Hello,
I have trained, tuned, and prepared an eval set, trying to get Moses
to decode Arabic to English. However, trying with either filtered or
unfiltered phrase tables, I get the following error:
moses: StaticData.cpp:772: bool
Moses::StaticData::LoadPhraseTables(): Assertion `weightAll.size() >=
weightAllOffset + numScoreComponent' failed.
Apparently something in my configuration is wrong, but I'm not sure
which weights/scores/components/offsets I've got misconfigured. Any
help appreciated.
Below is the complete moses output (this one using the filtered phrase
tables, but it's essentially the same when I use the moses.ini file
for the unfiltered tables):
Sun May 16 15:57:41 PDT 2010
Defined parameters (per moses.ini or switch):
config: /home/david/projects/mt-diac/evaluation/raw/moses.ini
distortion-limit: 6
input-factors: 0
input-file: /home/david/projects/mt-diac/data/testing.raw.ar
/home/david/projects/mt-diac/evaluation/raw/raw.tuned-filtered.output
/home/david/projects/mt-diac/evaluation/raw/raw.tuned-filtered.decode.out
lmodel-file: 0 0 3 /home/david/projects/mt-diac/data/ar-en.en.lm
mapping: 0 T 0
ttable-file: 0 0 0 5
/home/david/projects/mt-diac/evaluation/raw/phrase-table.0-0.1.1
ttable-limit: 20
weight-d: 0.093497
weight-l: 0.173478
weight-t: 0.0650884 0.129613 0.0552512 0.15415
weight-w: -0.328827
Loading lexical distortion models...have 0 models
Start loading LanguageModel
/home/david/projects/mt-diac/data/ar-en.en.lm : [0.000] seconds
/home/david/projects/mt-diac/data/ar-en.en.lm: line 2473: warning:
non-zero probability for <unk> in closed-vocabulary LM
Finished loading LanguageModels : [2.000] seconds
moses: StaticData.cpp:772: bool Moses::StaticData::LoadPhraseTables():
Assertion `weightAll.size() >= weightAllOffset + numScoreComponent'
failed.
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support
