moses-smt/mosesdecoder
1
1
Fork 0
mirror of https://github.com/moses-smt/mosesdecoder.git synced 2024-09-11 11:25:40 +03:00
Code Issues Packages Projects Releases Wiki Activity

handles hiero models too

Browse Source
This commit is contained in:
Hieu Hoang 2016-10-05 21:33:19 +01:00
parent 5069aa4370
commit bcea640c9a
1 changed files with 31 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

41
scripts/generic/binarize4moses2.perl
View File

@ -12,22 +12,23 @@ my $mosesDir = "$RealBin/../..";
my $ptPath;
my $lexRoPath;
my $outPath;
my $numScores = 4;
my $numLexScores;
my $pruneNum = 0;
my $scfg = 0;
GetOptions("phrase-table=s" => \$ptPath,
"lex-ro=s" => \$lexRoPath,
"output-dir=s" => \$outPath,
"num-scores=s" => \$numScores,
"num-lex-scores=i" => \$numLexScores,
"prune=i" => \$pruneNum
"prune=i" => \$pruneNum,
"scfg" => \$scfg
) or exit 1;
#print STDERR "scfg=$scfg \n";
die("ERROR: please set --phrase-table") unless defined($ptPath);
die("ERROR: please set --lex-ro") unless defined($lexRoPath);
#die("ERROR: please set --lex-ro") unless defined($lexRoPath);
die("ERROR: please set --output-dir") unless defined($outPath);
die("ERROR: please set --num-lex-scores") unless defined($numLexScores);
#die("ERROR: please set --num-lex-scores") unless defined($numLexScores);
my $cmd;
@ -37,13 +38,33 @@ my $tempPath = dirname($outPath) ."/tmp.$$";
$cmd = "gzip -dc $ptPath | $mosesDir/contrib/sigtest-filter/filter-pt -n $pruneNum | gzip -c > $tempPath/pt.gz";
systemCheck($cmd);
$cmd = "$mosesDir/bin/processLexicalTableMin -in $lexRoPath -out $tempPath/lex-ro -T . -threads all";
systemCheck($cmd);
if (defined($lexRoPath)) {
die("ERROR: please set --num-lex-scores") unless defined($numLexScores);
$cmd = "$mosesDir/bin/addLexROtoPT $tempPath/pt.gz $tempPath/lex-ro.minlexr | gzip -c > $tempPath/pt.withLexRO.gz";
systemCheck($cmd);
$cmd = "$mosesDir/bin/processLexicalTableMin -in $lexRoPath -out $tempPath/lex-ro -T . -threads all";
systemCheck($cmd);
$cmd = "$mosesDir/bin/addLexROtoPT $tempPath/pt.gz $tempPath/lex-ro.minlexr | gzip -c > $tempPath/pt.withLexRO.gz";
systemCheck($cmd);
$cmd = "ln -s pt.withLexRO.gz $tempPath/pt.txt.gz";
systemCheck($cmd);
}
else {
$cmd = "ln -s pt.gz $tempPath/pt.txt.gz";
systemCheck($cmd);
}
$cmd = "$mosesDir/bin/CreateProbingPT2 --log-prob --input-pt $tempPath/pt.txt.gz --output-dir $outPath";
if (defined($lexRoPath)) {
$cmd .= " --num-lex-scores $numLexScores";
}
if ($scfg) {
$cmd .= " --scfg";
}
$cmd = "$mosesDir/bin/CreateProbingPT2 --num-scores $numScores --num-lex-scores $numLexScores --log-prob --input-pt $tempPath/pt.withLexRO.gz --output-dir $outPath";
systemCheck($cmd);
exit(0);