Moses-support Digest, Vol 123, Issue 6

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. Some Wrong answers about the train-transliteration-module.pl
? ( WULongski )


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

Message: 1
Date: Tue, 10 Jan 2017 15:16:33 +0800
From: " WULongski " <1251817914@qq.com>
Subject: [Moses-support] Some Wrong answers about the
train-transliteration-module.pl ?
To: " moses-support " <moses-support@mit.edu>, " hieuhoang "
<hieuhoang@gmail.com>
Message-ID: <tencent_74A258044BC62AF9203DE7F2@qq.com>
Content-Type: text/plain; charset="gb18030"

hi , professor
Now I want to translate the German to English, I first run the baseline system .Then i want to use the transliteration-module to handle the UNK:
I do as the website(http://www.statmt.org/moses/?n=Advanced.OOVs) says:
first I run this scripts:
as this (https://docs.google.com/document/d/1G9RjczZXWGHU6byJFORf6uToItph1jU_piL53wQhGXg/edit)
"
/nobackup/f1/asr/wulong/MOSESHEJUNQIN/Moses/mosesdecoder/scripts/Transliteration/train-transliteration-module.pl --corpus-f /nobackup/f1/asr/wulong/MOSESHEJUNQIN/Moses/mosesdecoder/HEJUNQIN/corpus/train.clean.de --corpus-e /nobackup/f1/asr/wulong/MOSESHEJUNQIN/Moses/mosesdecoder/HEJUNQIN/corpus/train.clean.en --alignment /nobackup/f1/asr/wulong/MOSESHEJUNQIN/Moses/mosesdecoder/HEJUNQIN/working/train/model/aligned.grow-diag-final-and --moses-src-dir /nobackup/f1/asr/wulong/MOSESHEJUNQIN/Moses/mosesdecoder/ --external-bin-dir /nobackup/f1/asr/wulong/MOSESHEJUNQIN/Moses/mosesdecoder/mgizacombinedtools/ --input-extension de --output-extension en --srilm-dir /nobackup/f1/asr/wulong/srilm-1.7.2/bin/i686-m64/ --out-dir translit-en-hi > translit.out 2>&1
"

I use all Absolute paths. But there erros in the mail attachments . It says " mosesdecoder/mgizacombinedtools/mkcls -c50 -n2 -p/working/translit-en-hi/training/corpus.de"Then I check the file and find that ./working/translit-en-hi/training/corpus.de is empty .then I check "/working/translit-en-hi/"and I find that "1-1.de-en.cleaned? 1-1.de-en.mined-pairs? 1-1.de-en.pair-probs" are also empty which lead to corpus.de to be empty .
Then I check the train-transliteration-module.pl and find the reason why "/working/translit-en-hi/"and I find that "1-1.de-en.cleaned? 1-1.de-en.mined-pairs? 1-1.de-en.pair-probs" are empty is
the line 268 " `$MOSES_SRC_DIR/scripts/Transliteration/clean.pl $OUT_DIR/1-1.$inp_ext-$op_ext > $OUT_DIR/1-1.$inp_ext-$op_ext.cleaned`;"
then I check $MOSES_SRC_DIR/scripts/Transliteration/clean.pl and find that all the print...... statements have become #print ..... so it cannot output to $OUT_DIR/1-1.$inp_ext-$op_ext.cleaned.
so I think if there is a bug in the clean.pl .

Thank you very much !
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20170110/8bb1c962/attachment.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: clean.pl
Type: application/octet-stream
Size: 6537 bytes
Desc: not available
Url : http://mailman.mit.edu/mailman/private/moses-support/attachments/20170110/8bb1c962/attachment.obj
-------------- next part --------------
A non-text attachment was scrubbed...
Name: translit.out
Type: application/octet-stream
Size: 29351 bytes
Desc: not available
Url : http://mailman.mit.edu/mailman/private/moses-support/attachments/20170110/8bb1c962/attachment-0001.obj

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

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


End of Moses-support Digest, Vol 123, Issue 6
*********************************************

Tweet