Dear all, I tried tree-based model with/without syntactic tree. (hierarchical model and tree-to-string model)
I got the "moses_chart and tried following two commands: 1) For hierarchical model #> perl train-model.perl -root-dir work_dir -hierarchical -glue-grammar --coropus train.token -f en -e kr -lm 0:3:srilm_file --parallel 2) For tree-to-stirng model #> perl train-model.perl -root-dir work_dir -hierarchical -glue-grammar -source-syntax --corpus train.xml -f en -e kr -lm 0:3:srilm_file --parallel "train.xml.en" file is a XML formatted charniak parse result. These two commands generated glue-grammar and rule_table.gz files. But the extracted rules are like followings: all [X] ||| 모든 [X] ||| ||| 0.869565 0.627907 0.645161 0.243243 2.718 ||| 23 31 [X][X] trace ' [X] ||| [X][X] 추적 ' [X] ||| 0-0 ||| 0.727273 0.444625 1 0.172348 2.718 ||| 0.366667 0.266667 [X][X] to know consumers [X] ||| 소비자 를 알 기위하 [X][X] [X] ||| 0-4 ||| 1 0.00991236 1 0.00402288 2.718 ||| 0.2 0.2 As my knowledge, rules should have "input and output non-termial": [X] [X] ||| all [X] ||| 모든 [X] ||| ||| 0.869565 0.627907 0.645161 0.243243 2.718 ||| 23 31 or (in case of tree-to-string) [NP] [X] ||| all [NP] ||| 모든 [X] ||| ||| 0.869565 0.627907 0.645161 0.243243 2.718 ||| 23 31 Anyone know why the rules are not contain parent symbol (input and output non-terminal)? Thank You with regards _______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
