Hi everyone, who can build huge LM with IRSTLM?
___
Moses-support mailing list
Moses-support@mit.edu
http://mailman.mit.edu/mailman/listinfo/moses-support
I'm trying to switch between (M)GIZA and BerkeleyAligner without
switching to the EMS. I'm not sure where to start.
The train-model.perl
script runs through 9 stages when creating the phrase and reordering
tables. Step 2 is the "GIZA++" step.
Is using BerkeleyAligner as simple
as replacing st
How big is "huge?"
You can use build-lm.sh with the -k option to split the training job into
smaller parts and conserve memory.
Moses with irstlm supports memory mapped irstlm files that reduces memory
use during decoding.
If these two techniques stretch your resources, you might consider using
there's been some problems with the website. It'll be back soon
**cough**, so keep checking everyday
On 12/11/2010 00:03, Shibamouli Lahiri wrote:
Thank you very much for the solution. The only problem now is that
statmt website is down I guess, so I'll have to wait for some time.
Regards,
Shi
Hello,
I think there might be a problem with some new code in mert-moses.pl.
I got tuning to run by commenting out line 1225 where values for
weight-d are pushed into the used_triples array.
--
Regards,
John J Morgan
___
Moses-support mailing list
Mose
Hi,
I had a query with regard to use of lattice input in moses.
There is a little difference in the translations generated when I run moses
using the 'normal' input format and when I run it with 'lattice input'
format.
The translations weren't radically different - only a few phrases were
differen
Hi,
I want to know if the sequence of 5 different phrase translation scores
in the translation table is important in the process of training and tuning.
These 5 scores, as mentioned in tutorial, are φ(f|e),lex(f|e),φ(e|f),
lex(e|f),phrase penalty.
I tried to change the position of the first scor
Hi,
The order has no significance, but due to the random elements in MERT,
slightly different results are possible after tuning.
-phi
On 12 Nov 2010 18:10, "RaEn" wrote:
>
> Hi,
>
> I want to know if the sequence of 5 different phrase translation scores
> in the translation table is important i