ACK gotcha ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Chris Mattmann, Ph.D. Chief Architect Instrument Software and Science Data Systems Section (398) NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA Office: 168-519, Mailstop: 168-527 Email: chris.a.mattm...@nasa.gov WWW: http://sunset.usc.edu/~mattmann/ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Director, Information Retrieval and Data Science Group (IRDS) Adjunct Associate Professor, Computer Science Department University of Southern California, Los Angeles, CA 90089 USA WWW: http://irds.usc.edu/ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
On 4/5/16, 5:06 PM, "Matt Post" <p...@cs.jhu.edu> wrote: >Chris — I think this is the official posted language pack. From an email I >sent to you and Lewis on 2015-03-16: > >> Finally, this is ready: >> >> http://joshua-decoder.org/language-packs/es-en-phrase/ >> >> I apologize for the delay! It took some work implementing a few new features >> to be able to decode with a large unfiltered model like this efficiently. >> You will need to download the 6.0.1 release (which contains some bug fixes >> and also enables our efficient compiled grammar format for phrase-based >> models): >> >> http://joshua-decoder.org/releases/6.0/ >> >> There is a README in the language pack tarball. Please let me know how >> things go! >> >> matt > > > >> On Apr 5, 2016, at 1:55 PM, Mattmann, Chris A (3980) >> <chris.a.mattm...@jpl.nasa.gov> wrote: >> >> Matt you had built me a specialized one and I think it lives on >> our xdata machine somewhere I’ll take a look but if you still have >> it that would be grand >> >> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ >> Chris Mattmann, Ph.D. >> Chief Architect >> Instrument Software and Science Data Systems Section (398) >> NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA >> Office: 168-519, Mailstop: 168-527 >> Email: chris.a.mattm...@nasa.gov >> WWW: http://sunset.usc.edu/~mattmann/ >> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ >> Director, Information Retrieval and Data Science Group (IRDS) >> Adjunct Associate Professor, Computer Science Department >> University of Southern California, Los Angeles, CA 90089 USA >> WWW: http://irds.usc.edu/ >> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ >> >> >> >> >> >> >> >> >> >> On 4/5/16, 12:46 PM, "Matt Post" <p...@cs.jhu.edu> wrote: >> >>> >>> >>> >>> Can you not use the posted Spanish → English language pack? >>> >>> >>> >>> >>> On Apr 5, 2016, at 12:34 PM, Lewis John Mcgibbney >>> <lewis.mcgibb...@gmail.com> wrote: >>> >>> That was my concern Matt. Thanks for the explanation. >>> Basically I've been passed a packed grammar for Spanish --> English >>> translation (possibly one you built for Chris M. While back) and wondered >>> if I could package it into a language pack. It seems that without the >>> accompanying parameters we've discussed >>> its usefulness is questionable. >>> >>> On Tuesday, April 5, 2016, Matt Post <p...@cs.jhu.edu> wrote: >>> >>> Hi Lewis, >>> >>> >>> I would suggest taking another language pack and just copying that over. >>> You can delete the language model and grammars from it, and then edit the >>> joshua.config to do the following: >>> >>> >>> - Remove the language model lines: >>> >>> >>> feature-function = StateMinimizingLanguageModel ... >>> feature-function = LanguageModel ... >>> >>> >>> - Modify the ^tm lines to point to your new grammar. You will want to keep >>> the glue grammar. Also, the tm specification should match the type of >>> grammar you learned (set to "thrax" if it was a hiero grammar, or to >>> "phrase" or "moses" if a phrase >>> or moses one). >>> >>> >>> That said, what are you doing this for? It's unlikely to work very well >>> without a language model, and without tuned parameters. >>> >>> >>> matt >>> >>> >>> >>> >>> >>> On Apr 5, 2016, at 12:54 AM, Lewis John Mcgibbney >>> <lewis.mcgibb...@gmail.com<javascript:_e(%7B%7D,'cvml','lewis.mcgibb...@gmail.com');>> >>> wrote: >>> >>> Hi Folks, >>> >>> Does anyone know if it is possible to bundle a language pack without input >>> for the mandatory arguments as documented at [0]? >>> >>> Say for example I was passed a packed grammar which was built from a >>> pipeline execution however no accompanying configuration file, the tuned >>> configuration file, etc. was provided as well. >>> >>> Any help is appreciated. >>> >>> Thanks >>> >>> Lewis >>> >>> [0] >>> http://joshua-decoder.org/6.0/bundle.html >>> <http://joshua-decoder.org/6.0/bundle.html> >>> >>> -- >>> Lewis >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> -- >>> Lewis >
