kaldi-developers — Kaldi Developers

Re: [Kaldi-developers] KALDI Speech Recognition POC

[Daniel P. <dp...@gm...>]

We don't really have any "application code"-- we generally run Kaldi
executables from the UNIX command line, in a batch/off-line mode, although
the libraries can in principle be used independently.  But all
speech-recognition is complex and you need to be willing to spend a while
learning.  I don't think there is any very easy off-the-shelf solution in
existence (for free).
Dan

On Wed, Jun 20, 2012 at 4:55 AM, Suresh Gorakala <
sur...@em...> wrote:

> Hi,
>
> Hi Would like to create a Proof Of Concept on Speech recognition.
> The application would be speaker independent,  Isolated word.
> AS I have been going through the documentation, I read that "KALDI is
> not for beginners",
> But I'm a novice in this area.
> Please do share any samples application code for easy understanding.
>
>
> Regards,
> Suresh Gorakala
>
> -----Original Message-----
> From: Arnab Ghoshal [mailto:ar...@gm...]
> Sent: 20 June 2012 12:37
> To: Suresh Gorakala
> Cc: kal...@li...
> Subject: Re: [Kaldi-developers] KALDI Speech Recognition POC
>
> What is a POC? To build a speech recognition system, you should look at
> the trunk/egs directory. Probably the best place for you to start will
> be egs/tidigits/s5. -Arnab
>
> On Tue, Jun 19, 2012 at 11:29 AM, Suresh Gorakala
> <sur...@em...> wrote:
> > Hi Team,
> >
> >
> >
> > I'm planning to do a POC of Speech recognition- speaker independent,
> > Isolated word - application, using KALDI toolkit.
> >
> > I'm new to speech recognition. Please advise me of how to go forward.
> >
> >
> >
> > Regards,
> >
> > Suresh G,
> >
> > Software Engineer,
> >
> > Innovation Team,
> >
> > Emirates Group -IT.
> >
> >
> >
> >
> > ----------------------------------------------------------------------
> > --------
> > 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-developers mailing list
> > Kal...@li...
> > https://lists.sourceforge.net/lists/listinfo/kaldi-developers
> >
>
>
> ------------------------------------------------------------------------------
> 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-developers mailing list
> Kal...@li...
> https://lists.sourceforge.net/lists/listinfo/kaldi-developers
>

[Kaldi-developers] Potential issue (and fix) in WSJ recipe s3

[Andrew R. <an...@cs...>]

Hi all,

I've run into a problem with language model training in  the wsj
training recipe s3.  During the LM training an error shows up that i'm
not quite sure how to fix.

Train LM 3gram min count
Getting raw N-gram counts
generating n grams
discount_ngrams: for n-gram order 1, D=0.000000, tau=0.000000 phi=1.000000
discount_ngrams: for n-gram order 2, D=0.000000, tau=0.000000 phi=1.000000
discount_ngrams: for n-gram order 3, D=1.000000, tau=0.000000 phi=1.000000
error: histories are not in sorted order, "?? ۰" > "? ??"
merge_ngrams: merge_ngrams.cc:141: void process_line(char*): Assertion
`comp > 0 || (comp == 0 &&
entry.predicted.compare(stack.back().predicted) >= 0)' failed.

Looking at the "error: ..." line in emacs rather than the console, to
see what the '?' characters actually were, it was clear that the issue
was with the sorting of the ngram tokens.

The recipe runs error free until
local/wsj_train_lms.sh

within this the line that gives the problem (first) is

train_lm.sh --arpa --lmtype 3gram-mincount $dir

digging deeper into train_lm.sh the line that generates the error is

gunzip -c $dir/train.gz | tail -n +$heldout_sents | \
 get_raw_ngrams 3 | sort | uniq -c  | uniq_to_ngrams | \
 sort | discount_ngrams $subdir/config.get_ngrams | \
 sort | merge_ngrams | gzip -c > $subdir/ngrams.gz

Suspecting this is a problem with how sort is working,  i tried sort
-n to see if it would fix the issue to no avail.

Ultimately the fix was to "export LC_ALL=C" to ensure POSIX style sorting.

This is clearly an environment problem, but I figured you guys would
want to know about it.  I'm doing this in bash on CentOS. (if more
environment information would be useful, let me know.)

Thanks very much for putting this tool together.  I'm really enjoying
getting to know it.

Andrew

Re: [Kaldi-developers] KALDI Speech Recognition POC

[Suresh G. <sur...@em...>]

Hi,

Hi Would like to create a Proof Of Concept on Speech recognition. 
The application would be speaker independent,  Isolated word.
AS I have been going through the documentation, I read that "KALDI is
not for beginners",
But I'm a novice in this area.
Please do share any samples application code for easy understanding.


Regards,
Suresh Gorakala

-----Original Message-----
From: Arnab Ghoshal [mailto:ar...@gm...] 
Sent: 20 June 2012 12:37
To: Suresh Gorakala
Cc: kal...@li...
Subject: Re: [Kaldi-developers] KALDI Speech Recognition POC

What is a POC? To build a speech recognition system, you should look at
the trunk/egs directory. Probably the best place for you to start will
be egs/tidigits/s5. -Arnab

On Tue, Jun 19, 2012 at 11:29 AM, Suresh Gorakala
<sur...@em...> wrote:
> Hi Team,
>
>
>
> I'm planning to do a POC of Speech recognition- speaker independent, 
> Isolated word - application, using KALDI toolkit.
>
> I'm new to speech recognition. Please advise me of how to go forward.
>
>
>
> Regards,
>
> Suresh G,
>
> Software Engineer,
>
> Innovation Team,
>
> Emirates Group -IT.
>
>
>
>
> ----------------------------------------------------------------------
> --------
> 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-developers mailing list
> Kal...@li...
> https://lists.sourceforge.net/lists/listinfo/kaldi-developers
>

Re: [Kaldi-developers] KALDI Speech Recognition POC

[Arnab G. <ar...@gm...>]

What is a POC? To build a speech recognition system, you should look
at the trunk/egs directory. Probably the best place for you to start
will be egs/tidigits/s5. -Arnab

On Tue, Jun 19, 2012 at 11:29 AM, Suresh Gorakala
<sur...@em...> wrote:
> Hi Team,
>
>
>
> I’m planning to do a POC of Speech recognition- speaker independent,
> Isolated word – application, using KALDI toolkit.
>
> I’m new to speech recognition. Please advise me of how to go forward.
>
>
>
> Regards,
>
> Suresh G,
>
> Software Engineer,
>
> Innovation Team,
>
> Emirates Group –IT.
>
>
>
>
> ------------------------------------------------------------------------------
> 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-developers mailing list
> Kal...@li...
> https://lists.sourceforge.net/lists/listinfo/kaldi-developers
>

[Kaldi-developers] KALDI Speech Recognition POC

[Suresh G. <sur...@em...>]

Hi Team,

 

I'm planning to do a POC of Speech recognition- speaker independent,
Isolated word - application, using KALDI toolkit.

I'm new to speech recognition. Please advise me of how to go forward.

 

Regards,

Suresh G,

Software Engineer,

Innovation Team,

Emirates Group -IT.

 

Re: [Kaldi-developers] queries on Kaldi

[Tony R. <to...@ca...>]

Hi Vassil,

On 06/15/2012 01:07 PM, Vassil Panayotov wrote:
> I think the VoxForge recipe is a very good idea. I am not sure what
> language model should be used for the decoding part of the recipe, but
> since it will be just a demo system with free data, it doesn't matter
> that much and probably just training a "cheating" LM on all
> transcripts is OK.

I would suggest that the AM data and the LM data should be just the same 
as used in the HTK and Sphinx recipes that Ken already supports, that 
way you can compare the performance and there is no extra work needed to 
maintain the LM.


Tony

-- 
Dr A J Robinson, Founder and Director of Cantab Research Limited.
St Johns Innovation Centre, Cowley Road, Cambridge, CB4 0WS, UK.
Company reg no 05697423 (England and Wales), VAT reg no 925606030.

Re: [Kaldi-developers] queries on Kaldi

[Vassil P. <vas...@gm...>]

Hi Robert,

I think the VoxForge recipe is a very good idea. I am not sure what
language model should be used for the decoding part of the recipe, but
since it will be just a demo system with free data, it doesn't matter
that much and probably just training a "cheating" LM on all
transcripts is OK.

I agree with Arnab, that it would be the best if you base your recipe
on wsj/s5. In fact I think it shouldn't be very difficult to modify it
to use the VoxForge data, because for example in egs/rm/s5 just the
data normalization scripts("local/") are specific to RM, and steps/
and utils/ directories are just symlinks to wsj/s5.

Vassil

On Fri, Jun 15, 2012 at 2:21 PM, Arnab Ghoshal <ar...@gm...> wrote:
> On Mon, Jun 11, 2012 at 3:09 PM, Robert Mullins <r_p...@ya...> wrote:
>> I downloaded and installed the Kaldi code on my machine, however I do not
>> have the data from the LDC RM discs. I was just wondering whether anyone has
>> used the voxforge speech corpus for training your system? And if not is
>> there a reason and would it be a worthwhile exercise?
>
> Hi Robert,
>
> you could use the egs/rm/s4 recipe that uses a free subset of RM. Here
> is a blog post about it:
> http://vpanayotov.blogspot.co.uk/2012/02/poor-mans-kaldi-recipe-setup.html
>
> There may be people on this list who have used voxforge. You are
> certainly welcome to contribute a recipe for it.
>
>> I would also be interested in seeing an example of how to use the kaldi
>> system as an end to end speech recognition system
>
> That's exactly what Kaldi is meant to be. Once you go through the
> steps of rm/s4 you will get a hang of it. If you want to start
> building a voxforge recipe, I will recommend that you start from
> egs/wsj/s5.
>
> -Arnab

Re: [Kaldi-developers] queries on Kaldi

[Arnab G. <ar...@gm...>]

On Mon, Jun 11, 2012 at 3:09 PM, Robert Mullins <r_p...@ya...> wrote:
> I downloaded and installed the Kaldi code on my machine, however I do not
> have the data from the LDC RM discs. I was just wondering whether anyone has
> used the voxforge speech corpus for training your system? And if not is
> there a reason and would it be a worthwhile exercise?

Hi Robert,

you could use the egs/rm/s4 recipe that uses a free subset of RM. Here
is a blog post about it:
http://vpanayotov.blogspot.co.uk/2012/02/poor-mans-kaldi-recipe-setup.html

There may be people on this list who have used voxforge. You are
certainly welcome to contribute a recipe for it.

> I would also be interested in seeing an example of how to use the kaldi
> system as an end to end speech recognition system

That's exactly what Kaldi is meant to be. Once you go through the
steps of rm/s4 you will get a hang of it. If you want to start
building a voxforge recipe, I will recommend that you start from
egs/wsj/s5.

-Arnab

[Kaldi-developers] queries on Kaldi

[Robert M. <r_p...@ya...>]

Hi ,

I downloaded and installed the Kaldi code on my machine, however I do not have the data from the LDC RM discs. I was just wondering whether anyone has used the voxforge speech corpus for training your system? And if not is there a reason and would it be a worthwhile exercise?

I would also be interested in seeing an example of how to use the kaldi system as an end to end speech recognition system, perhaps as an alternative to using the CMU sphinx system. Is this possible to do or is it meant to be used in this way?

Best Regards,
Robert

 
Robert Mullins
email : r_p...@ya...
phone : +353 87 2409338

Re: [Kaldi-developers] about function "PushArc" in trivial-factor-weight.h

[Daniel P. <dp...@gm...>]

Hi,

This function was in a base-class that was part of the OpenFst code,
see ../src/openfst/include/fst/cache.h
I have added a "using" statement to the code that might help it
compile.  I am surprised this was the only problem.
The problem on Windows is that the example scripts will not work
because they are written in bash.  Unless this is for a specific
purpose, it's probably a bad idea to try.

Dan

On Fri, May 18, 2012 at 4:14 AM, 司玉景 <siy...@hc...> wrote:
> Hi,
>   Recently, I tried to install kaldi on the windows. Everything is all
> right except that I can not found the declaration and definition of
> "PushArc". However, "PushArc" is directly used in the file
> "trivial-factor-weight.h". I doubt whether it is a bug or not. I am
> a newbie about kaldi. If you could give me some explanations, I would be
> very grateful.
>
> Best wishes for you!
>
>
> Yours sincerely, yujing si.
>
>
> --
> 中国科学院声学研究所语言声学与内容理解重点实验室
> Key Laboratory of Acoustics and Natural language understanding,
> Institute of Acoustics, Chinese Academy of Sciences，
> Beijing 100190, P.R.China
>
>
> ------------------------------------------------------------------------------
> 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-developers mailing list
> Kal...@li...
> https://lists.sourceforge.net/lists/listinfo/kaldi-developers
>

[Kaldi-developers] Request to be added to the Kaldi-users email list

[Charl v. H. <cvh...@gm...>]

Hi,

Would you please add me to the Kaldi-users email list? Also, is there an
archive which I can search through before posting questions?

Kind regards,
Charl van Heerden

[Kaldi-developers] about function "PushArc" in trivial-factor-weight.h

[司玉景 <siy...@hc...>]

Hi,
  Recently, I tried to install kaldi on the windows. Everything is all
right except that I can not found the declaration and definition of
"PushArc". However, "PushArc" is directly used in the file
"trivial-factor-weight.h". I doubt whether it is a bug or not. I am
a newbie about kaldi. If you could give me some explanations, I would be
very grateful.

Best wishes for you!


Yours sincerely, yujing si.


-- 
中国科学院声学研究所语言声学与内容理解重点实验室
Key Laboratory of Acoustics and Natural language understanding,
Institute of Acoustics, Chinese Academy of Sciences，
Beijing 100190, P.R.China

[Kaldi-developers] kaldi-users

[Partha P. <sa...@mi...>]

I would like to subscribe to the kaldi-users email list.

Thanks.
Partha

Re: [Kaldi-developers] Poor man's Kaldi recipe

[Vassil P. <vas...@gm...>]

> OK thanks.  It looks good.
> I do have a few minor comments:
>  Try to use int32 in place of int-- partly because we tend to use that
> style, and partly because you're reading and writing int-sized stuff
> to disk and I want it to work if we compile with 64 as the default
> int.
>  There is a function MachineIsLittleEndian() that you can use.  There
> is a macro KALDI_SWAP4 that you can use.
>  I think it's fine to put it with the main tools.
> In fact, I notice that copy-feats already has the ability to read HTK
> features, so why don't you add functionality to that (just add your
> name to the copyright header), i.e. add a new binary argument
> --read-sphinx, and put your Holder class wherever the HTK one is
> already.

Thank you - these suggestions are really helpful! I will implement the
changes(most probably over the weekend) and will write here when
ready.

Vassil

Re: [Kaldi-developers] Poor man's Kaldi recipe

[Daniel P. <dp...@gm...>]

> True. I was just wary of commiting directly in trunk, but will try to
> make sure that not too many commits will be necessary(hopefully just
> one) to get it right.

It's not a problem.  I often commit trivial stuff and/or fixes.

>> RE the binary that converts CMU's
>> feature files into a table-- do you have source for this?  Is it
>> written in the Kaldi style?
>
> You can see the source at:
> https://github.com/vdp/kaldi-rm1-mod/blob/master/cxx/sphinx/pack-sphinx-feats.cc

OK thanks.  It looks good.
I do have a few minor comments:
 Try to use int32 in place of int-- partly because we tend to use that
style, and partly because you're reading and writing int-sized stuff
to disk and I want it to work if we compile with 64 as the default
int.
 There is a function MachineIsLittleEndian() that you can use.  There
is a macro KALDI_SWAP4 that you can use.
 I think it's fine to put it with the main tools.
In fact, I notice that copy-feats already has the ability to read HTK
features, so why don't you add functionality to that (just add your
name to the copyright header), i.e. add a new binary argument
--read-sphinx, and put your Holder class wherever the HTK one is
already.


Dan

Re: [Kaldi-developers] Poor man's Kaldi recipe

[Vassil P. <vas...@gm...>]

Please, consider my comments below.

On Wed, Feb 22, 2012 at 5:19 PM, Daniel Povey <dp...@gm...> wrote:
> I think for something like this it's not necessary to create a
> branch-- you can just get it working on your local copy, and check it
> in when it's done.  Since the stuff you're doing wouldn't break any
> functionality in anything else, I don't think we need to test it too
> thoroughly before adding it.

True. I was just wary of commiting directly in trunk, but will try to
make sure that not too many commits will be necessary(hopefully just
one) to get it right.

> RE the binary that converts CMU's
> feature files into a table-- do you have source for this?  Is it
> written in the Kaldi style?

You can see the source at:
https://github.com/vdp/kaldi-rm1-mod/blob/master/cxx/sphinx/pack-sphinx-feats.cc
. This was the first C++ Kaldi tool, that I wrote and I wasn't
particularly concerned about making it comformant in style at the
time, but as far as I could say it is mostly Kaldi-style.
As you can see it is basically a very trivial Holder class, and Sphinx
feature format is very simple itself(which in this case is not a good
thing because just the number of vectors are recorded in the header
but nothing else, not even their dimensionality). If you have
recommendations I will try to unify the style more and I could
probably think about adding some kind of automatic endianess detection
instead of hardcoding it as template parameters. Also perhaps the name
can be changed to, for example, "copy-sphinx-feats".

> We could just include it in the tools,
> e.g. in featbin/

Yes, that's where I placed it during my experiments, but frankly I
wonder if it should be stored along with the core tools, because most
likely the majority of the Kaldi users won't use it.

> BTW, the Apache copyright notice should have just your name-- it's
> whoever wrote the tool.  Hopefully you don't work for a company that
> claims ownership of whatever you do.

No, that's not an issue in my case - just my name will be in the
header. Is the notice of the file above OK?

Regards,
Vassil

Re: [Kaldi-developers] Poor man's Kaldi recipe

[Daniel P. <dp...@gm...>]

I think for something like this it's not necessary to create a
branch-- you can just get it working on your local copy, and check it
in when it's done.  Since the stuff you're doing wouldn't break any
functionality in anything else, I don't think we need to test it too
thoroughly before adding it.  RE the binary that converts CMU's
feature files into a table-- do you have source for this?  Is it
written in the Kaldi style?  We could just include it in the tools,
e.g. in featbin/
BTW, the Apache copyright notice should have just your name-- it's
whoever wrote the tool.  Hopefully you don't work for a company that
claims ownership of whatever you do.
Dan

On Wed, Feb 22, 2012 at 4:59 AM, Vassil Panayotov
<vas...@gm...> wrote:
> Hello Dan,
>
> I am glad you like the idea!
> Of course, from maintanance and visibility point of view, it is always
> better for a recipe to be part of the official distribution. I am
> ready to help with the integration and maintanance of this recipe. How
> about setting a SVN branch, where these scripts will be integrated and
> merging it into trunk after it's ready?
> I agree, that the more exotic tools if needed should be kept
> separately from the core ones. The only binary tool that is essential
> here is the one that packs CMU's feature files into a Kaldi Table.
> Even this tool can be avoided, if needed, by shipping pre-built
> ark/scp.
>
> Vassil
>
> On Tue, Feb 21, 2012 at 5:44 PM, Daniel Povey <dp...@gm...> wrote:
>> Actually, I wonder whether it would make more sense to include your
>> recipe as an alternative within the examples in the sourceforge
>> repository?  E.g. egs/rm/s2 ?  We did need a recipe that didn't rely
>> on LDC data.  Perhaps we could add you as a committer on Sourceforge,
>> and you can check it in?  BTW, if you want to add any new tools e.g.
>> in the tools/ directory, I prefer if this is done by the scripts in
>> the egs/rm/s2 directory (or whatever it is called), rather than the
>> install scripts in tools/ itself, since those are supposed to be
>> things that are more universally used.
>>
>> Dan
>>
>> On Tue, Feb 21, 2012 at 10:35 AM, Daniel Povey <dp...@gm...> wrote:
>>> That's excellent!  I'll try to include a link to your blog from
>>> somewhere Kaldi-related.
>>> RE the pre-built acoustic models-- we'll have to think about this...
>>> it seems like a good idea, but it's a question of how to select them
>>> and update them, and to provide the extra stuff people would need
>>> (e.g. instructions how to use them).
>>>
>>> Dan
>>>
>>> On Mon, Feb 20, 2012 at 6:20 AM, Vassil Panayotov
>>> <vas...@gm...> wrote:
>>>> Hi everyone,
>>>>
>>>> Just wanted to let you know, that I just wrote a post on my brand new
>>>> blog about using Kaldi with the free RM1 subset available from CMU
>>>> (http://vpanayotov.blogspot.com/2012/02/poor-mans-kaldi-recipe-setup.html).
>>>> I do understand that the toolkit is targeted mostly toward the
>>>> researchers, but if other hobbyists want to try it, why not...
>>>>
>>>> I would also like to ask if it would be possible for you to make some
>>>> pre-built acoustic models available on Kaldi's SourceForge site or
>>>> elsewhere? I mean something like this:
>>>> http://www.keithv.com/software/htk/us/ and
>>>> http://www.keithv.com/software/sphinx/us/ .
>>>>
>>>> Regards,
>>>> Vassil
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Keep Your Developer Skills Current with LearnDevNow!
>>>> The most comprehensive online learning library for Microsoft developers
>>>> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
>>>> Metro Style Apps, more. Free future releases when you subscribe now!
>>>> http://p.sf.net/sfu/learndevnow-d2d
>>>> _______________________________________________
>>>> Kaldi-developers mailing list
>>>> Kal...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/kaldi-developers

Re: [Kaldi-developers] Poor man's Kaldi recipe

[Vassil P. <vas...@gm...>]

Hello Dan,

I am glad you like the idea!
Of course, from maintanance and visibility point of view, it is always
better for a recipe to be part of the official distribution. I am
ready to help with the integration and maintanance of this recipe. How
about setting a SVN branch, where these scripts will be integrated and
merging it into trunk after it's ready?
I agree, that the more exotic tools if needed should be kept
separately from the core ones. The only binary tool that is essential
here is the one that packs CMU's feature files into a Kaldi Table.
Even this tool can be avoided, if needed, by shipping pre-built
ark/scp.

Vassil

On Tue, Feb 21, 2012 at 5:44 PM, Daniel Povey <dp...@gm...> wrote:
> Actually, I wonder whether it would make more sense to include your
> recipe as an alternative within the examples in the sourceforge
> repository?  E.g. egs/rm/s2 ?  We did need a recipe that didn't rely
> on LDC data.  Perhaps we could add you as a committer on Sourceforge,
> and you can check it in?  BTW, if you want to add any new tools e.g.
> in the tools/ directory, I prefer if this is done by the scripts in
> the egs/rm/s2 directory (or whatever it is called), rather than the
> install scripts in tools/ itself, since those are supposed to be
> things that are more universally used.
>
> Dan
>
> On Tue, Feb 21, 2012 at 10:35 AM, Daniel Povey <dp...@gm...> wrote:
>> That's excellent!  I'll try to include a link to your blog from
>> somewhere Kaldi-related.
>> RE the pre-built acoustic models-- we'll have to think about this...
>> it seems like a good idea, but it's a question of how to select them
>> and update them, and to provide the extra stuff people would need
>> (e.g. instructions how to use them).
>>
>> Dan
>>
>> On Mon, Feb 20, 2012 at 6:20 AM, Vassil Panayotov
>> <vas...@gm...> wrote:
>>> Hi everyone,
>>>
>>> Just wanted to let you know, that I just wrote a post on my brand new
>>> blog about using Kaldi with the free RM1 subset available from CMU
>>> (http://vpanayotov.blogspot.com/2012/02/poor-mans-kaldi-recipe-setup.html).
>>> I do understand that the toolkit is targeted mostly toward the
>>> researchers, but if other hobbyists want to try it, why not...
>>>
>>> I would also like to ask if it would be possible for you to make some
>>> pre-built acoustic models available on Kaldi's SourceForge site or
>>> elsewhere? I mean something like this:
>>> http://www.keithv.com/software/htk/us/ and
>>> http://www.keithv.com/software/sphinx/us/ .
>>>
>>> Regards,
>>> Vassil
>>>
>>> ------------------------------------------------------------------------------
>>> Keep Your Developer Skills Current with LearnDevNow!
>>> The most comprehensive online learning library for Microsoft developers
>>> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
>>> Metro Style Apps, more. Free future releases when you subscribe now!
>>> http://p.sf.net/sfu/learndevnow-d2d
>>> _______________________________________________
>>> Kaldi-developers mailing list
>>> Kal...@li...
>>> https://lists.sourceforge.net/lists/listinfo/kaldi-developers

Re: [Kaldi-developers] Poor man's Kaldi recipe

[Daniel P. <dp...@gm...>]

Actually, I wonder whether it would make more sense to include your
recipe as an alternative within the examples in the sourceforge
repository?  E.g. egs/rm/s2 ?  We did need a recipe that didn't rely
on LDC data.  Perhaps we could add you as a committer on Sourceforge,
and you can check it in?  BTW, if you want to add any new tools e.g.
in the tools/ directory, I prefer if this is done by the scripts in
the egs/rm/s2 directory (or whatever it is called), rather than the
install scripts in tools/ itself, since those are supposed to be
things that are more universally used.

Dan

On Tue, Feb 21, 2012 at 10:35 AM, Daniel Povey <dp...@gm...> wrote:
> That's excellent!  I'll try to include a link to your blog from
> somewhere Kaldi-related.
> RE the pre-built acoustic models-- we'll have to think about this...
> it seems like a good idea, but it's a question of how to select them
> and update them, and to provide the extra stuff people would need
> (e.g. instructions how to use them).
>
> Dan
>
> On Mon, Feb 20, 2012 at 6:20 AM, Vassil Panayotov
> <vas...@gm...> wrote:
>> Hi everyone,
>>
>> Just wanted to let you know, that I just wrote a post on my brand new
>> blog about using Kaldi with the free RM1 subset available from CMU
>> (http://vpanayotov.blogspot.com/2012/02/poor-mans-kaldi-recipe-setup.html).
>> I do understand that the toolkit is targeted mostly toward the
>> researchers, but if other hobbyists want to try it, why not...
>>
>> I would also like to ask if it would be possible for you to make some
>> pre-built acoustic models available on Kaldi's SourceForge site or
>> elsewhere? I mean something like this:
>> http://www.keithv.com/software/htk/us/ and
>> http://www.keithv.com/software/sphinx/us/ .
>>
>> Regards,
>> Vassil
>>
>> ------------------------------------------------------------------------------
>> Keep Your Developer Skills Current with LearnDevNow!
>> The most comprehensive online learning library for Microsoft developers
>> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
>> Metro Style Apps, more. Free future releases when you subscribe now!
>> http://p.sf.net/sfu/learndevnow-d2d
>> _______________________________________________
>> Kaldi-developers mailing list
>> Kal...@li...
>> https://lists.sourceforge.net/lists/listinfo/kaldi-developers

Re: [Kaldi-developers] Poor man's Kaldi recipe

[Daniel P. <dp...@gm...>]

That's excellent!  I'll try to include a link to your blog from
somewhere Kaldi-related.
RE the pre-built acoustic models-- we'll have to think about this...
it seems like a good idea, but it's a question of how to select them
and update them, and to provide the extra stuff people would need
(e.g. instructions how to use them).

Dan

On Mon, Feb 20, 2012 at 6:20 AM, Vassil Panayotov
<vas...@gm...> wrote:
> Hi everyone,
>
> Just wanted to let you know, that I just wrote a post on my brand new
> blog about using Kaldi with the free RM1 subset available from CMU
> (http://vpanayotov.blogspot.com/2012/02/poor-mans-kaldi-recipe-setup.html).
> I do understand that the toolkit is targeted mostly toward the
> researchers, but if other hobbyists want to try it, why not...
>
> I would also like to ask if it would be possible for you to make some
> pre-built acoustic models available on Kaldi's SourceForge site or
> elsewhere? I mean something like this:
> http://www.keithv.com/software/htk/us/ and
> http://www.keithv.com/software/sphinx/us/ .
>
> Regards,
> Vassil
>
> ------------------------------------------------------------------------------
> Keep Your Developer Skills Current with LearnDevNow!
> The most comprehensive online learning library for Microsoft developers
> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
> Metro Style Apps, more. Free future releases when you subscribe now!
> http://p.sf.net/sfu/learndevnow-d2d
> _______________________________________________
> Kaldi-developers mailing list
> Kal...@li...
> https://lists.sourceforge.net/lists/listinfo/kaldi-developers

[Kaldi-developers] Poor man's Kaldi recipe

[Vassil P. <vas...@gm...>]

Hi everyone,

Just wanted to let you know, that I just wrote a post on my brand new
blog about using Kaldi with the free RM1 subset available from CMU
(http://vpanayotov.blogspot.com/2012/02/poor-mans-kaldi-recipe-setup.html).
I do understand that the toolkit is targeted mostly toward the
researchers, but if other hobbyists want to try it, why not...

I would also like to ask if it would be possible for you to make some
pre-built acoustic models available on Kaldi's SourceForge site or
elsewhere? I mean something like this:
http://www.keithv.com/software/htk/us/ and
http://www.keithv.com/software/sphinx/us/ .

Regards,
Vassil

[Kaldi-developers] Deleting sandbox/discrim branch

[Arnab G. <ar...@gm...>]

Hello all,

the contents of the sandbox/discrim branch have been merged back into
the trunk and the branch should no longer be used for development. We
will delete the branch in the coming days. If you have any local
changes in that branch which are not checked in yet please let me know
soon since any changes will have to be merged manually.

Thanks and happy new year,
-Arnab

Re: [Kaldi-developers] Error while compiling sclite

[Daniel P. <dp...@gm...>]

I have modified the install.sh to fix this automatically-- do "svn
update" and see if it works now.
Dan


On Sun, Dec 25, 2011 at 1:26 PM, Daniel Povey <dp...@gm...> wrote:
> Hi,
> Just tested this with g++-4.6
> Replace Filter::Filter with ::Filter
> There are also two instances in recording.cc
> [ it also works if you replace it with "typename Filter::Filter" but
> I'm not sure that this is valid C++ so I would avoid it]
> Cc'ing John Fiscus so he can fix it in the original version.
> Dan
>
> On Sat, Dec 24, 2011 at 11:35 PM, LI Peng <lip...@gm...> wrote:
>> Hi,
>>
>> The following error rises when I build sclite:
>>
>> recording.h:122:29: error: ‘Filter::Filter’ cannot appear in a
>> constant-expression.
>>
>> The g++ version is 4.6.1, and the OS is Ubuntu 11.10.
>>
>> Best regards,
>>
>> LI Peng
>>
>> ------------------------------------------------------------------------------
>> Write once. Port to many.
>> Get the SDK and tools to simplify cross-platform app development. Create
>> new or port existing apps to sell to consumers worldwide. Explore the
>> Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join
>> http://p.sf.net/sfu/intel-appdev
>> _______________________________________________
>> Kaldi-developers mailing list
>> Kal...@li...
>> https://lists.sourceforge.net/lists/listinfo/kaldi-developers
>>

Re: [Kaldi-developers] Error while compiling sclite

[Daniel P. <dp...@gm...>]

Hi,
Just tested this with g++-4.6
Replace Filter::Filter with ::Filter
There are also two instances in recording.cc
[ it also works if you replace it with "typename Filter::Filter" but
I'm not sure that this is valid C++ so I would avoid it]
Cc'ing John Fiscus so he can fix it in the original version.
Dan

On Sat, Dec 24, 2011 at 11:35 PM, LI Peng <lip...@gm...> wrote:
> Hi,
>
> The following error rises when I build sclite:
>
> recording.h:122:29: error: ‘Filter::Filter’ cannot appear in a
> constant-expression.
>
> The g++ version is 4.6.1, and the OS is Ubuntu 11.10.
>
> Best regards,
>
> LI Peng
>
> ------------------------------------------------------------------------------
> Write once. Port to many.
> Get the SDK and tools to simplify cross-platform app development. Create
> new or port existing apps to sell to consumers worldwide. Explore the
> Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join
> http://p.sf.net/sfu/intel-appdev
> _______________________________________________
> Kaldi-developers mailing list
> Kal...@li...
> https://lists.sourceforge.net/lists/listinfo/kaldi-developers
>

[Kaldi-developers] Error while compiling sclite

[LI P. <lip...@gm...>]

Hi,

The following error rises when I build sclite:

recording.h:122:29: error: ‘Filter::Filter’ cannot appear in a
constant-expression.

The g++ version is 4.6.1, and the OS is Ubuntu 11.10.

Best regards,

LI Peng