Hi,
> 1. Recently I've been trying to train my corpus on my dual core processor. > But it seemed that my system only used half the resource. On the process > manager, the CPU only loaded 50% of its full capacity. > > So, can anybody tell me how to maximize my system during training the > GIZA, run the mert-moses.pl, and doing the decoding? > I mean to use both core in the processor. > Is there any options to pass before I run those jobs? > You can use the --parallel option with train-factored-phrase-model.perlwhich will cause it to execute the two GIZA runs in parallel, which will probably be more efficient on a dual core system. GIZA itself does not support multi-core processing. > > 2. If I train the same corpus twice, using 2 different word alignment, e.g., > union and grow-diag-final, will I get different weight after running > mert-moses.pl; or it will be the same because I used exactly the same > corpus? > MERT is a non-deterministic algorithm and so you'll see different weights from run to run, even with the exact same alignment heuristics. But, yes, in general, changing any aspect of the system, including the alignment heuristic used, will result in different weights.
_______________________________________________ Moses-support mailing list Moses-support@mit.edu http://mailman.mit.edu/mailman/listinfo/moses-support
