Hi, parse-en-collins.perl does indeed try to run Collins' parser, so if you want to use the Berkeley parser, you will have to write a new wrapper script parse-en-berkeley.perl which is probably easiest by adapting the Collins parse script.
-phi On Tue, Jan 31, 2012 at 5:58 PM, <[email protected]> wrote: > > Hi all, > > I'm trying to train a tree-to-string model. I have parse the source > side with Berkeley parser, then edited it to match Collins parse > format, so that I can wrap it into Moses syntax tree format using > parse-en-collins.perl as instructed here: > http://www.statmt.org/moses/?n=Moses.SyntaxTutorial. > > However, I found that this script tries to locate Collins parser to parse > the input, which in fact I don't need since I have parses. I was wondering > if anyone could clarify this whether I'm misunderstanding the instruction > and script usage or advice me what the correct way of accomplishing this > goal is. > > Thank you very much in advance. > > Best Regards, > Rasul. > _______________________________________________ > 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
