Commit Graph

24 Commits

Author SHA1 Message Date
Philipp Koehn
b95c372e3a updates to EMS: mteval-v13a.pl, parallel preparation, better paths and defaults in examples 2011-12-21 04:26:27 +00:00
Philipp Koehn
d6035a0612 first cimmit to github: enable parallel data preparation in ems 2011-12-21 01:04:11 +00:00
Phil Williams
ee0a6dbd5c extract-ghkm: add the features required for use as a drop-in replacement
for extract-rules: composed rules, unaligned source word attachment,
non-lexical unary rule elimination, glue rule generation, unknown word
label generation, and EMS integration.
2011-11-21 16:21:04 +00:00
pjwilliams
a1ca2722df Add --MinNonInitialRuleCount option to filter-model-given-input.pl. This
prunes non-initial rules (i.e. rules with non-terminals) from the rule table
based on their frequency counts.  In Zollmann, Venugopal, Och, and Ponte (2008),
pruning hierarchical rules that occur only once was found to significantly
decrease rule table size without harming translation quality.

Also, add TUNING:filter-settings and EVALUATION[:<set>]:filter-settings
variables so that this can be enabled in the EMS.

git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@4033 1f5c12ca-751b-0410-a591-d2e778427230
2011-06-24 16:36:27 +00:00
phkoehn
6acd6a8684 improvements to ems analysis
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@4026 1f5c12ca-751b-0410-a591-d2e778427230
2011-06-21 21:52:13 +00:00
phkoehn
4285a8e236 various experiment.perl improvements: split filter and decode/tune; extensions to analysis.perl, especially precision by coverage graphs
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@4018 1f5c12ca-751b-0410-a591-d2e778427230
2011-06-16 23:43:29 +00:00
bhaddow
c7603add22 Option to specify language model type
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@4013 1f5c12ca-751b-0410-a591-d2e778427230
2011-06-13 12:13:55 +00:00
phkoehn
4c11bcd617 extensions to phrase table scoring options
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3893 1f5c12ca-751b-0410-a591-d2e778427230
2011-02-23 10:27:54 +00:00
phkoehn
7334d49191 minor experiment.perl fixes
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3668 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-27 12:42:34 +00:00
phkoehn
ace33d16dd bug fix
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3636 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-22 07:30:52 +00:00
phkoehn
85a5a13e4c improvements to web analysis, fixes to syntax wrappers
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3633 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-21 09:49:27 +00:00
phkoehn
f34b37bad3 added hierarchical alignment view to web analysis tool
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3514 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-17 13:28:04 +00:00
suzyh
fa4eca6ccb Added loop to check_if_crashed in EMS experiment.perl to wait in case the .STDERR file is slow in appearing after the step has completed. Reinstated --old-sge and --filterfile command-line arguments to mert-moses.pl.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3420 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-09 11:40:40 +00:00
phkoehn
4a85fd95ce srilm as setting in interpolation
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3416 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-03 12:57:16 +00:00
bhaddow
12269b062c Only add queue-flags if non-empty.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3414 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-02 09:01:39 +00:00
phkoehn
4e0bc582f6 minor improvements: binarizing rule tables in filter script, multiple reference translation in ems analysis
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3284 1f5c12ca-751b-0410-a591-d2e778427230
2010-05-28 22:19:58 +00:00
pjwilliams
05eb33d5ac In train-model.perl, write an 'unknown-lhs' line out to the config file if
an unknown word label file has been generated.  Also, disable this option
by default since it can greatly increase number of hypotheses generated
and hasn't been shown to help translation yet.


git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3277 1f5c12ca-751b-0410-a591-d2e778427230
2010-05-27 15:11:16 +00:00
phkoehn
b271862d7c various updates, mostly related to experiment.perl
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3262 1f5c12ca-751b-0410-a591-d2e778427230
2010-05-18 17:39:16 +00:00
phkoehn
447dccfc59 more analysis in experiment.perl
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3234 1f5c12ca-751b-0410-a591-d2e778427230
2010-05-07 11:28:55 +00:00
phkoehn
45ecfa72d2 minor changes to experiment.perl
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3225 1f5c12ca-751b-0410-a591-d2e778427230
2010-05-05 16:06:34 +00:00
phkoehn
2ed6804f12 official release of experiment.perl
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3224 1f5c12ca-751b-0410-a591-d2e778427230
2010-05-04 23:04:10 +00:00
pjwilliams
4c6c4b71cf Remove redundant --ProperConditioning option from extract.
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@3118 1f5c12ca-751b-0410-a591-d2e778427230
2010-04-12 10:41:32 +00:00
phkoehn
244e334b3d bug fix
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@2690 1f5c12ca-751b-0410-a591-d2e778427230
2010-01-27 14:55:17 +00:00
phkoehn
def35604af initial release of experiment.perl
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@2669 1f5c12ca-751b-0410-a591-d2e778427230
2010-01-25 17:38:53 +00:00