@li... >>https://lists.sourceforge.net/lists/listinfo/kaldi-users > > > > ________________________________ > 网易Lofter，专注兴趣，分享创作！ > ------------------------------------------------------------------------------ > For Developers, A Lot Can Happen In A Second. > Boundary is the first to Know...and Tell You. > Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! > http://p.sf.net/sfu/Boundary-d2dvs2 > > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-04-23 01:27:06
|
Hi, Dan
I do appreciate your direction.
Could you tell me how to setup word insertion penalty in gmm-decode-faster?
Best regards,
hlwang
At 2012-04-20 10:46:19,"Mailing list used for User Communication and Updates" <kal...@li...> wrote:
>> I am going to compare the decoding performance using difference
>> parameters by Kaldi.
>
>That's great!
>
>> I have got some results, but I can't obtain an exact conclusion.
>> 1) What is the biggest difference between gmm-decode-faster and
>> gmm-decode-kaldi? gmm-decode-kaldi can generate lattice, whilst
>> gmm-decode-faster can't. I found the performances of gmm-decode-faster and
>> gmm-decode-kaldi are almost same.
>
>I think you mean the other way round, no? Anyway the normal method to
>do lattice
>generation is gmm-latgen-faster, which is derived from
>gmm-decode-faster. gmm-decode-kaldi was
>intended to be a highly optimized decoder, but in the end some of
>those optimization techniques did
>not work as well as we had hoped.
>
>> 2) How do the decoding parameters of gmm-decode-faster effect the
>> performance? I found the acoustic-scale can obviously effect the
>> performance, but other parameters almost have no effect. In general,
>> max-active=16000, beam=16.0, is the best one.
>
>I would be cautious about saying that something is the best one.
>Surely it depends
>what point in the speed/accuracy tradeoff you want.
>
>> 3) The beam in gmm-decode-faster is difference with that in Hvite in
>> the scale. How to adjust the beam to get an optimal result.
>
>The difference is the acoustic scale. The beam divided by the
>acoustic scale should be
>comparable to a HVite beam.
>
>> 4) How to optimize the generated WFST to improve decoding
>> performance? What are the important matters needing to pay attention in
>> composition, determination and minimization operations in Kaldi's WFST
>> generation?
>
>This is a very big subject. There are many methods.
>One is simply using a smaller WFST by using a smaller
>language model, and then rescoring with a larger one. There are two paths here:
>either do lattice rescoring, or use the "biglm" decoder. I would
>appreciate it if you could
>find out which approach is faster, as I have not had time. In
>general, I think such
>experiments should take the form of generating speed/accuracy curves
>and seeing which
>curve lies above or below the other.
>
>Search for "hbka.pdf" and read it, to learn the general principles and
>theory of WFSTs-- although
>that is a rather technical presentation.
>
>Dan
>
>
>> hlwang
>>
>>
>>
>> ------------------------------------------------------------------------------
>> For Developers, A Lot Can Happen In A Second.
>> Boundary is the first to Know...and Tell You.
>> Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
>> http://p.sf.net/sfu/Boundary-d2dvs2
>> _______________________________________________
>> Kaldi-users mailing list
>> Kal...@li...
>> https://lists.sourceforge.net/lists/listinfo/kaldi-users
>>
>
>------------------------------------------------------------------------------
>For Developers, A Lot Can Happen In A Second.
>Boundary is the first to Know...and Tell You.
>Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
>http://p.sf.net/sfu/Boundary-d2dvs2
>_______________________________________________
>Kaldi-users mailing list
>Kal...@li...
>https://lists.sourceforge.net/lists/listinfo/kaldi-users |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-04-20 02:46:26
|
> I am going to compare the decoding performance using difference > parameters by Kaldi. That's great! > I have got some results, but I can't obtain an exact conclusion. > 1) What is the biggest difference between gmm-decode-faster and > gmm-decode-kaldi? gmm-decode-kaldi can generate lattice, whilst > gmm-decode-faster can't. I found the performances of gmm-decode-faster and > gmm-decode-kaldi are almost same. I think you mean the other way round, no? Anyway the normal method to do lattice generation is gmm-latgen-faster, which is derived from gmm-decode-faster. gmm-decode-kaldi was intended to be a highly optimized decoder, but in the end some of those optimization techniques did not work as well as we had hoped. > 2) How do the decoding parameters of gmm-decode-faster effect the > performance? I found the acoustic-scale can obviously effect the > performance, but other parameters almost have no effect. In general, > max-active=16000, beam=16.0, is the best one. I would be cautious about saying that something is the best one. Surely it depends what point in the speed/accuracy tradeoff you want. > 3) The beam in gmm-decode-faster is difference with that in Hvite in > the scale. How to adjust the beam to get an optimal result. The difference is the acoustic scale. The beam divided by the acoustic scale should be comparable to a HVite beam. > 4) How to optimize the generated WFST to improve decoding > performance? What are the important matters needing to pay attention in > composition, determination and minimization operations in Kaldi's WFST > generation? This is a very big subject. There are many methods. One is simply using a smaller WFST by using a smaller language model, and then rescoring with a larger one. There are two paths here: either do lattice rescoring, or use the "biglm" decoder. I would appreciate it if you could find out which approach is faster, as I have not had time. In general, I think such experiments should take the form of generating speed/accuracy curves and seeing which curve lies above or below the other. Search for "hbka.pdf" and read it, to learn the general principles and theory of WFSTs-- although that is a rather technical presentation. Dan > hlwang > > > > ------------------------------------------------------------------------------ > For Developers, A Lot Can Happen In A Second. > Boundary is the first to Know...and Tell You. > Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! > http://p.sf.net/sfu/Boundary-d2dvs2 > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-04-20 01:23:22
|
Hi, all,
I am going to compare the decoding performance using difference parameters by Kaldi.
I have got some results, but I can't obtain an exact conclusion.
1) What is the biggest difference between gmm-decode-faster and gmm-decode-kaldi? gmm-decode-kaldi can generate lattice, whilst gmm-decode-faster can't. I found the performances of gmm-decode-faster and gmm-decode-kaldi are almost same.
2) How do the decoding parameters of gmm-decode-faster effect the performance? I found the acoustic-scale can obviously effect the performance, but other parameters almost have no effect. In general, max-active=16000, beam=16.0, is the best one.
3) The beam in gmm-decode-faster is difference with that in Hvite in the scale. How to adjust the beam to get an optimal result.
4) How to optimize the generated WFST to improve decoding performance? What are the important matters needing to pay attention in composition, determination and minimization operations in Kaldi's WFST generation?
Could anybody who have used these parameters setup give me some instruction？
Best regards,
hlwang |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-03-22 14:07:16
|
Hi, Dan
Thanks for your answer!
Best Regards,
hlwang
At 2012-03-22 21:49:21,"Mailing list used for User Communication and Updates" <kal...@li...> wrote:
>Kaldi does not support within-word triphones. Probably this is the
>reason why it failed. There is no way to do this-- you can only do
>across-word triphones.
>
>Dan
>
>2012/3/22 Mailing list used for User Communication and Updates
><kal...@li...>:
>> Hi, all,
>> Have anybody used within-word triphone trained by HTK in Kaldi?
>> I successfully converted the within-word triphone trained by HTK into
>> Kaldi format. But when using the model to decode, an error occurred as
>> following:
>> KALDI_ASSERT: at
>> gmm-decode-faster:LogLikelihoodZeroBased:decodable-am-diag-gmm.cc:28,
>> failed: static_cast<size_t>(state) < static_cast<size_t>(NumIndices())
>> Stack trace is:
>> kaldi::KaldiGetStackTrace()
>> kaldi::KaldiAssertFailure_(char const*, char const*, int, char
>> const*)
>> kaldi::DecodableAmDiagGmmUnmapped::LogLikelihoodZeroBased(int, int)
>> kaldi::DecodableAmDiagGmmScaled::LogLikelihood(int, int)
>> kaldi::FasterDecoder::ProcessEmitting(kaldi::DecodableInterface*,
>> int)
>> kaldi::FasterDecoder::Decode(kaldi::DecodableInterface*)
>> ../../src/gmmbin/gmm-decode-faster(main+0x883) [0x55caa9]
>> /lib/libc.so.6(__libc_start_main+0xfd) [0x2b0633438c4d]
>> ../../src/gmmbin/gmm-decode-faster() [0x55bbe9]
>> Command terminated by signal 6
>>
>> Can anyone tell me the reason and solution for the problem?
>> By the way, can Kaldi tool be used to train within-word triphone
>> model? How to do this?
>>
>> Best regards,
>>
>> hlwang
>>
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> This SF email is sponsosred by:
>> Try Windows Azure free for 90 days Click Here
>> http://p.sf.net/sfu/sfd2d-msazure
>> _______________________________________________
>> Kaldi-users mailing list
>> Kal...@li...
>> https://lists.sourceforge.net/lists/listinfo/kaldi-users
>>
>
>------------------------------------------------------------------------------
>This SF email is sponsosred by:
>Try Windows Azure free for 90 days Click Here
>http://p.sf.net/sfu/sfd2d-msazure
>_______________________________________________
>Kaldi-users mailing list
>Kal...@li...
>https://lists.sourceforge.net/lists/listinfo/kaldi-users
|
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-03-22 13:49:31
|
Kaldi does not support within-word triphones. Probably this is the reason why it failed. There is no way to do this-- you can only do across-word triphones. Dan 2012/3/22 Mailing list used for User Communication and Updates <kal...@li...>: > Hi, all, > Have anybody used within-word triphone trained by HTK in Kaldi? > I successfully converted the within-word triphone trained by HTK into > Kaldi format. But when using the model to decode, an error occurred as > following: > KALDI_ASSERT: at > gmm-decode-faster:LogLikelihoodZeroBased:decodable-am-diag-gmm.cc:28, > failed: static_cast<size_t>(state) < static_cast<size_t>(NumIndices()) > Stack trace is: > kaldi::KaldiGetStackTrace() > kaldi::KaldiAssertFailure_(char const*, char const*, int, char > const*) > kaldi::DecodableAmDiagGmmUnmapped::LogLikelihoodZeroBased(int, int) > kaldi::DecodableAmDiagGmmScaled::LogLikelihood(int, int) > kaldi::FasterDecoder::ProcessEmitting(kaldi::DecodableInterface*, > int) > kaldi::FasterDecoder::Decode(kaldi::DecodableInterface*) > ../../src/gmmbin/gmm-decode-faster(main+0x883) [0x55caa9] > /lib/libc.so.6(__libc_start_main+0xfd) [0x2b0633438c4d] > ../../src/gmmbin/gmm-decode-faster() [0x55bbe9] > Command terminated by signal 6 > > Can anyone tell me the reason and solution for the problem? > By the way, can Kaldi tool be used to train within-word triphone > model? How to do this? > > Best regards, > > hlwang > > > > > > ------------------------------------------------------------------------------ > This SF email is sponsosred by: > Try Windows Azure free for 90 days Click Here > http://p.sf.net/sfu/sfd2d-msazure > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-03-22 12:39:17
|
Hi, all,
Have anybody used within-word triphone trained by HTK in Kaldi?
I successfully converted the within-word triphone trained by HTK into Kaldi format. But when using the model to decode, an error occurred as following:
KALDI_ASSERT: at gmm-decode-faster:LogLikelihoodZeroBased:decodable-am-diag-gmm.cc:28, failed: static_cast<size_t>(state) < static_cast<size_t>(NumIndices())
Stack trace is:
kaldi::KaldiGetStackTrace()
kaldi::KaldiAssertFailure_(char const*, char const*, int, char const*)
kaldi::DecodableAmDiagGmmUnmapped::LogLikelihoodZeroBased(int, int)
kaldi::DecodableAmDiagGmmScaled::LogLikelihood(int, int)
kaldi::FasterDecoder::ProcessEmitting(kaldi::DecodableInterface*, int)
kaldi::FasterDecoder::Decode(kaldi::DecodableInterface*)
../../src/gmmbin/gmm-decode-faster(main+0x883) [0x55caa9]
/lib/libc.so.6(__libc_start_main+0xfd) [0x2b0633438c4d]
../../src/gmmbin/gmm-decode-faster() [0x55bbe9]
Command terminated by signal 6
Can anyone tell me the reason and solution for the problem?
By the way, can Kaldi tool be used to train within-word triphone model? How to do this?
Best regards,
hlwang
|
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-03-09 02:17:25
|
OK, that is very old. Do "svn up" from the trunk directory (i.e. the top-level one). Dan 2012/3/8 Mailing list used for User Communication and Updates <kal...@li...>: > Hi, Dan > I am using the trunk version and Its revision number is 767. > > Thanks! > > hlwang > > > At 2012-03-08 23:24:36,"Mailing list used for User Communication and Updates" <kal...@li...> wrote: >>This is a script bug that was fixed at some point. What version >>(trunk or v1) and revision number are you using? >>Dan >> >>2012/3/8 Mailing list used for User Communication and Updates >><kal...@li...>: >>> Hi, my friends, >>> I meet a problem when I convert HTK acoustic model into Kaldi format. >>> My HTK acoustic model is cross-word triphone. >>> When I run the script "convert_htk.sh", a error occurs in the step of >>> perform command "gmm-init-trans". The error information is as follows: >>> gmm-init-trans kaldi.topo kaldi.am_gmm kaldi.tree kaldi.mdl >>> ERROR (gmm-init-trans:Read():kaldi-vector.cc:888) Failed to read vector >>> from stream. Expected "[" but got FV File position at start is 59, >>> currently 61 >>> ERROR (gmm-init-trans:Read():kaldi-vector.cc:888) Failed to read vector >>> from stream. Expected "[" but got FV File position at start is 59, >>> currently 61 >>> >>> What is the reason causing the error? Could anybody give me some cues >>> about the problem? >>> >>> >>> best regards, >>> >>> hlwang >>> >>> >>> >>> >>> ________________________________ >>> 三八节，网易微博替男人伸冤！ >>> ------------------------------------------------------------------------------ >>> Virtualization & Cloud Management Using Capacity Planning >>> Cloud computing makes use of virtualization - but cloud computing >>> also focuses on allowing computing to be delivered as a service. >>> http://www.accelacomm.com/jaw/sfnl/114/51521223/ >>> _______________________________________________ >>> Kaldi-users mailing list >>> Kal...@li... >>> https://lists.sourceforge.net/lists/listinfo/kaldi-users >>> >> >>------------------------------------------------------------------------------ >>Virtualization & Cloud Management Using Capacity Planning >>Cloud computing makes use of virtualization - but cloud computing >>also focuses on allowing computing to be delivered as a service. >>http://www.accelacomm.com/jaw/sfnl/114/51521223/ >>_______________________________________________ >>Kaldi-users mailing list >>Kal...@li... >>https://lists.sourceforge.net/lists/listinfo/kaldi-users > > > > > ------------------------------------------------------------------------------ > Virtualization & Cloud Management Using Capacity Planning > Cloud computing makes use of virtualization - but cloud computing > also focuses on allowing computing to be delivered as a service. > http://www.accelacomm.com/jaw/sfnl/114/51521223/ > _______________________________________________ > Kaldi-users mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-users > |
|
From: Mailing l. u. f. U. C. a. U. <kal...@li...> - 2012-03-09 01:00:02
|
Hi, Dan
I am using the trunk version and Its revision number is 767.
Thanks!
hlwang
At 2012-03-08 23:24:36,"Mailing list used for User Communication and Updates" <kal...@li...> wrote:
>This is a script bug that was fixed at some point. What version
>(trunk or v1) and revision number are you using?
>Dan
>
>2012/3/8 Mailing list used for User Communication and Updates
><kal...@li...>:
>> Hi, my friends,
>> I meet a problem when I convert HTK acoustic model into Kaldi format.
>> My HTK acoustic model is cross-word triphone.
>> When I run the script "convert_htk.sh", a error occurs in the step of
>> perform command "gmm-init-trans". The error information is as follows:
>> gmm-init-trans kaldi.topo kaldi.am_gmm kaldi.tree kaldi.mdl
>> ERROR (gmm-init-trans:Read():kaldi-vector.cc:888) Failed to read vector
>> from stream. Expected "[" but got FV File position at start is 59,
>> currently 61
>> ERROR (gmm-init-trans:Read():kaldi-vector.cc:888) Failed to read vector
>> from stream. Expected "[" but got FV File position at start is 59,
>> currently 61
>>
>> What is the reason causing the error? Could anybody give me some cues
>> about the problem?
>>
>>
>> best regards,
>>
>> hlwang
>>
>>
>>
>>
>> ________________________________
>> 三八节，网易微博替男人伸冤！
>> ------------------------------------------------------------------------------
>> Virtualization & Cloud Management Using Capacity Planning
>> Cloud computing makes use of virtualization - but cloud computing
>> also focuses on allowing computing to be delivered as a service.
>> http://www.accelacomm.com/jaw/sfnl/114/51521223/
>> _______________________________________________
>> Kaldi-users mailing list
>> Kal...@li...
>> https://lists.sourceforge.net/lists/listinfo/kaldi-users
>>
>
>------------------------------------------------------------------------------
>Virtualization & Cloud Management Using Capacity Planning
>Cloud computing makes use of virtualization - but cloud computing
>also focuses on allowing computing to be delivered as a service.
>http://www.accelacomm.com/jaw/sfnl/114/51521223/
>_______________________________________________
>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.