Commit Graph

15 Commits

Author SHA1 Message Date
Rico Sennrich
5dce1463e7 documentation: -phrase-word-alignment is on by default. 2013-04-09 11:15:28 +02:00
Rico Sennrich
50a8e94a8f tmcombine: better error message for unsupported phrase tables 2013-02-05 11:44:50 +01:00
Rico Sennrich
647bc11c0a fix display of error message 2013-01-11 11:24:41 +01:00
Rico Sennrich
06126ae6e2 tmcombine: ignore phrase pairs with 0 probability in optimization 2012-10-29 10:29:51 +01:00
Rico Sennrich
9e03090ca2 fix regression from having all 3 counts in phrase table. 2012-10-18 16:09:38 +02:00
Rico Sennrich
0f3de74934 possible fix for tmcombine for models without alignment info 2012-09-11 14:33:30 +02:00
Rico Sennrich
4e2fc82854 new training option -write-lexical-counts
(creates additional files lex.counts.e2f and lex.counts.f2e)
2012-09-06 11:48:54 +02:00
Rico Sennrich
dd2cf70f3f tmcombine: fix combination of lexical reordering models 2012-07-19 10:31:46 +02:00
Rico Sennrich
9dd42dd9b6 tmcombine: fix crash with option --recompute-lexweights 2012-07-16 13:26:39 +02:00
Rico Sennrich
ea4d0a4785 fix encoding issue
Error that occurred for phrase tables without phrase alignment:

  File "tmcombine.py", line 377, in write_phrase_table
    line = b"%s ||| %s ||| %s 2.718 %s||| %s ||| %s\n" %(src,target,features,origin_features,alignment,comments)
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 26: ordinal not in range(128)  File "tmcombine.py", line 377, in write_phrase_table
2012-07-01 20:48:12 +02:00
Rico Sennrich
fcf05e63c1 make translation model combination more general: allow non-default number and position of phrase table features 2012-05-31 14:23:38 +02:00
Rico Sennrich
b99ebd81cd documentation 2012-04-17 14:03:41 +02:00
Rico Sennrich
7927979298 use specified tmpdir when sorting 2012-02-28 10:05:30 +01:00
Rico Sennrich
e9d960d73e tmpdir option and bugfix to reordering 2012-02-27 14:11:47 +01:00
Rico Sennrich
8ee3ed6d64 tmcombine (translation model combination) 2012-01-31 10:50:20 +01:00