Code should solve the current problem... and send a warning to the next one.
*grin*
~amittai
http://www.boxresear.ch
liling tan on February 29, 2016 at 11:45am wrote:
Thanks Ammittai!!
Lolzzz I like the chained perl -pe =)
Regards,Liling
amittai axelrod on
Last I checked (18+ months ago), EMS didn't handle the 2TM-2LM setup natively. I had to modify the moses.ini file manually, and then launch the tuning step manually as well. It's straightforward, but annoying, so I wrote a bash wrapper to do it. I just uploaded it to github for you:
https://github
> The reason for using Witten-Bell was because
> Kneser-Ney wasn’t able to cope up with the counts being generated for
> coarse language models.
that is indeed an annoyance with kndiscount. however, you can now try
using "--discount_fallback" with kenlm. it works for me, even with tens
of cla
whoops, forgot link. see "class-based models" section in:
http://kheafield.com/code/kenlm/estimation/
~amittai
On 1/23/16 13:08, amittai axelrod wrote:
> > The reason for using Witten-Bell was because
> > Kneser-Ney wasn’t able to cope up with the counts being gen
Do you mean "sense spotting" as in the 2013 Carpuat et al paper:
http://www.aclweb.org/anthology/P/P13/P13-1141.pdf ?
There is already an implementation here that you can use:
www.github.com/hal3/IntrinsicPSDEvaluation/
~amittai
On 6/27/15 21:25, Marwa Refaie wrote:
>
>
> Hi all ,
> I try to im
> James
>
>
> From: amittai axelrod
> Sent: Friday, June 19, 2015 7:52 PM
> To: Read, James C; Lane Schwartz
> Cc: moses-support@mit.edu; Philipp Koehn
> Subject: Re: [Moses-support] Major bug found in Moses
>
> if we don't
if we don't understand the problem, how can we possibly fix it?
all the relevant code is open source. go for it!
~amittai
On 6/19/15 12:49, Read, James C wrote:
> So, all I did was filter out the less likely phrase pairs and the BLEU
> score shot up. Was that such a stroke of genius? Was that not
ist rather than acknowledge point 2" ?
heh -- i was right the first time:
On 6/17/15 13:20, amittai axelrod wrote:
> also, your argument could be easily mis-interpreted as "this behavior
is unexpected to me, ergo this is unexpected behavior", and that will
unfortunately bi
speaking of cobbling together a good translation from imperfect parts:
google:
A motorist heard on the radio the announcement: "Caution Caution On the
N9 you will encounter a ghost driver Please drive far right and do not
overtake!.!"
The driver: "What do you mean a dozens dozens?!""
microsoft
y would that rank
> on your desirable qualities of a TM list?
>
> James
>
>
> From: amittai axelrod
> Sent: Wednesday, June 17, 2015 8:20 PM
> To: Read, James C; Hieu Hoang; Kenneth Heafield; moses-support@mit.edu
> Cc: Arnold, Doug
>
hi --
you might not be aware, but your emails sound almost belligerently
confrontational. i can see how you would be frustrated, but starting a
conversation with "i have found a major bug" and then repeatedly saying
that "clearly" everything is broken -- that may not be the best way to
convinc
On 10 April 2012 08:07, Barry Haddow wrote:
> If we move to github, with the primary documentation written in Latex, then it
> seems to make it harder to contribute. Not everyone knows Latex, it's harder
> to link across documents with Latex, and you have to wait at least until you
> check it in b
Hi --
> In MakefileWIN32 I wrote:
> TARGETDIR=/home/administrator/bin/moses-script
that should be "...bin/moses-scripts" with an "s".
I don't know that that will fix the issue, though. Are you running the
exact Moses setup mentioned in the guide? (i.e. last stable release from
2010-08-13 ?).
P
No clue; never tried.
I'm guessing this is more a generic Cygwin/Java mismatch than a Moses-specific
issue, so try looking for that error message on your favorite search engine.
~amittai
From: Rasha R [mailto:rasha.ka...@gmail.com]
Sent: Monday, September 12, 2011 1:01 AM
To: Amittai Axelr
If you have Moses installed & working correctly under Cygwin, then just run it
from Cygwin.
If it turns out that you do not have Moses installed & working correctly under
Cygwin, or have not installed all the necessary packages in Cygwin, then check
out the handy-dandy guide:
http://ssli.ee.was
Even for the last stable release of Moses? I thought that predates the kenlm
release by a couple months.
~amittai
-Original Message-
From: moses-support-boun...@mit.edu [mailto:moses-support-boun...@mit.edu] On
Behalf Of Kenneth Heafield
Sent: Saturday, July 16, 2011 9:55 PM
To: moses-s
t looked into it, and am
unlikely to do so in the near future.
So to clarify: If anyone has experience regarding getting publicly-available LM
generation toolkits installed on Cygwin, let me know.
Cheers,
~amittai
-Original Message-
From: Amittai Axelrod
Sent: Saturday, July 16, 2
Windows, ie. compiles with Visual Studio.
Every C++ part of Moses already does but it bugs me not to have the decoder run
natively on windows
On 16/07/2011 05:32, Amittai Axelrod wrote:
Hullo Moses-folk:
I've written what I think is the first step-by-step guide to installing and
compilin
Hullo Moses-folk:
I've written what I think is the first step-by-step guide to installing and
compiling Moses on Windows (7), using Cygwin.
Hopefully this will make it easier for people to use the most common
open-source MT system on the most common OS.
Plus, it addresses this FAQ: http://www.st
2009/2/20 Aditya :
> I'm trying to install moses onto a 64-bit linux machine. I have compiled
> srilm with the MACHINE_TYPE variable set to "i686-m64". Now, when compiling
as far as i remember from my last 64bit install, Moses gets the
architecture type from the result of an SRILM script.
in the S
2008/12/16 Philipp Koehn :
> the only ones that I am aware of are:
> - IWSLT sets, althogh they are in a very specific tourist domain
Is the IWSLT/BTEC data actually freely-available?
>From reading the latest license agreement, it sounds like it isn't.
http://www.slc.atr.jp/IWSLT2008/downloads/IWS
2008/8/13 Sara Stymne <[EMAIL PROTECTED]>:
> When we installed SRILM and Moses on our 64-bit Ubuntu machine we had
> some troubles with getting the machine type right. What solved it in the
> end was to hack the machine-type script (found in srilm/sbin), so that
> it gave the correct machine type,
2008/8/5 John D. Burger <[EMAIL PROTECTED]>:
> I'm starting to think it's a lost cause to try to get one LM
> implementation to act very much like the other. Thanks for the
> insights, though!
I also spent some time unsuccessfully trying to exactly match the
SRILM toolkit's output. Aside from the
> 5. I thought building a bi-lingual corpus was having
> a text file with a sentence in the source language
> along with the sentence in the target language.
Yes. The basics of a parallel corpus are sentence pairs, where one
sentence is in one language, the other is in the other, and the two
a
24 matches
Mail list logo
