El dc 31 de 08 de 2011 a les 14:37 +0530, en/na amba kulkarni va
escriure:
> Thanks Fran. This is what I was also suspecting.
> 
> Well, I would like to know about the performance -- if I split it into
> several dix files, and build more several bins out of it, run them
> serially to get the analysis, will the performance be of similar order
> compared to  a single bin?

This can't be done at the moment I think, because the analyser can't
accept input of LUs (^.../...$) for analysis -- although it would be an
interesting feature. 

> I have  4GB RAM, and I could get a single bin with lt-comp. The size
> of this big bin file is 17M.
> Recently I added some more data and then lt-expand stopped working.
> But before that I had 117MB word forms, and still I have not covered
> many issues in derivational morph.

Do you have access to a machine with more RAM that you could test it
on ? 

Regarding lt-expand, if the dictionary compiles, then you could just
e.g. strip out the interesting lemmas that you want to expand. Or is
there a reason you need to expand the whole thing ? 

Fran




------------------------------------------------------------------------------
Special Offer -- Download ArcSight Logger for FREE!
Finally, a world-class log management solution at an even better 
price-free! And you'll get a free "Love Thy Logs" t-shirt when you
download Logger. Secure your free ArcSight Logger TODAY!
http://p.sf.net/sfu/arcsisghtdev2dev
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff

Reply via email to