f973293 | evahasler | 10 December 2007, 18:41:09 UTC | git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/tags/eva_2d_pruning@1522 1f5c12ca-751b-0410-a591-d2e778427230 | 10 December 2007, 18:41:09 UTC |
78728d7 | evahasler | 27 November 2007, 16:25:19 UTC | delete CompressedRepr() from Bitmap git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1521 1f5c12ca-751b-0410-a591-d2e778427230 | 27 November 2007, 16:25:19 UTC |
858f5c1 | evahasler | 22 November 2007, 16:48:57 UTC | fixed error in WordsBitmap::GetCompressedRepresentation() git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1516 1f5c12ca-751b-0410-a591-d2e778427230 | 22 November 2007, 16:48:57 UTC |
6fd3dac | hieuhoang1972 | 08 November 2007, 22:40:14 UTC | changes for visual studio compilation git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1503 1f5c12ca-751b-0410-a591-d2e778427230 | 08 November 2007, 22:40:14 UTC |
e6adb61 | evahasler | 08 November 2007, 19:47:47 UTC | cube pruning fixed for maxDistortion != -1 git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1502 1f5c12ca-751b-0410-a591-d2e778427230 | 08 November 2007, 19:47:47 UTC |
312afd5 | evahasler | 07 November 2007, 19:01:33 UTC | implemented cube pruning for maxDistortion = -1 git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1501 1f5c12ca-751b-0410-a591-d2e778427230 | 07 November 2007, 19:01:33 UTC |
4155cd3 | evahasler | 06 November 2007, 12:01:35 UTC | -- building a 3*3 grid of a coverage set and possible extensions git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1500 1f5c12ca-751b-0410-a591-d2e778427230 | 06 November 2007, 12:01:35 UTC |
0df5fb4 | evahasler | 06 November 2007, 12:00:24 UTC | building a 3*3 grid of a coverage set and possible extensions git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1499 1f5c12ca-751b-0410-a591-d2e778427230 | 06 November 2007, 12:00:24 UTC |
e615f31 | evahasler | 31 October 2007, 19:42:18 UTC | not much git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1495 1f5c12ca-751b-0410-a591-d2e778427230 | 31 October 2007, 19:42:18 UTC |
9f11e5a | evahasler | 31 October 2007, 02:11:08 UTC | added HypothesisStack::GetCoverageSet(..) added Manager::ProcessCoverageSet(..) git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1493 1f5c12ca-751b-0410-a591-d2e778427230 | 31 October 2007, 02:11:08 UTC |
26b26ac | evahasler | 30 October 2007, 16:07:17 UTC | test junk git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1492 1f5c12ca-751b-0410-a591-d2e778427230 | 30 October 2007, 16:07:17 UTC |
72abf00 | hieuhoang1972 | 26 October 2007, 15:37:31 UTC | made a copy git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/eva_cube_pruning/moses@1489 1f5c12ca-751b-0410-a591-d2e778427230 | 26 October 2007, 15:37:31 UTC |
8070922 | hieuhoang1972 | 21 October 2007, 20:56:23 UTC | visual studio project & minor code clean git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1475 1f5c12ca-751b-0410-a591-d2e778427230 | 21 October 2007, 20:56:23 UTC |
ab1d8e0 | redpony | 28 September 2007, 16:43:33 UTC | 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/moses@1468 1f5c12ca-751b-0410-a591-d2e778427230 | 28 September 2007, 16:43:33 UTC |
135e7b6 | redpony | 28 September 2007, 15:44:53 UTC | fix tracking of source phrases in CNs/Word Lattices git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1467 1f5c12ca-751b-0410-a591-d2e778427230 | 28 September 2007, 15:44:53 UTC |
d15af77 | redpony | 27 September 2007, 23:22:58 UTC | is this the last of the word lattice distortion errors? maybe. i hope so. the nbest list is at least in the right order now. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1466 1f5c12ca-751b-0410-a591-d2e778427230 | 27 September 2007, 23:22:58 UTC |
7f361e2 | hieuhoang1972 | 27 September 2007, 14:16:28 UTC | floor score git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1465 1f5c12ca-751b-0410-a591-d2e778427230 | 27 September 2007, 14:16:28 UTC |
a890458 | redpony | 27 September 2007, 12:14:42 UTC | stricter checks on word lattice distortion git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1464 1f5c12ca-751b-0410-a591-d2e778427230 | 27 September 2007, 12:14:42 UTC |
1179cd9 | redpony | 27 September 2007, 05:03:06 UTC | fixed a subtle bug in the word lattice decoder. it now acts sane. This experience also gave me more evidence for thinking that numbering the words in a phrase is not good-- we should be numbering the spaces between them. this would fix oh so many nasty things we've got (like the fact that we represent word drops with [n,n-1] etc). git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1463 1f5c12ca-751b-0410-a591-d2e778427230 | 27 September 2007, 05:03:06 UTC |
7ec496e | hieuhoang1972 | 17 September 2007, 17:51:11 UTC | visual studio proj git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1462 1f5c12ca-751b-0410-a591-d2e778427230 | 17 September 2007, 17:51:11 UTC |
3f3fd67 | redpony | 17 September 2007, 17:09:35 UTC | add timing logging git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1461 1f5c12ca-751b-0410-a591-d2e778427230 | 17 September 2007, 17:09:35 UTC |
78b6929 | redpony | 17 September 2007, 02:49:08 UTC | fix distortion limiting functionality for WordLattices. Now, the shortest path to a another node is used to determine whether the threshold has been exceeded. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1460 1f5c12ca-751b-0410-a591-d2e778427230 | 17 September 2007, 02:49:08 UTC |
d6aef33 | redpony | 16 September 2007, 19:55:06 UTC | refactor distortion limit code in preparation for fixing word lattice behavior git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1459 1f5c12ca-751b-0410-a591-d2e778427230 | 16 September 2007, 19:55:06 UTC |
6f959cb | redpony | 14 September 2007, 07:31:56 UTC | The primary change here is that the distortion distance in a word lattice is now defined to be the number of words on the shortest path between the previous phrase's ending span and the next phrase's starting span. Originally, the distance was defined to be the difference in node number, which could be extremely large even for words that are, in some sense, adjacent. Only the distortion scoring has been fixed for word lattices, but the distortion limit is still broken. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1458 1f5c12ca-751b-0410-a591-d2e778427230 | 14 September 2007, 07:31:56 UTC |
48c6f3f | jdschroeder | 05 September 2007, 10:46:18 UTC | updated xcode projects to include new source files. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1457 1f5c12ca-751b-0410-a591-d2e778427230 | 05 September 2007, 10:46:18 UTC |
8e94903 | hieuhoang1972 | 13 August 2007, 19:37:43 UTC | if using distinct n-best list, create deviant paths for hypothesis even if its not distinct. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1451 1f5c12ca-751b-0410-a591-d2e778427230 | 13 August 2007, 19:37:43 UTC |
8814be1 | hieuhoang1972 | 13 August 2007, 18:22:19 UTC | code clean. get ready for bug fix to distinct n-best code git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1450 1f5c12ca-751b-0410-a591-d2e778427230 | 13 August 2007, 18:22:19 UTC |
f5bb69d | redpony | 08 August 2007, 14:24:10 UTC | add error msgs when PCN parse errors happen git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1448 1f5c12ca-751b-0410-a591-d2e778427230 | 08 August 2007, 14:24:10 UTC |
05e874d | hieuhoang1972 | 08 August 2007, 14:20:30 UTC | don't skip empty lines git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1447 1f5c12ca-751b-0410-a591-d2e778427230 | 08 August 2007, 14:20:30 UTC |
8b2ae69 | phkoehn | 27 July 2007, 12:08:59 UTC | performance fix: saner behaviour for very very long sentences (hundreds of words) by using a maximum phrase length, which is set to 20 by default and can be changed with the switch "-max-phrase-length". git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1440 1f5c12ca-751b-0410-a591-d2e778427230 | 27 July 2007, 12:08:59 UTC |
f901f70 | hieuhoang1972 | 24 July 2007, 22:54:08 UTC | vs.net files & weird mem leak again git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1434 1f5c12ca-751b-0410-a591-d2e778427230 | 24 July 2007, 22:54:08 UTC |
2f40686 | hieuhoang1972 | 24 July 2007, 22:34:55 UTC | code cleanup. moved DecodeGraph to its own class git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1433 1f5c12ca-751b-0410-a591-d2e778427230 | 24 July 2007, 22:34:55 UTC |
7726993 | hieuhoang1972 | 24 July 2007, 21:48:06 UTC | code cleanup git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1432 1f5c12ca-751b-0410-a591-d2e778427230 | 24 July 2007, 21:48:06 UTC |
6ad0cf8 | hieuhoang1972 | 24 July 2007, 20:59:26 UTC | weird mem leak in vs.net git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1431 1f5c12ca-751b-0410-a591-d2e778427230 | 24 July 2007, 20:59:26 UTC |
8931ebe | hieuhoang1972 | 20 July 2007, 18:19:32 UTC | turn trans opt on by default git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1430 1f5c12ca-751b-0410-a591-d2e778427230 | 20 July 2007, 18:19:32 UTC |
82deab2 | hieuhoang1972 | 20 July 2007, 13:03:12 UTC | cache for translation options for speed-up. default to true. only works for sentence input git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1429 1f5c12ca-751b-0410-a591-d2e778427230 | 20 July 2007, 13:03:12 UTC |
4daf26c | phkoehn | 19 July 2007, 19:03:40 UTC | fixed memort blowup with caching bin ttable / bin reordering table git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1428 1f5c12ca-751b-0410-a591-d2e778427230 | 19 July 2007, 19:03:40 UTC |
f995581 | hieuhoang1972 | 17 July 2007, 20:45:51 UTC | build with SRILM in windows under different configuration git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1427 1f5c12ca-751b-0410-a591-d2e778427230 | 17 July 2007, 20:45:51 UTC |
d604374 | redpony | 16 July 2007, 22:27:33 UTC | prevent silent failure when there aren't enough distortion weights given with lexicalized reordering models git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1426 1f5c12ca-751b-0410-a591-d2e778427230 | 16 July 2007, 22:27:33 UTC |
a988546 | hieuhoang1972 | 10 July 2007, 13:18:09 UTC | UINT32 for windows comp git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1424 1f5c12ca-751b-0410-a591-d2e778427230 | 10 July 2007, 13:18:09 UTC |
151e67f | redpony | 10 July 2007, 12:44:40 UTC | fix for binary reordering format on 64-bit platforms. thanks to Jaakko Väyrynen for identifying the problem git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1423 1f5c12ca-751b-0410-a591-d2e778427230 | 10 July 2007, 12:44:40 UTC |
7da6ac8 | redpony | 06 June 2007, 04:49:17 UTC | reduce memory consumption by removing data members from Factor object. These were not used anywhere and are determined by context. Since the were not used, it is not even clear if they were being reliably set, so it's better to remove them. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1415 1f5c12ca-751b-0410-a591-d2e778427230 | 06 June 2007, 04:49:17 UTC |
a945808 | jdschroeder | 29 May 2007, 13:49:52 UTC | Deleted old xcode files. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1409 1f5c12ca-751b-0410-a591-d2e778427230 | 29 May 2007, 13:49:52 UTC |
3ce45a7 | jdschroeder | 29 May 2007, 09:03:00 UTC | Added Xcode 2.4 projects for moses and moses-cmd. Uses links to external srilm and irstlm libraries. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1406 1f5c12ca-751b-0410-a591-d2e778427230 | 29 May 2007, 09:03:00 UTC |
5557d51 | jdschroeder | 24 May 2007, 13:23:36 UTC | Added support for pharaoh-style XML tags in Sentence input to supplement phrase table translation options. Use command line flag -xml-input. Possible values are: -inclusive (places XML options into mix with regular phrase table choices) -exclusive (no phrase table entries that overlap any of the XML source words are allowed) -ignore (XML markup is removed from input but not used during decoding) -pass-through (default, previous moses behavior: XML data is erroneously treated as words) git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1398 1f5c12ca-751b-0410-a591-d2e778427230 | 24 May 2007, 13:23:36 UTC |
f711c24 | maurocettolo | 24 May 2007, 09:07:40 UTC | updated interface to new irstlm toolkit, able to handle micro-macro tag maps git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1397 1f5c12ca-751b-0410-a591-d2e778427230 | 24 May 2007, 09:07:40 UTC |
63c7591 | hieuhoang1972 | 16 May 2007, 13:31:21 UTC | assert number of phrase table weights git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1391 1f5c12ca-751b-0410-a591-d2e778427230 | 16 May 2007, 13:31:21 UTC |
264f643 | hieuhoang1972 | 16 May 2007, 13:08:55 UTC | rename LatticePath class and associated classes to TrellisPath, to avoid confusion with WordLattice input git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1389 1f5c12ca-751b-0410-a591-d2e778427230 | 16 May 2007, 13:08:55 UTC |
a94eb01 | hieuhoang1972 | 16 May 2007, 12:52:49 UTC | code tidy up git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1388 1f5c12ca-751b-0410-a591-d2e778427230 | 16 May 2007, 12:52:49 UTC |
8fbe9e2 | abarun | 15 May 2007, 21:34:21 UTC | Code for MBR decoding git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1384 1f5c12ca-751b-0410-a591-d2e778427230 | 15 May 2007, 21:34:21 UTC |
19bc4a1 | abarun | 15 May 2007, 20:54:01 UTC | Code for MBR decoding git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1382 1f5c12ca-751b-0410-a591-d2e778427230 | 15 May 2007, 20:54:01 UTC |
1b937a6 | hieuhoang1972 | 04 May 2007, 17:29:48 UTC | clean up lattice path code git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1378 1f5c12ca-751b-0410-a591-d2e778427230 | 04 May 2007, 17:29:48 UTC |
cbb85cc | hieuhoang1972 | 04 May 2007, 17:19:50 UTC | errm, pass regression git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1377 1f5c12ca-751b-0410-a591-d2e778427230 | 04 May 2007, 17:19:50 UTC |
5754fde | hieuhoang1972 | 04 May 2007, 16:37:23 UTC | compile error git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1376 1f5c12ca-751b-0410-a591-d2e778427230 | 04 May 2007, 16:37:23 UTC |
bebee22 | hieuhoang1972 | 04 May 2007, 16:12:49 UTC | compile error git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1375 1f5c12ca-751b-0410-a591-d2e778427230 | 04 May 2007, 16:12:49 UTC |
7fca0f6 | hieuhoang1972 | 04 May 2007, 15:57:43 UTC | clean up lattice path code git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1374 1f5c12ca-751b-0410-a591-d2e778427230 | 04 May 2007, 15:57:43 UTC |
3d9c9f7 | hieuhoang1972 | 04 May 2007, 15:43:41 UTC | fixed mem leak for distinct n-best git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1373 1f5c12ca-751b-0410-a591-d2e778427230 | 04 May 2007, 15:43:41 UTC |
332e55e | hieuhoang1972 | 26 April 2007, 11:45:56 UTC | make dependency between bin phrase table & confusion networks explicit git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1371 1f5c12ca-751b-0410-a591-d2e778427230 | 26 April 2007, 11:45:56 UTC |
009433c | hieuhoang1972 | 18 April 2007, 18:10:24 UTC | visual studio project git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1364 1f5c12ca-751b-0410-a591-d2e778427230 | 18 April 2007, 18:10:24 UTC |
8373597 | redpony | 18 April 2007, 16:49:51 UTC | fix bug with unknown words/CNs/lexical reordering git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1362 1f5c12ca-751b-0410-a591-d2e778427230 | 18 April 2007, 16:49:51 UTC |
82c4607 | hieuhoang1972 | 18 April 2007, 14:20:51 UTC | make lex reordering work with confusion net git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1361 1f5c12ca-751b-0410-a591-d2e778427230 | 18 April 2007, 14:20:51 UTC |
a2cef90 | redpony | 18 April 2007, 14:18:05 UTC | add word lattice type git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1360 1f5c12ca-751b-0410-a591-d2e778427230 | 18 April 2007, 14:18:05 UTC |
56b46ca | redpony | 18 April 2007, 14:08:46 UTC | Support for the decoding of arbitrary word lattices. Must be given in the form of a "plf" file, which is a little tricky. I'll add documentation at some point; for now, refer to the example plf file in the "lattice-surface" regression test. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1359 1f5c12ca-751b-0410-a591-d2e778427230 | 18 April 2007, 14:08:46 UTC |
439fbc4 | hieuhoang1972 | 13 April 2007, 07:19:05 UTC | check for .gz generation files git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1356 1f5c12ca-751b-0410-a591-d2e778427230 | 13 April 2007, 07:19:05 UTC |
c74ed14 | hieuhoang1972 | 12 April 2007, 13:44:40 UTC | fixed bug in creating bin phrase table git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1353 1f5c12ca-751b-0410-a591-d2e778427230 | 12 April 2007, 13:44:40 UTC |
b594944 | hieuhoang1972 | 11 April 2007, 19:56:43 UTC | Add alignment info to phrase table git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1351 1f5c12ca-751b-0410-a591-d2e778427230 | 11 April 2007, 19:56:43 UTC |
60427fb | hieuhoang1972 | 08 April 2007, 22:15:56 UTC | ability to load phrase table with alignment info git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1347 1f5c12ca-751b-0410-a591-d2e778427230 | 08 April 2007, 22:15:56 UTC |
2f28588 | phkoehn | 06 April 2007, 18:41:45 UTC | bug fix for lexicalized reordering: weights were wrongly assigned git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1345 1f5c12ca-751b-0410-a591-d2e778427230 | 06 April 2007, 18:41:45 UTC |
def71dd | hieuhoang1972 | 06 April 2007, 10:45:34 UTC | visual studio project git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1340 1f5c12ca-751b-0410-a591-d2e778427230 | 06 April 2007, 10:45:34 UTC |
c2aec8e | hieuhoang1972 | 06 April 2007, 00:31:40 UTC | remove mutable from m_reordering. store tran opt in list as non-const git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1339 1f5c12ca-751b-0410-a591-d2e778427230 | 06 April 2007, 00:31:40 UTC |
a392c4d | hieuhoang1972 | 06 April 2007, 00:24:25 UTC | re-introduce konrad lexicalized reordering but fix bugs with help from philipp. speed improvements by caching probabilities in Trans Opt git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1338 1f5c12ca-751b-0410-a591-d2e778427230 | 06 April 2007, 00:24:25 UTC |
f26cdaa | hieuhoang1972 | 26 March 2007, 19:40:35 UTC | roll back new Lexical Reordering git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1330 1f5c12ca-751b-0410-a591-d2e778427230 | 26 March 2007, 19:40:35 UTC |
bed8fc5 | konrad_rawlik | 15 March 2007, 12:43:27 UTC | queryLexicalTable git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1314 1f5c12ca-751b-0410-a591-d2e778427230 | 15 March 2007, 12:43:27 UTC |
7008234 | hieuhoang1972 | 15 March 2007, 10:11:28 UTC | change unsigned in to size_t git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1312 1f5c12ca-751b-0410-a591-d2e778427230 | 15 March 2007, 10:11:28 UTC |
d619614 | hieuhoang1972 | 14 March 2007, 22:44:36 UTC | set svn properties git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1310 1f5c12ca-751b-0410-a591-d2e778427230 | 14 March 2007, 22:44:36 UTC |
7c448ca | hieuhoang1972 | 14 March 2007, 20:07:58 UTC | get rid of idx arg in GetScoreProducerDescription() git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1306 1f5c12ca-751b-0410-a591-d2e778427230 | 14 March 2007, 20:07:58 UTC |
bcda8c3 | hieuhoang1972 | 14 March 2007, 13:02:00 UTC | lexicalised reodering working with visual studio git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1302 1f5c12ca-751b-0410-a591-d2e778427230 | 14 March 2007, 13:02:00 UTC |
99323b8 | konrad_rawlik | 14 March 2007, 12:42:35 UTC | Added lexical reordering changes git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1301 1f5c12ca-751b-0410-a591-d2e778427230 | 14 March 2007, 12:42:35 UTC |
d16119a | hieuhoang1972 | 07 March 2007, 22:15:03 UTC | change DecodeStep* to const DecodeStep* git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1287 1f5c12ca-751b-0410-a591-d2e778427230 | 07 March 2007, 22:15:03 UTC |
787b45a | hieuhoang1972 | 07 March 2007, 15:37:38 UTC | make naming of hypo stacks classes consistent git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1282 1f5c12ca-751b-0410-a591-d2e778427230 | 07 March 2007, 15:37:38 UTC |
b6be04a | hieuhoang1972 | 06 March 2007, 15:57:30 UTC | gcc compile error git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1276 1f5c12ca-751b-0410-a591-d2e778427230 | 06 March 2007, 15:57:30 UTC |
7e19dcf | hieuhoang1972 | 06 March 2007, 12:15:51 UTC | add comments and Reset() git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1275 1f5c12ca-751b-0410-a591-d2e778427230 | 06 March 2007, 12:15:51 UTC |
4570576 | hieuhoang1972 | 04 March 2007, 09:01:50 UTC | visual studio proj file git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1267 1f5c12ca-751b-0410-a591-d2e778427230 | 04 March 2007, 09:01:50 UTC |
adde31a | hieuhoang1972 | 04 March 2007, 00:43:45 UTC | make output of decimal places consistent by not format anywhere but in Main.cpp git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1266 1f5c12ca-751b-0410-a591-d2e778427230 | 04 March 2007, 00:43:45 UTC |
535ec78 | hieuhoang1972 | 04 March 2007, 00:17:54 UTC | make output of decimal places consistent by not format anywhere but in Main.cpp git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1265 1f5c12ca-751b-0410-a591-d2e778427230 | 04 March 2007, 00:17:54 UTC |
d29bc38 | hieuhoang1972 | 02 March 2007, 13:37:12 UTC | code cleanup git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1257 1f5c12ca-751b-0410-a591-d2e778427230 | 02 March 2007, 13:37:12 UTC |
c2360bd | hieuhoang1972 | 28 February 2007, 11:59:32 UTC | get rid of warning message in srilm class git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1251 1f5c12ca-751b-0410-a591-d2e778427230 | 28 February 2007, 11:59:32 UTC |
601a14e | phkoehn | 26 February 2007, 20:59:41 UTC | added option -include-alignment-in-n-best to include the word alignment for each sentence in the n-best list file git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1246 1f5c12ca-751b-0410-a591-d2e778427230 | 26 February 2007, 20:59:41 UTC |
0e1e674 | hieuhoang1972 | 26 February 2007, 16:49:32 UTC | visual studio output paths git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1245 1f5c12ca-751b-0410-a591-d2e778427230 | 26 February 2007, 16:49:32 UTC |
87fe54a | hieuhoang1972 | 24 February 2007, 20:32:32 UTC | move calling InitializeBeforeSentenceProcessing() & CleanUpAfterSentenceProcessing() in Manager class git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1241 1f5c12ca-751b-0410-a591-d2e778427230 | 24 February 2007, 20:32:32 UTC |
dd47f46 | hieuhoang1972 | 23 February 2007, 00:00:09 UTC | code cleanup git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1234 1f5c12ca-751b-0410-a591-d2e778427230 | 23 February 2007, 00:00:09 UTC |
77a21eb | hieuhoang1972 | 22 February 2007, 23:54:59 UTC | more verbose=0 git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1233 1f5c12ca-751b-0410-a591-d2e778427230 | 22 February 2007, 23:54:59 UTC |
9ebc482 | hieuhoang1972 | 22 February 2007, 23:44:38 UTC | verbose=0 nothing goes to stderr except for real, aborting errors git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1232 1f5c12ca-751b-0410-a591-d2e778427230 | 22 February 2007, 23:44:38 UTC |
4e063d1 | hieuhoang1972 | 22 February 2007, 12:04:53 UTC | fixed mem leak git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1230 1f5c12ca-751b-0410-a591-d2e778427230 | 22 February 2007, 12:04:53 UTC |
8d6fb61 | hieuhoang1972 | 21 February 2007, 20:05:27 UTC | minor gcc compile error. also, no longer use IRSTLM as a subsitutute for SRILM, and vice versa. They don't give identitcal results - avoids confusion. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1229 1f5c12ca-751b-0410-a591-d2e778427230 | 21 February 2007, 20:05:27 UTC |
7207c93 | hieuhoang1972 | 21 February 2007, 19:48:53 UTC | change unknown word processing to be closer to the way pharaoh does it - create unknown word whenever single word is not in translation table but penalise hypothesis for using it. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1227 1f5c12ca-751b-0410-a591-d2e778427230 | 21 February 2007, 19:48:53 UTC |
5372a33 | hieuhoang1972 | 20 February 2007, 11:08:55 UTC | hack to fix hypo collection where all hypo scores are -inf. need to rethink pruning or creation of trans opt for unknown word git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1225 1f5c12ca-751b-0410-a591-d2e778427230 | 20 February 2007, 11:08:55 UTC |
baea585 | hieuhoang1972 | 16 February 2007, 18:15:06 UTC | minor gcc compile error git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1219 1f5c12ca-751b-0410-a591-d2e778427230 | 16 February 2007, 18:15:06 UTC |
4a52442 | hieuhoang1972 | 16 February 2007, 18:08:37 UTC | code cleanup - make FactorCollection and StaticData totally accessible only globally git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1218 1f5c12ca-751b-0410-a591-d2e778427230 | 16 February 2007, 18:08:37 UTC |
ce36dc0 | lexi_birch | 16 February 2007, 15:56:44 UTC | Merging from branch. From now on, can have multiple decoder step lists to accomodate backoff Specify this as an extra parameter in the [mapping] option in the ini file This is backwards compatible. Before (and still accepted): [mapping] T 0 Now you can have: [mapping] 0 T 0 1 T 1 1 G 0 Imagine for instance the translation table 0 is words - words, and the table 1 is stems - stems, and the generation table 0 is stems - words. This will allow us to backoff to stems if words are not found. It is not really backoff because all the options from both decoder step lists get included into the translation option collection, which is then used to create the hypotheses. The different paths must have their weights carefully balanced. MERT might not be enough to discover the best weights for all the combined parameters. git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk/moses@1217 1f5c12ca-751b-0410-a591-d2e778427230 | 16 February 2007, 15:56:44 UTC |