yes, it does tree-to-tree, string-to-tree, tree-to-string. I also think it does SAMT. The basic stuff are there for people to play with & extend

the filter script
filter-model-given-input.pl -Hierarchical
works with the new file format when u use the -Hierarchical argument

On 22/05/2010 11:24, dongxinghua0213 wrote:
thank you for your reply, the problem is solved. there are many translation models on tree-string , string-based ,or tree-tree, on which moses implemented is based ? Could you give me some details,or some papersto refer to ?
another problem,
When I train the hierarchical phrase-based model using the following command: tools/moses-scripts/scripts-20100517-1641/training/train-model.perl -scripts-root-dir tools/moses-scripts/scripts-20100517-1641 -root-dir hier-dir -corpus hier-dir/corpus/raw -f ch -e uy -alignment grow-diag-final-and -parallel -hierarchical -glue-grammar -max-phrase-length 5 -lm 0:3:/home/top/programming/language_model/hier-dir/lm/raw.uy.lm:0

errors occurred,seeming this:
......
(6.2) creating table half hier-dir/model/rule-table.half.f2e @ Thu May 20 17:08:43 CST 2010 tools/moses-scripts/scripts-20100517-1641/training/phrase-extract/score hier-dir/model/extract.sorted hier-dir/model/lex.f2e hier-dir/model/rule-table.half.f2e --Hierarchical Executing: tools/moses-scripts/scripts-20100517-1641/training/phrase-extract/score hier-dir/model/extract.sorted hier-dir/model/lex.f2e hier-dir/model/rule-table.half.f2e --Hierarchical
Score v2.0 written by Philipp Koehn
scoring methods for extracted rules
processing hierarchical rules
Loading lexical translation table from hier-dir/model/lex.f2e.
.score: score.cpp:454: void outputPhrasePair(std::vector<PhraseAlignment*, std::allocator<PhraseAlignment*> >&, float): Assertion `bestAlignment->alignedToT[ j ].size() == 1' failed. ERROR: Execution of: tools/moses-scripts/scripts-20100517-1641/training/phrase-extract/score hier-dir/model/extract.sorted hier-dir/model/lex.f2e hier-dir/model/rule-table.half.f2e --Hierarchical
died with signal 6, without coredump
I have no idear about it , but when I used the clean script to filter my corpus ,it made sentence length shorter than 15,it worked! Are the corpus too big to fit into the RAM of the machine,or the problem my corpus itself ? is there another script to filter it ?
thank you again ! best wishes!



??2010-05-22 09:58:13??"Hieu Hoang" <[email protected] <mailto:[email protected]>> ??????

    you should binarise your phrase table, otherwise it takes up too
    much memory. This is described at the bottom of this page
    http://www.statmt.org/moses/?n=Moses.SyntaxTutorial
    eg.
    ~/CreateOnDiskPt 1 1 5 100 0 rule-table.gz rules.bin
    then change the ini file so the rules are read from the binary
    phrase table

    [ttable-file]
    2 0 0 5
    /home/top/programming/language_model/working-dir/model/rules.bin
    6 0 0 1
    /home/top/programming/language_model/working-dir/model/glue-grammar



    On 22/05/2010 02:47, dongxinghua0213 wrote:
    the contents in moses.ini are as follows,but we can successed
    when decoding in sample-models !
    #########################
    ### MOSES CONFIG FILE ###
    #########################
    # input factors
    [input-factors]
    0
    # mapping steps
    [mapping]
    0 T 0
    1 T 1
    # translation tables: source-factors, target-factors, number of
    scores, file
    [ttable-file]
    6 0 0 5
    /home/top/programming/language_model/working-dir/model/rule-table.gz
    6 0 0 1
    /home/top/programming/language_model/working-dir/model/glue-grammar
    # no generation models, no generation-file section
    # language models: type(srilm/irstlm), factors, order, file
    [lmodel-file]
    0 0 3 /home/top/programming/language_model/working-dir/lm/raw.ch.lm

    # limit on how many phrase translations e for each phrase f are
    loaded
    # 0 = all elements loaded
    [ttable-limit]
    20

    # language model weights
    [weight-l]
    0.5000

    # translation model weights
    [weight-t]
    0.2
    0.2
    0.2
    0.2
    0.2
    1.0
    # no generation models, no weight-generation section
    # word penalty
    [weight-w]
    -1
    [cube-pruning-pop-limit]
    1000
    [glue-rule-type]
    0
    [non-terminals]
    X
    [search-algorithm]
    3
    [inputtype]
    3
    [max-chart-span]
    20
    1000


    ??2010-05-21 23:28:48??"Hieu Hoang" <[email protected]> ??????

        can you send me your ini file, a few line from each of your
        phrase tables and a few lines from you input. I'll try &
        figure out what the problem is

        2010/5/21 dongxinghua0213 <[email protected]
        <mailto:[email protected]>>

            when I use the following command to decoder, an error
            occurred !
            moses1/moses-chart-cmd/src/moses_chart -config
            working-dir/tuning/filtered/moses.ini -input-file
            working-dir/tuning/input> hier-dir/tuning/output


            Defined parameters (per moses.ini or switch):
            config: working-dir/tuning/filtered/moses.ini
            cube-pruning-pop-limit: 1000
            glue-rule-type: 0
            input-factors: 0
            input-file: working-dir/tuning/input
            inputtype: 3
            lmodel-file: 0 0 3
            /home/top/programming/language_model/working-dir/lm/raw.ch.lm

            mapping: 0 T 0 1 T 1
            max-chart-span: 20 1000
            non-terminals: X
            search-algorithm: 3
            ttable-file: 6 0 0 5
            
/home/top/programming/language_model/working-dir/tuning/filtered/phrase-table.0-0.1.1
            6 0 0 1
            
/home/top/programming/language_model/working-dir/tuning/filtered/phrase-table.0-0.2.1

            ttable-limit: 20
            weight-l: 0.5000
            weight-t: 0.2 0.2 0.2 0.2 0.2 1.0
            weight-w: -1
            Loading lexical distortion models...have 0 models
            Start loading LanguageModel
            /home/top/programming/language_model/working-dir/lm/raw.ch.lm
            : [0.000] seconds
            Finished loading LanguageModels : [0.000] seconds
            Using uniform ttable-limit of 20 for all translation tables.
            Start loading PhraseTable
            
/home/top/programming/language_model/working-dir/tuning/filtered/phrase-table.0-0.1.1
            : [0.000] seconds
            filePath:
            
/home/top/programming/language_model/working-dir/tuning/filtered/phrase-table.0-0.1.1
            Start loading new format pt model : [0.000] seconds
            moses_chart: Phrase.cpp:246: void
            Moses::Phrase::CreateFromStringNewFormat(Moses::FactorDirection,
            const std::vector<unsigned int, std::allocator<unsigned
            int> >&, const std::string&, const std::string&,
            Moses::Word&): Assertion `nextPos != string::npos' failed.
            thank you for your attention!


            
------------------------------------------------------------------------
            ??????????????????????????????????????????
            <http://ym.163.com/?from=od3>
            _______________________________________________
            Moses-support mailing list
            [email protected] <mailto:[email protected]>
            http://mailman.mit.edu/mailman/listinfo/moses-support




    ------------------------------------------------------------------------
    ??????????????????????????????????????????
<http://ym.163.com/?from=od3>



------------------------------------------------------------------------
?????????????????????????????????????????? <http://ym.163.com/?from=od3>
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support

Reply via email to