Commit Graph

1190 Commits

Author SHA1 Message Date
sinnatagg
234ffc4727 Working
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/andre_lynum@1851 1f5c12ca-751b-0410-a591-d2e778427230
2008-06-11 13:51:33 +00:00
sinnatagg
3baf774ca5 git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/andre_lynum@1590 1f5c12ca-751b-0410-a591-d2e778427230 2008-03-25 08:20:46 +00:00
sinnatagg
483327a646 git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/andre_lynum@1589 1f5c12ca-751b-0410-a591-d2e778427230 2008-03-21 18:24:29 +00:00
sinnatagg
388bccaf79 Work on missing word score penalty.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/andre_lynum@1570 1f5c12ca-751b-0410-a591-d2e778427230
2008-02-23 11:03:24 +00:00
sinnatagg
844a05ea0f Easy test cases now work.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/andre_lynum@1551 1f5c12ca-751b-0410-a591-d2e778427230
2008-02-07 17:41:04 +00:00
sinnatagg
8dd8406ac9 Added translation option pruning and unknown word handling. Changed Manager::GetBestHypothesis to search in stack representing end of translated input.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/andre_lynum@1550 1f5c12ca-751b-0410-a591-d2e778427230
2008-02-06 20:51:24 +00:00
sinnatagg
77525d057a Added Scorer.h with scorer conditionals.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/andre_lynum@1546 1f5c12ca-751b-0410-a591-d2e778427230
2008-01-27 12:07:49 +00:00
sinnatagg
ca5772d3c2 Initial work on bilingual scorer.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/andre_lynum@1545 1f5c12ca-751b-0410-a591-d2e778427230
2008-01-27 12:02:29 +00:00
hieuhoang1972
9b7d13baba made a copy
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/andre_lynum@1541 1f5c12ca-751b-0410-a591-d2e778427230
2008-01-18 14:21:29 +00:00
phkoehn
1044b16522 bug fixes to distinct n-best list, mbr decoding; also changed switches for mbr
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1540 1f5c12ca-751b-0410-a591-d2e778427230
2008-01-16 20:25:01 +00:00
abarun
aa6f605a2d Updated xcode project
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1518 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-26 15:32:05 +00:00
nicolabertoldi
ae319da62b revert to /bin/sh for enhanced-mert; use of setenv (instead of export) in the csh scripts created by qsub-wrapper.pl
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1515 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-21 14:31:21 +00:00
nicolabertoldi
0176c5f8ec use fo csh instead of sh
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1514 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-21 07:59:17 +00:00
hieuhoang1972
dee505b823 don't use hypo pool until delayed destructor is sorted out
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1513 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-16 00:09:21 +00:00
hieuhoang1972
e2d8dfb420 bug fix for multiple phrase tables
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1512 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-13 16:18:42 +00:00
hieuhoang1972
2672b8cb34 variable renaming - less confusing & to make it comparable to async branch
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1511 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-12 20:21:44 +00:00
redpony
9521e78fbf add functionalty to support adding direct translation (ie, passthrouh) even when there are other translation options available for the span.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1510 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-10 23:19:41 +00:00
hieuhoang1972
80d66757c6 win32 compile
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1509 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-10 09:32:07 +00:00
redpony
180d9bac5d add support for unary XML options. refer to regression-testing/tests/xml-markup/to-translate.txt for examples.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1508 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-10 04:40:18 +00:00
redpony
3b008f67a9 add support for multiple xml options for the same span, start refactoring to add support for overlapping xml options.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1507 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-09 23:56:57 +00:00
maurocettolo
c4b566dcd7 make effective the use of mapped IRSTLM
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1506 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-09 08:06:09 +00:00
redpony
78cb04f13e use the built-in lgamma function instead of my stolen log_gamma function
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1505 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-09 01:14:18 +00:00
redpony
d4fa70a6ac enable filtering only based on p(e|f) [this is useful when you don't want to or can't build a suffix array of the training corpus, but still want to do some filtering]
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1504 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-09 01:08:45 +00:00
bojar
89ea9828ba added ttable iterator to this script, too
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1498 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-06 03:33:41 +00:00
bojar
09d8b5e657 improving documentation and allowing environment variables to override the
default paths


git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1497 1f5c12ca-751b-0410-a591-d2e778427230
2007-11-06 03:16:17 +00:00
hieuhoang1972
a2e805109a hide future cost - not needed by anyone else
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1496 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-31 22:40:59 +00:00
hieuhoang1972
d99aea83f5 cleanup scoring in TranslationOption
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1494 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-31 17:59:58 +00:00
nicolabertoldi
568f92b310 bug fixed
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1491 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-30 15:42:26 +00:00
jdschroeder
e52040bc12 added str length check to stop std::out_of_range error in a few more spots - similar bug to one corrected in v. 1319
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1488 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-26 12:42:55 +00:00
nicolabertoldi
fd3ecd4334 bug fixed
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1487 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-25 14:34:49 +00:00
nicolabertoldi
1b0576ba6c small bug fixed: temporary concatenated sorted file is now deleted only at the end
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1486 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-24 07:50:51 +00:00
nicolabertoldi
8710cc9bc9 features can be activated using a comma- or blank-separated list
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1485 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-23 16:55:02 +00:00
nicolabertoldi
9e70b5ffd8 Features are activated using their names
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1484 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-23 16:30:02 +00:00
nicolabertoldi
8fe62f2b95 some small bugs fixed and clean up
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1483 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-23 14:15:19 +00:00
nicolabertoldi
4720d1cb9f bug fixed
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1482 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-22 17:01:53 +00:00
nicolabertoldi
918dae011a bug fixed
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1481 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-22 14:00:10 +00:00
nicolabertoldi
e7ac20d4d6 bug fixed in the name of a temporary file
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1480 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-22 13:34:12 +00:00
nicolabertoldi
db9d0fc539 Added a more time-efficient (but more memory-consumptive) method to rescore nbest list
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1479 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-22 10:08:07 +00:00
nicolabertoldi
b827d51870 changes to cope with the new mert suite (enhanced-mert)
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1478 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-22 09:45:19 +00:00
jdschroeder
a969197e16 Fixed passing decoder parameters when tuning on single machine.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1477 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-22 09:27:09 +00:00
nicolabertoldi
5759005857 Suite of scripts to perform MERT on a subset of fetures.
Look at the directory example to learn about its use.


git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1476 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-22 09:04:03 +00:00
hieuhoang1972
05bb17422c visual studio project & minor code clean
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1475 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-21 20:56:23 +00:00
redpony
0cf583e249 add --hmm-align option. Allows using Giza++'s HMM word alignment model as the underlying word alignment. It is much faster than Model 4 alignment and not much worse.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1474 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-19 03:44:05 +00:00
nicolabertoldi
901823d83a explicit export of PYTHONPATH variable
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1473 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-18 15:33:04 +00:00
nicolabertoldi
81b439d728 minor changes in passing parameters to moses-parallel
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1472 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-18 12:36:49 +00:00
hieuhoang1972
4e1cad4bbe fixed sync/async bug
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1471 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-04 17:31:31 +00:00
redpony
57dcaa8e80 performance fixes for scorer
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1470 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-02 21:43:54 +00:00
hieuhoang1972
9cbc2922b4 separate word penalty for each decode step for async
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1469 1f5c12ca-751b-0410-a591-d2e778427230
2007-10-02 12:52:00 +00:00
redpony
7dbcef925d Implemented -r (-recover-input-path) to recover the actual path through the CN/word lattice that was used in the best translation. Also cleaned up handling of source Phrases in the context of a particular hypothesis, which may improve performance translating with lexicalized reordering models.
Removed PhraseReference since the concept was broken for CN/lattice inputs.



git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1468 1f5c12ca-751b-0410-a591-d2e778427230
2007-09-28 16:43:33 +00:00
redpony
d2928a0653 fix tracking of source phrases in CNs/Word Lattices
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1467 1f5c12ca-751b-0410-a591-d2e778427230
2007-09-28 15:44:53 +00:00