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. Can moses be used to do system combination
(xinzyang@mail.ustc.edu.cn)
2. Re: Factored models and xml-input (Standa K)
3. Re: Factored models and xml-input (Hieu Hoang)
4. Re: Factored models and xml-input (Standa K)
5. Re: Factored models and xml-input (Hieu Hoang)
6. Re: Factored models and xml-input (Marcin Junczys-Dowmunt)
----------------------------------------------------------------------
Message: 1
Date: Thu, 21 May 2015 10:19:50 +0800 (GMT+08:00)
From: xinzyang@mail.ustc.edu.cn
Subject: [Moses-support] Can moses be used to do system combination
To: moses-support <moses-support@mit.edu>
Message-ID:
<102c023b.52ef.14d7445076c.Coremail.xinzyang@mail.ustc.edu.cn>
Content-Type: text/plain; charset="utf-8"
Hi,
I am studying moses to do some research abou SMT. I have read confusion networks decoding of moses at the website http://www.statmt.org/moses/?n=Moses.ConfusionNetworks . As we know that confusion network can do system combination, so could moses be used to do system combination?
Thanks and Best Regards,
Zhenxin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150521/71310b12/attachment-0001.htm
------------------------------
Message: 2
Date: Thu, 21 May 2015 04:51:09 +0000 (UTC)
From: Standa K <standa.kurik@gmail.com>
Subject: Re: [Moses-support] Factored models and xml-input
To: moses-support@mit.edu
Message-ID: <loom.20150521T064203-787@post.gmane.org>
Content-Type: text/plain; charset=us-ascii
Hello Hieu,
the output I get is:
...
Loading LM0
Loading LM1
Loading LM2
Loading LM3
Loading LM4
Loading TranslationModel0
Created input-output object : [0.510] seconds
Segmentation fault (core dumped)
This is the input:
<glossary translation="jabadaba">confirm|confirm|V|V</glossary>
dialog|dialog|N|N
This is the command:
cat input | moses
-f moses.ini
-xml-input exclusive
-search-algorithm 1
-print-alignment-info
Moses.ini (paths shortened):
# MERT optimized configuration
# decoder /home/moses/mt//mosesdecoder/bin//moses
# BLEU 0.484687 on dev /storage/moses/trained-models/EN_CS_2015-04-
09T16_51_55_915978/data/dev.EN
# We were before running iteration 7
# finished Thu Apr 9 18:04:20 CEST 2015
### MOSES CONFIG FILE ###
#########################
# input factors
[input-factors]
0
1
2
# mapping steps
[mapping]
0 T 0
[distortion-limit]
6
# feature functions
[feature]
UnknownWordPenalty
WordPenalty
PhrasePenalty
PhraseDictionaryCompact name=TranslationModel0 num-features=4
path=phrase-table.compact.minphr
input-factor=0,2 output-factor=0,2
Distortion
KENLM lazyken=0 name=LM0 factor=0
path=factored.CS.reduced.0.arpa.binary order=5
KENLM lazyken=0 name=LM1 factor=2
path=factored.CS.reduced.2.arpa.binary order=4
KENLM lazyken=0 name=LM2 factor=0
path=extralm.0.0.binary order=5
KENLM lazyken=0 name=LM3 factor=2
path=extralm.0.2.binary order=4
KENLM lazyken=0 name=LM4 factor=0
path=lm.binary order=5
# dense weights for feature functions
[v]
0
[th]
7
[search-algorithm]
1
[weight]
Distortion0= 0.0898804
LM0= 0.101411
LM1= 0.0023427
LM2= 0.00440753
LM3= 0.027301
LM4= 0.0663638
WordPenalty0= -0.296714
PhrasePenalty0= -0.0296737
TranslationModel0= 0.0277221 0.10007 0.252912 -0.00120226
UnknownWordPenalty0= 1
------------------------------
Message: 3
Date: Thu, 21 May 2015 09:58:34 +0400
From: Hieu Hoang <hieuhoang@gmail.com>
Subject: Re: [Moses-support] Factored models and xml-input
To: moses-support@mit.edu, Stanislav Ku??k <standa.kurik@gmail.com>
Message-ID: <555D740A.3070406@gmail.com>
Content-Type: text/plain; charset="utf-8"
try
<glossary translation="jabadaba|*WHATEVER|WHATEVER*">confirm|confirm|V|V</glossary> dialog|dialog|N|N
On 21/05/2015 08:51, Standa K wrote:
> -xml-input exclusive
--
Hieu Hoang
Researcher
New York University, Abu Dhabi
http://www.hoang.co.uk/hieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150521/21172b8c/attachment-0001.htm
------------------------------
Message: 4
Date: Thu, 21 May 2015 06:05:16 +0000 (UTC)
From: Standa K <standa.kurik@gmail.com>
Subject: Re: [Moses-support] Factored models and xml-input
To: moses-support@mit.edu
Message-ID: <loom.20150521T080449-787@post.gmane.org>
Content-Type: text/plain; charset=us-ascii
Yes, I tried that as well, it gives the same error.
------------------------------
Message: 5
Date: Thu, 21 May 2015 10:18:45 +0400
From: Hieu Hoang <hieuhoang@gmail.com>
Subject: Re: [Moses-support] Factored models and xml-input
To: moses-support@mit.edu
Message-ID: <555D78C5.8090904@gmail.com>
Content-Type: text/plain; charset="windows-1252"
it works for me. My input and ini files are attached
On 21/05/2015 10:05, Standa K wrote:
> Yes, I tried that as well, it gives the same error.
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>
--
Hieu Hoang
Researcher
New York University, Abu Dhabi
http://www.hoang.co.uk/hieu
-------------- next part --------------
#########################
### MOSES CONFIG FILE ###
#########################
# input factors
[input-factors]
0
1
# indicates that factors are separated by '++' in the input
#[factor-delimiter]
#++
# mapping steps
[mapping]
T 0
G 0
T 1
[verbose]
2
[feature]
KENLM factor=0 order=3 num-features=1 lazyken=0 path=/Users/Hieu/workspace/github/moses-regression-tests/lm/europarl.en.srilm.gz
Generation input-factor=0 output-factor=1 num-features=2 path=/Users/Hieu/workspace/github/moses-regression-tests/models/multi-factor/generation.0-1.gz
Distortion
WordPenalty
PhraseDictionaryMemory input-factor=0 output-factor=0 path=/Users/Hieu/workspace/github/moses-regression-tests/models/multi-factor/phrase-table.0-0.gz num-features=5 table-limit=20
PhraseDictionaryBinary input-factor=1 output-factor=1 path=/Users/Hieu/workspace/github/moses-regression-tests/models/multi-factor/phrase-table.1-1.gz num-features=5 table-limit=0
UnknownWordPenalty
[weight]
WordPenalty0= -1
KENLM0= 0.5
Generation0= 0.3 0
Distortion0= 0.6
PhraseDictionaryMemory0= 0.2 0.2 0.2 0.2 0.2
PhraseDictionaryBinary0= 0.2 0.2 0.2 0.2 0.2
-------------- next part --------------
es|PPER <glossary translation="jabadaba|dsfsdf">gibt|VVFIN</glossary> verschiedene|ADJ andere|ADJ meinungen|NN .|PE
------------------------------
Message: 6
Date: Thu, 21 May 2015 08:22:19 +0200
From: Marcin Junczys-Dowmunt <junczys@amu.edu.pl>
Subject: Re: [Moses-support] Factored models and xml-input
To: moses-support@mit.edu
Message-ID: <555D799B.3000300@amu.edu.pl>
Content-Type: text/plain; charset=windows-1252; format=flowed
Confirmed. I have the same problem and was planning to look at it. It
also happens with generation models and lms that depend on that
generated factors. XML input causes segfaults.
W dniu 21.05.2015 o 08:05, Standa K pisze:
> Yes, I tried that as well, it gives the same error.
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
------------------------------
_______________________________________________
Moses-support mailing list
Moses-support@mit.edu
http://mailman.mit.edu/mailman/listinfo/moses-support
End of Moses-support Digest, Vol 103, Issue 54
**********************************************
Subscribe to:
Post Comments (Atom)
0 Response to "Moses-support Digest, Vol 103, Issue 54"
Post a Comment