Hi John, Thanks for your response. Replies inline... On Wed, Oct 26, 2016 at 9:49 PM, < dev-digest-h...@joshua.incubator.apache.org> wrote:
> > From: John Hewitt <john...@seas.upenn.edu> > To: dev@joshua.incubator.apache.org > Cc: > Date: Thu, 27 Oct 2016 00:49:34 -0400 > Subject: Re: Pipeline Mystery > It seems like MERT isn't writing it's final config file (which is typical > of MERT, in my experience). I recall giving up and using kbmira. This final > config file is the one used in test, so I can see why skipping to test ends > up failing pretty quick. > >From my understanding, in order to use --tuner kbmira, I need to download, configure and run Moses. Is this correct? I would REALLY prefer not to do this if at all possible. In the meantime, it looks like I'm going to try another fresh pipeline run and see where I get. Sometimes starting afresh has lead to surprising and delightful results :) > > To answer your question, though, I haven't tried. Not in my bandwidth right > now. No problems. In all honesty, an entire pipeline execution on a small parallel dataset would be a killer smoke test(s) for any contributions coming into Joshua. Language pack creation is so important and having confidence in the overall process is something which I really look forward to building over the next while. Thanks