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. Have found the problem Fw:A new problem when tuning in Errors
when training: No extract.o.sorted.gz (???)
----------------------------------------------------------------------
Message: 1
Date: Thu, 8 Jun 2017 00:23:04 +0800 (CST)
From: ??? <yufeifan7@126.com>
Subject: [Moses-support] Have found the problem Fw:A new problem when
tuning in Errors when training: No extract.o.sorted.gz
To: moses-support@mit.edu
Message-ID: <1e1dfa26.ccd4.15c835dd667.Coremail.yufeifan7@126.com>
Content-Type: text/plain; charset="gbk"
I checked the filterphrases.err and found that I moved the related files without changing the path in moses.ini. Thanks for paying attention.
-------- Forwarding messages --------
From: "???" <yufeifan7@126.com>
Date: 2017-06-07 21:26:38
To: moses-support@mit.edu
Subject: A new problem when tuning in Errors when training: No extract.o.sorted.gz
Dear all:
I tried running from the first step and successfully completed training. But when tuning, an error comes out immediately:
The command is:
mert-moses.pl /home/yuff/mosesdecoder/projectdata/accessor_variety/dev-split-trainingmodel.clean.right /home/yuff/mosesdecoder/projectdata/accessor_variety/dev-split-trainingmodel.clean.left /home/yuff/mosesdecoder/bin/moses \
train/model/moses.ini --mertdir /home/yuff/mosesdecoder/bin/ --mertargs="--sctype TER" --decoder-flag="-threads-8"
Thanks for paying attention and trying to give suggestions.
Dear all:
I am a new hand at moses and machine translation. I was trying to completing a task of translation between two unknown languages.I had completed the word segmentation and clean the corpus with the command: clean-corpus-n.perl train-split.txt right left train-split.clean 1 120
(PS:the two pararrel corpus with extension left and right)
Then I ran the train-model perl as follows:
train-model.perl -root-dir train --cores 8 --corpus /home/yuff/mosesdecoder/projectdata/accessor_variety/train-split.clean --f right --e left --alignment grow-diag-final-and --first-step 2 --reordering msd-bidirectional-fe --lm 0:3:/home/yuff/mosesdecoder/projectdata/accessor_variety/train-split.blm.left --mgiza -mgiza-cpus 2 --parallel --external-bin-dir /home/yuff/mosesdecoder/word_align_tools/
There were a lot of warnings beginning with "Use of Uninitialized Value in " coming out, without ending the executing of perl. However, after about 20 minutes, I got the error as follows:
"Executing: /home/yuff/mosesdecoder/scripts/../bin/lexical-reordering-score /home/yuff/mosesdecoder/projectdata/accessor_variety/translating_working/train/model/extract.o.sorted.gz 0.5 /home/yuff/mosesdecoder/projectdata/accessor_variety/translating_working/train/model/reordering-table. --model "wbe msd wbe-msd-bidirectional-fe"
Lexical Reordering Scorer
scores lexical reordering models of several types (hierarchical, phrase-based and word-based-extraction
terminate called after throwing an instance of 'util::ErrnoException'
what(): util/file.cc:76 in int util::OpenReadOrThrow(const char*) threw ErrnoException because `-1 == (ret = open(name, 00))'.
No such file or directory while opening /home/yuff/mosesdecoder/projectdata/accessor_variety/translating_working/train/model/extract.o.sorted.gz
Aborted (core dumped)
Exit code: 134
ERROR: Lexical reordering scoring failed at /home/yuff/mosesdecoder/scripts/training/train-model.perl line 1924."
The files under the directory "/train" are extremely small or even empty. I searched for the problem and found some others asking about that in 2012: He got the reply like this
However, '[',']' are both characters in these two languages, and uppercase and downcase letters are recognized as different ones. I'm not sure if the first step of training would downcase the letters, so I started with the second step.
I will appreciate it so much if some ones can give some suggestions. Thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20170607/0a771cf0/attachment.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 45643 bytes
Desc: not available
Url : http://mailman.mit.edu/mailman/private/moses-support/attachments/20170607/0a771cf0/attachment.png
------------------------------
_______________________________________________
Moses-support mailing list
Moses-support@mit.edu
http://mailman.mit.edu/mailman/listinfo/moses-support
End of Moses-support Digest, Vol 128, Issue 12
**********************************************
Subscribe to:
Post Comments (Atom)
Hi.
ReplyDeleteI was working on building an SMT from Amharic(Official Language of Ethiopia) to English and I faced the same problem.(No such file or directory extract.o.sorted.gz with ERROR: Lexical reordering scoring failed at line 1924 on the train-mode.perl file)
I resolved the issue by following tutorial on this site http://bit.ly/deploy-smt
My problem was that I was executing mossesdecoder's train-model.perl file not the one was that was mentioned in the above tutorial link. So when you reach the step to train your translation system use the train-model.perl file located in the bin folder of the tool mentioned in the above site.
In short try to train the translation system from the zip file found at http://bit.ly/smt-moses-tool by navigating to scripts/training/train-model.perl.
Hope it helps and good luck!