Commit Graph

363 Commits

Author SHA1 Message Date
Arianna Bisazza
ff276e9911 Fixed several bugs in LRscore-MERT. Namely, solved a float-to-int conversion; added hypothesis counter to the scores file to enable later computation of average reordering score; fixed special case of 1-word hypothesis; enabled reading of word-based alignments from n-best-list. 2012-09-24 15:40:18 +02:00
Colin Cherry
ae6ac1c2ae Merge branch 'master' of https://github.com/moses-smt/mosesdecoder 2012-09-14 14:07:53 -04:00
Colin Cherry
3fa95c022b Addad a "--safe-hope" option to kbmira.
This will limit the influence of model score on oracle (hope) selection.
Good for cases whith extremely large feature values. May make it the defult.
2012-09-14 13:58:28 -04:00
Barry Haddow
021f5702a7 remove obsolete file 2012-09-13 22:21:31 +01:00
bhaddow
d0f1c15105 enable single character option 2012-07-12 19:47:57 +01:00
Barry Haddow
c303142ab2 option to skip duplicate removal 2012-07-12 19:08:55 +01:00
Colin Cherry
662e7e7f64 As requested by my bosses: added NRC copyright to kbmira. 2012-07-10 13:13:50 -04:00
Hieu Hoang
7d664b745e Integrate Lexi's LR Score into tuning 2012-07-10 09:25:00 +01:00
Hieu Hoang
75e038f4cf create namespace for all classes 2012-07-02 17:05:11 +01:00
Hieu Hoang
b5aa04feb7 compile error 2012-07-02 10:23:26 +01:00
Hieu Hoang
121e258e84 namespace all classes in mert directory 2012-06-30 21:39:10 +01:00
Hieu Hoang
e3dd3a8d2c namespace all classes in mert directory 2012-06-30 20:23:45 +01:00
Colin Cherry
65df386581 Merge branch 'master' of https://github.com/moses-smt/mosesdecoder 2012-06-26 17:07:27 -04:00
Colin Cherry
58c3280c2c HypPackEnumerator now stores MiraFeatureVectors, as opposed to
FeatureDataItems. Uses roughly half the memory.
2012-06-26 17:02:32 -04:00
Hieu Hoang
3c7b7ac9f5 rollback 2012-06-26 16:31:38 -04:00
Colin Cherry
32299593fa Added debugging info to kbmira. 2012-06-26 16:29:20 -04:00
Hieu Hoang
153e80053c lock m_vocab variable access in Encode() and Lookup(). Other functions are still not threadsafe 2012-06-26 13:33:50 -04:00
Hieu Hoang
00f018a477 Merge https://github.com/moses-smt/mosesdecoder into lrscore 2012-06-25 16:57:17 -04:00
Hieu Hoang
2a03f275a3 change regression data download to git instead of download from edin server.
Minor change in mert/trimStr() function to prevent warning
2012-06-25 16:03:11 -04:00
Hieu Hoang
8498b17a41 gcc version-specific error 2012-06-25 14:45:45 +01:00
Hieu Hoang
0fd0adc1f6 merge Lexi Birch's LRScore from mert_mtm5 branch. Compiles and run. Hack, must double check with barry or lexi 2012-06-23 22:58:18 -04:00
Hieu Hoang
0cb63edcb9 merge Lexi Birch's LRScore from mert_mtm5 branch. Compiles and run. Hack, must double check with barry or lexi 2012-06-23 22:51:48 -04:00
Hieu Hoang
f48c348508 typo 2012-06-22 22:23:14 -04:00
Hieu Hoang
b1ca36387f mert now compiles with PermScorer. However, didn't implement score() - assert(false). Update Jamfile 2012-06-22 21:07:05 -04:00
Hieu Hoang
7d19fe13ae merge Lexi Birch's LRScore from mert_mtm5 branch 2012-06-22 18:19:16 +01:00
Colin Cherry
a8a5f896db Fixed some bugs in BatchMira's sparse feature handling. 2012-06-14 14:09:06 -04:00
Tetsuo Kiso
1dbd8e5ec5 Merge branch 'master' of github.com:moses-smt/mosesdecoder 2012-06-09 19:33:31 +09:00
Tetsuo Kiso
2599ef6dc3 Bug fix: kbmira failed to load dense weights. 2012-06-09 18:03:12 +09:00
Tetsuo Kiso
713ff8c5e2 Delete mert/init.opt.
It looks like the file was no longer used.
2012-06-01 02:25:25 +09:00
Hieu Hoang
465c5cbf97 move all executables into bin 2012-05-31 12:55:05 +01:00
Tetsuo Kiso
beb2256dba Move 'using namespace std' out from .h.
Add "std" to size_t, too.
2012-05-30 23:11:09 +09:00
Tetsuo Kiso
01eb60f350 Add "virtual" destructor to the HypPackEnumerator class. 2012-05-30 22:59:23 +09:00
Hieu Hoang
d25805858d xcode build supports threads. move 'using namespace' out from .h file to stop namespace pollution 2012-05-30 13:04:02 +01:00
Hieu Hoang
45870348ff xcode build supports threads. move 'using namespace' out from .h file to stop namespace pollution 2012-05-30 12:47:20 +01:00
Colin Cherry
fd577d7a65 Batch k-best MIRA is written and integrated into mert-moses.pl
Regression tests all check out, and kbmira seems to work fine
on a Hansard French->English task.

HypPackEnumerator class may be of interest to pro.cpp and future
optimizers, as it abstracts a lot of the boilerplate involved in
enumerating multiple k-best lists.

MiraWeightVector is not really mira-specific - just a weight vector
that enables efficient averaging. Could be useful to a perceptron
as well. Same goes for MiraFeatureVector.

Interaction with sparse features is written, but untested.
2012-05-29 13:38:57 -04:00
Matous Machacek
a77cca4f86 Fixed CderScorer name bug 2012-05-15 00:35:08 +02:00
Matous Machacek
3943112eb3 Fixed bug in SemposScorer.cpp 2012-05-13 11:11:13 +02:00
Matous Machacek
97f82a3e4d Fixed interpolated scorer 2012-05-12 16:11:33 +02:00
Christian Federmann
c360b5a329 Merge remote-tracking branch 'upstream/master' 2012-05-10 09:49:57 +02:00
Tetsuo Kiso
9c9d88a78a Avoid "using namespace std" in headers. 2012-05-10 07:51:05 +09:00
Tetsuo Kiso
aa269bdb4d Merge branch 'master' of github.com:moses-smt/mosesdecoder 2012-05-10 07:31:47 +09:00
Tetsuo Kiso
5f7967402a Reduce compilation dependencies. 2012-05-10 07:16:38 +09:00
Matous Machacek
31edaec4c2 Bigger buffer 2012-05-10 00:08:54 +02:00
Tetsuo Kiso
afa356aec4 Small changes to just improve the quality.
- Use forward declaration to reduce dependencies.
- Add "virtual" to the destructor of _fdstream class.
- Avoid using namespace std in header.
  - We have already used in mert a lot, though. It should be fixed.
- Fix warnings "-Wreorder".
- Fix the usage of enum.
2012-05-10 06:57:44 +09:00
Christian Federmann
c1cf0faf14 GzFileBuf.h got lost in the last merge ;) 2012-05-09 21:52:59 +02:00
Christian Federmann
25f43d13b8 Merged in upstream. 2012-05-09 21:10:52 +02:00
Matous Machacek
440650bd6e Added support for external unix filters to preprocess sentences in mert and evaluator 2012-05-09 19:21:41 +02:00
Tetsuo Kiso
dbfe766f2c Fix using directive refers to implicitly-defined namespace 'std'. 2012-05-06 05:27:04 +09:00
Tetsuo Kiso
b6f55f6bec Fix using namespace "std" before including headers in *.h. 2012-05-06 01:31:04 +09:00
Tetsuo Kiso
1b634488fe Prefix a private member of mert/CderScorer with "m_". 2012-05-03 05:38:35 +09:00