Hi all,
rather than having to search through email archive, as I guess we are
not the only one who won't use SRILM because it is proprietary (or
some other reason), I thought the best would be to modify the existing
script to be able to switch to IRSTLM when desired. I have just made a
pull reques
Hi,
2011/11/25 Philipp Koehn :
> Hi,
>
> check the FAQ:
> http://www.statmt.org/moses/?n=Moses.AdvancedFeatures#ntoc27
I think he is actually speaking of this page you mention. It explains
all the data preparation, then at the very end of the section, part
"update model", it just says:
«
Now tha
Hi all,
I've got a quick question regarding the tuning of hierarchical phrase-based
models. When calling mert, my terminal outputs an error I can't find in the
mailing lists:
main::create_extractor_script() called too early to check prototype at
./mert-moses.pl line 666
The script didn't stop a
Ah yes... the usual suspect ;)
Might be worth implementing though... let's see.
Cheers,
Christian
Christophe Servan hat am 24. November 2011 um
19:19 geschrieben:
> Hi,
> Maybe the lack of time ;-)
>
> Christophe
>
>
> Le 24/11/2011 19:10, Christian Federmann a écrit :
> > Hi all,
>
Hi,
Maybe the lack of time ;-)
Christophe
Le 24/11/2011 19:10, Christian Federmann a écrit :
> Hi all,
>
> is there a specific reason/obstacle that METEOR is not available for MERT?
>
> Cheers,
> Christian
>
>
>
>
> Christophe Servan hat am 24. November 2011 um
> 19:02 geschrieb
Hi all,
is there a specific reason/obstacle that METEOR is not available for MERT?
Cheers,
Christian
Christophe Servan hat am 24. November 2011 um
19:02 geschrieben:
> Hi Wilker,
> I confirm there is currently no scorer for METEOR in MERT.
> Do you plan do implement it ?
>
> Best reg
Hi Wilker,
I confirm there is currently no scorer for METEOR in MERT.
Do you plan do implement it ?
Best regards,
Christophe
--
Christophe SERVAN - PhD - Postdoc
Laboratoire Informatique de l'Université du Maine (LIUM)
Institut Claude Cha
Hi,
check the FAQ:
http://www.statmt.org/moses/?n=Moses.AdvancedFeatures#ntoc27
-phi
2011/11/21 Kádár Tamás (KTamas) :
> Hi
>
> I'm experimenting with incremental training. The documentation says:
> "Updating a running Moses instance is done via XML RPC". And, uh,
> that's it. I can't find any o
Hi all,
I checked on Moses manual and also the mailing list archives and I couldn't
quite find the information I need, so I'm trying with you.
As I understood, zmert was remove from Moses and MERT is now able to
optimize towards different metrics.
I know that a version of METEOR adapted for zmert
Hi,
if you get errors such as
ERROR: Forbidden zero sentence length 0
ERROR: Forbidden zero sentence length 0
the you have empty lines in your parallel corpus.
Please run the corpus cleaning script first.
-phi
On Tue, Nov 22, 2011 at 3:39 PM, Raja Bensalem
wrote:
> Hello
> I'm translating fr
the problem is that there are many ways to create five grams etc from
common words --words you will find in a large parallel set. you will
save on space, but it will not be significant.
nowadays there really is no need to use SRILM at all. before
considering drastic action, look at these other m
http://kheafield.com/code/filter.html filters ARPA files. This is a
separate project from KenLM. I think most LM toolkits don't
check/require that the ARPA is normalized. At least, KenLM, IRSTLM, and
SRILM work fine with filtered models.
On 11/24/11 13:07, Miles Osborne wrote:
> this can be don
this can be done, but it tends to not save much space. also it does
not help deal with OOVs, which the language model can still score even
though they are not in the parallel set.
if you are worried about saving space then you should either look at
KenLM or RandLM
Miles
On 24 November 2011 12:5
Dear all,
I hope that this is not too stupid a question, and that it hasn't been asked
recently.
In the MOSES EMS, when running experiments the phrase table is automatically
reduced to only those phrases that actually occur in the respective dev/test
set. Obviously this saves a lot of memory
14 matches
Mail list logo