Michael Denkowski
|
49f4eeca47
|
default-average-others option for PhraseDictionaryGroup
Average scores of other models to fill in gaps for unseen phrases
|
2015-11-12 16:33:52 -05:00 |
|
Michael Denkowski
|
652283c538
|
Consistent naming: zeros -> default-scores
(Matches LexicalReordering)
|
2015-11-10 15:48:41 -05:00 |
|
Michael Denkowski
|
fb7be206ed
|
Copy extra scores when merging phrases in PhraseDictionaryGroup
This adds compatibility for lr-func in member mmsapts
|
2015-11-06 12:54:14 -05:00 |
|
Michael Denkowski
|
9f28bf5619
|
Specify "zero" values for feature scores for PhraseDictionaryGroup
User-specified default scores for when models in the group do not
contain a phrase pair.
|
2015-10-27 14:35:45 -04:00 |
|
MosesAdmin
|
d757346b55
|
daily automatic beautifier
|
2015-10-19 00:00:40 +01:00 |
|
Ulrich Germann
|
b8b3d2111a
|
Merge branch 'master' of http://github.com/moses-smt/mosesdecoder into ranked-sampling
|
2015-10-18 21:45:27 +01:00 |
|
Ulrich Germann
|
bdb0227ee9
|
Life cycle of TargetPhraseCollection is now managed via shared pointers.
|
2015-10-18 21:27:58 +01:00 |
|
Hieu Hoang
|
55698d15bc
|
templatize Comparer for unordered maps/set
|
2015-10-18 01:02:38 +01:00 |
|
MosesAdmin
|
df5007ef47
|
daily automatic beautifier
|
2015-09-17 00:00:35 +01:00 |
|
Michael Denkowski
|
245539d90b
|
Update phrase table combination interface
- Move "all" modes from PhraseDictionaryMultiModel to PhraseDictionaryGroup
- Refactor combination logic
|
2015-09-15 15:01:07 +01:00 |
|