Phuong,
I'm not sure what you are asking here.
Are you trying to build your own online demo system? If you are
talking about the initial startup time for Moses, before translations
occur, then consider using binary phrase tables and language models.
Moses does not include a built-in server mode for web use -- here is
a link to a discussion about the same topic:
http://thread.gmane.org/gmane.comp.nlp.moses.user/610
If you are using someone else's online demonstration, what site were
you using?
Hope this helps,
Josh
On 22 Feb 2008, at 04:44, Phuong Thao wrote:
Hi all,
I had tested Moses online demonstrations and I see that the
translation speed is not quite different between the first and the
next run times. It means whenever I enter a sentence and click to
translate it, all the decoding procedures are operated, from loading
the model in to the memory to doing translation. So it takes much
loading time. Is there any way to keep the model in the memory and
so just wastes translating time when we run it on the Web?
Thank you.
P.Thao.
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support
