Hi everyone
I'm testing Moses Hiero on a parallel corpus. To minimize the tuning error,
I run MERT three times from flat weights. The bleu score for the first
tuning run, on the test dataset, is 15.5, while the bleu for the second and
third runs are 16.14 and 16.60.
I wonder if this significant d
Barry,
If I want the end result to be Compact Tables and not OnDisk.
Do I have to binarize first or can I convert directly to Compact ? (ie
can I skip the CreateOnDisk stuff)
if so is there a predefined script or should do it manually ?
thanks
Le 28/07/2015 15:44, Barry Haddow a écrit :
Hi V
Dear All,
all I need is to enter one word and get it is translation, is it available
in Moses and what must I do to use Moses for character-level translation
?
thank you,
Fatma El-Zahraa El -Taher
Teaching Assistant at Computer & System department
Faculty of Engineering, Azhar Universit
found it
I needed #!/usr/bin/perl -w
instead of
#!/usr/bin/env perl
but now I am confused.
I thought the second one was good.
why not working in the context of translate.cgi
Le 30/07/2015 16:48, Vincent Nguyen a écrit :
>
> still don't know why not working but I managed to patch the o
still don't know why not working but I managed to patch the original
tokenizer / detokenizer .perl files from Hervé.
I added the conversion ' => '
and reversed in his code.
not very clean but does the job.
Le 30/07/2015 14:53, Vincent Nguyen a écrit :
> nope.
>
> the tokenizer.perl in scrip
Are you passing the right filename to the "tar" command?
In your log it looks as if there's an extra word "tar" tacked onto the end of
the filename as written in that command line. That's probably just the output
getting mixed up with the command line though.
Jeroen
On July 30, 2015 8:01:29
Hi,
yes, that is correct. If there are non-zero valued scores listed with a
translation model feature, then this translation model was used for the
phrase pair.
-phi
On Wed, Jul 29, 2015 at 7:57 PM, Saumitra Yadav
wrote:
> Sir,
> Thank you for that option , it really helped. I just wanted to k
Hi,
according to the wget command, you did manage to download boost, but it may
have a different name.
Please check the directory and rename the file into boost_1_55_0.tar.gz
before unpacking it.
-phi
On Thu, Jul 30, 2015 at 8:48 AM, kalu mera wrote:
> Dear members
>
> Dear All,
> My name is
-- Forwarded message --
From: kalu mera
Date: 30 July 2015 at 15:48
Subject: request for help
To: moses-support@mit.edu
Dear members
Dear All,
My name is Akello Kalumera Christine, I am new to Moses
and moses-support mailing list. i have been trying to install moses by
following
nope.
the tokenizer.perl in scripts/share/tokenizer is much longer than the
one in contrib/web/bin.
I replaced it file for file, modified the path to nonbreaking_prefixes
tried with -b
but nothing happens in the ./daemon.pl window.
Le 30/07/2015 10:38, Barry Haddow a écrit :
> Try using the
Dear members
Dear All,
My name is Akello Kalumera Christine, I am new to Moses
and moses-support mailing list. i have been trying to install moses by
following the command from the moses website,
however, after installing boost i get this error.
kalumera@kalumera-Satellite-C50-A534:~$ tar zxvf bo
Hello Fatma,
Moses has been used for character-level translation, but the alignment
is not done by moses itself, but by external tools like (M)GIZA++ or
fast-align.
Jörg Tiedemann has some suggestions on how to represent your text to get
good character alignments with GIZA++:
http://www.acl
Try using the -b option in the tokenizer / detokenizer to disable buffering.
On 29/07/15 18:47, Vincent Nguyen wrote:
> Hi,
>
> As is, it was working fine except the tokenizer / detokenizer .perl code
> is outdated.
> It causes problem with the apostrophe in French.
>
> so I changed the translate.
13 matches
Mail list logo