Home » Uncategories » Moses-support Digest, Vol 83, Issue 7

Moses-support Digest, Vol 83, Issue 7

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. Using two translation tables for decoding (rohit dholakia)
2. Re: Using two translation tables for decoding (Philipp Koehn)
3. Re: size of training.out (Hieu Hoang)


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

Message: 1
Date: Wed, 4 Sep 2013 18:14:04 -0700
From: rohit dholakia <rdholaki@sfu.ca>
Subject: [Moses-support] Using two translation tables for decoding
To: moses-support@mit.edu
Message-ID:
<CAA==Lgs9b0oz7LOpiuQZn20+O3R7ZjygfYxSBj-5RO1DMbOPVA@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

Hi,

I am trying to use one of the features in moses that allows usage of two
translation tables for decoding. The directions are here :

http://www.statmt.org/moses/?n=Moses.AdvancedFeatures#ntoc18

Made the following changes to the moses.ini file :

# mapping steps
[mapping]
0 T 0
1 T 1

0 0 0 5 /global/scratch/rd/Thesis/UsingBothTables/baseline-table.gz
0 0 0 5 /global/scratch/rd/Thesis/UsingBothTables/augmented.gz

# translation model weights
[weight-t]
0.0440237
0.0427656
0.0508153
0.0441958
0.0043927
0.058598
0.0660405
0.108836
0.0028447
0.0856689

The distortion weights are untouched, am using the same language model, so,
left that alone as well. only made the changes mentioned above ( two
tables, mapping change and ten translation weights ).

When I run,

*$MOSES_DIR/moses -f moses.ini < /global/scratch/rd/Thesis/data/devtest/
devtest-clean.lower.ht > both.output.devtest.cl 2> err.both*

I get the following output, ( only the last part shown ),

*Loading table into memory...done.*
*Start loading text SCFG phrase table. Moses format : [14.000] seconds*
*Reading /global/scratch/rd/Thesis/UsingBothTables/baseline-table.gz*
*
----5---10---15---20---25---30---35---40---45---50---55---60---65---70---75---80---85---90---95--100
*
*
****************************************************************************************************
*
*Start loading text SCFG phrase table. Moses format : [41.000] seconds*
*Reading /global/scratch/rd/Thesis/UsingBothTables/augmented.gz*
*
----5---10---15---20---25---30---35---40---45---50---55---60---65---70---75---80---85---90---95--100
*
*
****************************************************************************************************
*
*Exception: std::bad_alloc*

Where am I going wrong ?

Thanks.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20130904/240faf6a/attachment-0001.htm

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

Message: 2
Date: Thu, 5 Sep 2013 11:20:21 +0200
From: Philipp Koehn <pkoehn@inf.ed.ac.uk>
Subject: Re: [Moses-support] Using two translation tables for decoding
To: rohit dholakia <rdholaki@sfu.ca>
Cc: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAAFADDAXs7AGinMAfdQyQpkTp40QNoNXMqse3YK5d7eO_vR-eA@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

Hi,

a "bad_alloc" is typically an indication that you ran out of RAM.
I would suggest to look into the binary and compact phrase table
implementations.

-phi


On Thu, Sep 5, 2013 at 3:14 AM, rohit dholakia <rdholaki@sfu.ca> wrote:

> Hi,
>
> I am trying to use one of the features in moses that allows usage of two
> translation tables for decoding. The directions are here :
>
> http://www.statmt.org/moses/?n=Moses.AdvancedFeatures#ntoc18
>
> Made the following changes to the moses.ini file :
>
> # mapping steps
> [mapping]
> 0 T 0
> 1 T 1
>
> 0 0 0 5 /global/scratch/rd/Thesis/UsingBothTables/baseline-table.gz
> 0 0 0 5 /global/scratch/rd/Thesis/UsingBothTables/augmented.gz
>
> # translation model weights
> [weight-t]
> 0.0440237
> 0.0427656
> 0.0508153
> 0.0441958
> 0.0043927
> 0.058598
> 0.0660405
> 0.108836
> 0.0028447
> 0.0856689
>
> The distortion weights are untouched, am using the same language model,
> so, left that alone as well. only made the changes mentioned above ( two
> tables, mapping change and ten translation weights ).
>
> When I run,
>
> *$MOSES_DIR/moses -f moses.ini < /global/scratch/rd/Thesis/data/devtest/
> devtest-clean.lower.ht > both.output.devtest.cl 2> err.both*
>
> I get the following output, ( only the last part shown ),
>
> *Loading table into memory...done.*
> *Start loading text SCFG phrase table. Moses format : [14.000] seconds*
> *Reading /global/scratch/rd/Thesis/UsingBothTables/baseline-table.gz*
> *
> ----5---10---15---20---25---30---35---40---45---50---55---60---65---70---75---80---85---90---95--100
> *
> *
> ****************************************************************************************************
> *
> *Start loading text SCFG phrase table. Moses format : [41.000] seconds*
> *Reading /global/scratch/rd/Thesis/UsingBothTables/augmented.gz*
> *
> ----5---10---15---20---25---30---35---40---45---50---55---60---65---70---75---80---85---90---95--100
> *
> *
> ****************************************************************************************************
> *
> *Exception: std::bad_alloc*
>
> Where am I going wrong ?
>
> Thanks.
>
> _______________________________________________
> 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/20130905/5028bfbd/attachment-0001.htm

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

Message: 3
Date: Thu, 5 Sep 2013 11:33:14 +0200
From: Hieu Hoang <Hieu.Hoang@ed.ac.uk>
Subject: Re: [Moses-support] size of training.out
To: Arefeh Kazemi <arefeh_kazemi@yahoo.com>
Cc: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAEKMkbh7XLV6c57FRufCE8YxSszbgwL6QUnsAOAenZCXFo9UVQ@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

the command looks ok. I'm not sure if 500k for training.out is normal or
not.

Is there any error messages in training.out? Are the models created by the
command you ran ok?




On 4 September 2013 11:52, Arefeh Kazemi <arefeh_kazemi@yahoo.com> wrote:

> Dear Hieu
> Thank you for the reply.
> the exact command is:
>
> nohup nice
>
>
> /opt/tools/mosesdecoder-RELEASE-1.0/mosesdecoder-RELEASE-1.0/scripts/training/train-model.perl
>
> -root-dir train -corpus /opt/tools/dataset/mizan/M_Tr_Clean -f en -e fa
>
> -alignment grow-diag-final-and -reordering msd-bidirectional-fe -lm
>
> 0:3:/opt/tools/lm2/Mizan.blm.en:8 -external-bin-dir
>
> /opt/tools/mosesdecoder-RELEASE-1.0/mosesdecoder-RELEASE-1.0/tools/ >
> training.out
>
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>
>


--
Hieu Hoang
Research Associate
University of Edinburgh
http://www.hoang.co.uk/hieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20130905/bd47c994/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 83, Issue 7
********************************************

0 Response to "Moses-support Digest, Vol 83, Issue 7"

Post a Comment

Subscribe to: Post Comments (Atom)