Moses-support Digest, Vol 101, Issue 48

Send Moses-support mailing list submissions to
moses-support@mit.edu

To subscribe or unsubscribe via the World Wide Web, visit
http://mailman.mit.edu/mailman/listinfo/moses-support
or, via email, send a message with subject or body 'help' to
moses-support-request@mit.edu

You can reach the person managing the list at
moses-support-owner@mit.edu

When replying, please edit your Subject line so it is more specific
than "Re: Contents of Moses-support digest..."


Today's Topics:

1. Re: Questions about kbmira code in Moses (Liangyou Li)
2. Re: lattice decoding (Hieu Hoang)
3. Re: Error compiling Moses with cmph 2.0 (Muhammad Danial Raza)


----------------------------------------------------------------------

Message: 1
Date: Fri, 13 Mar 2015 17:59:02 +0000
From: Liangyou Li <llysuda@gmail.com>
Subject: Re: [Moses-support] Questions about kbmira code in Moses
To: Barry Haddow <bhaddow@staffmail.ed.ac.uk>
Cc: moses-support@mit.edu
Message-ID:
<CAHeunNVwv=gGu83OFpv2rbFPN0TV7FKHVku7oPBzvWKLLwj9Tw@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Thanks Barry.

In my experiment, it doesn't affect kbmira.
Seemingly ToSparse is only called by hypergraph mira.

Cheers
Liangyou


On Fri, Mar 13, 2015 at 4:09 PM, Barry Haddow <bhaddow@staffmail.ed.ac.uk>
wrote:

> Hi Liangyou
>
> I looked into this a bit, and I think that ToSparse() should indeed be
> fixed along the lines you suggest. This would mean that hypergraph mira is
> currently broken for sparse features, although I don't think kbmira is
> affected.
>
> I'm quite busy at the moment, but I'll do some more testing and check in a
> fix as soon as I can - in the next week or two,
>
> cheers - Barry
>
> On 06/03/15 15:43, Liangyou Li wrote:
>
>> I have a question about the function : MiraWeightVector.ToSparse(SparseVector*
>> sparse) .
>>
>> It's defined as this:
>>
>> void MiraWeightVector::ToSparse(SparseVector* sparse) const
>> {
>> for (size_t i = 0; i < m_weights.size(); ++i) {
>> if(abs(m_weights[i])>1e-8) {
>> sparse->set(i,m_weights[i]);
>> }
>> }
>> }
>>
>> This means the same index i is used for both SparseVector and
>> MiraWeightVector.
>> However, according to the code in kbmira.cpp, there's a gap between the
>> indexes.
>>
>> for dense feature: size_t id = SparseVector::encode(names[i]);
>> then, for sparse feature: size_t id = SparseVector::encode(name) +
>> initDenseSize;
>>
>>
>> ?Why isn't the ToSparse function looks like this:
>>
>> void MiraWeightVector::ToSparse(SparseVector* sparse, size_t denseSize)
>> const
>> {
>> for (size_t i = 0; i < m_weights.size(); ++i) {
>> if (i < denseSize) {
>> sparse->set(i,m_weights[i]);
>> } else {
>> if(abs(m_weights[i])>1e-8) {
>> sparse->set(i-denseSize,m_weights[i]);
>> }
>> }
>> }
>> }
>>
>> I'm kind of confused, did I miss something ?
>> Many thanks.
>>
>> Liangyou
>>
>> --
>>
>> Liangyou Li
>> CNGL
>> School of Computing
>> Dublin City University
>>
>>
>>
>> _______________________________________________
>> Moses-support mailing list
>> Moses-support@mit.edu
>> http://mailman.mit.edu/mailman/listinfo/moses-support
>>
>
>
> --
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>
>


--

Liangyou Li
CNGL
School of Computing
Dublin City University
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150313/521b3aad/attachment-0001.htm

------------------------------

Message: 2
Date: Sat, 14 Mar 2015 13:33:53 +0000
From: Hieu Hoang <hieuhoang@gmail.com>
Subject: Re: [Moses-support] lattice decoding
To: Matt Post <post@cs.jhu.edu>
Cc: moses-support <moses-support@mit.edu>
Message-ID:
<CAEKMkbh2nz8mA_YVqtA0EEBiHy1_F+yUqBinqn5Rer_f1dZzZQ@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

There's only the web page at the moment. What kind of detail are you
looking for?
On 12 Mar 2015 21:37, "Matt Post" <post@cs.jhu.edu> wrote:

> Hi,
>
> Is there a technical writeup of Moses' phrase-based lattice decoding? The
> only real description I could find is Chris Dyer's "Generalizing Word
> Lattice Translation" paper, and that is quite high level.
>
> Matt
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150314/6ac2577e/attachment-0001.htm

------------------------------

Message: 3
Date: Sat, 14 Mar 2015 19:15:26 +0500
From: Muhammad Danial Raza <muhammad.danial.raza@gmail.com>
Subject: Re: [Moses-support] Error compiling Moses with cmph 2.0
To: Hieu Hoang <hieuhoang@gmail.com>
Cc: moses-support <moses-support@mit.edu>
Message-ID:
<CAEQzcOd2N4UB1ojsZ1grtzUWe7TfRaRTNMBEKskdG3B8f1_dwg@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Thank you, your trick worked
On Mar 13, 2015 6:25 PM, "Hieu Hoang" <hieuhoang@gmail.com> wrote:

> Try the full path, don't use ~
> On 13 Mar 2015 11:25, "Muhammad Danial Raza" <
> muhammad.danial.raza@gmail.com> wrote:
>
>> I have downloaded cmph-2.0 from sourceforge and have run the run the
>> following commands according to its documentation:
>>
>> - ./configure --prefix=$PWD
>> - make
>> - make install
>>
>> Now when I compile moses with the following command:
>>
>> -
>> - ./bjam --with-boost=~/mosesdecoder-master/boost_1_55_0
>> --with-cmph=~/mosesdecoder-master/cmph-2.0 -j8
>>
>> I receive the error "Missing File or Directory: cmph.h". Also I have seen
>> that "cmph.h" is inside the include folder in cmph directory. What am I
>> doing wrong here?
>>
>> P.S: I am using cygwin to build. The full output file is attached.
>>
>> _______________________________________________
>> Moses-support mailing list
>> Moses-support@mit.edu
>> http://mailman.mit.edu/mailman/listinfo/moses-support
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150314/90956d81/attachment-0001.htm

------------------------------

_______________________________________________
Moses-support mailing list
Moses-support@mit.edu
http://mailman.mit.edu/mailman/listinfo/moses-support


End of Moses-support Digest, Vol 101, Issue 48
**********************************************

Tweet