NAME¶

m2m-aligner - Parallel corpora aligner.

SYNOPSIS¶

m2m-aligner [OPTIONS]

DESCRIPTION¶

m2m-aligner

This tool aims at aligning parallel corpora. It can been used as a preliminary data preparation step in both machine translation and grapheme to phoneme training.

OPTIONS¶

--limit

Limit the alignment pair to used only from the initFile only (default false)

--errorInFile

Keep unaligned item in the output file (default false)

--initProb <long double>

Cut-off sum prior probability (default 0.5)

--init <string>

Initial mapping (model) filename (default null)

--nBest <int>

Generate n-best alignments (default n=1)

--inFormat <l2p|news>

Input file format [l2p, news] (default news)

--sepInChar <string>

Separated in-character used (default :)

--sepChar <string>

Separated character used (default |)

--nullChar <string>

Null character used (default _)

--pProcess <string>

Specify prefix output files

--pScore

Report score of each alignment (default false)

--cutoff <double>

Training threshold (default 0.01)

--maxFn <conXY|conYX|joint>

Maximization function [conXY, conYX, joint] (default conYX)

--eqMap

Allow mapping of |x| == |y| > 1 (default false)

--delY

Allow deletion of substring y (default false)

--delX

Allow deletion of substring x (default false)

--maxY <int>

Maximum length of substring y (default 2)

--maxX <int>

Maximum length of substring x (default 2)

--alignerIn <string>

Aligner model input filename

--alignerOut <string>

Aligner model output filename

-o <string>, --outputFile <string>

Output filename

-i <string>, --inputFile <string>

(required) Input filename

--, --ignore_rest

Ignores the rest of the labeled arguments following this flag.

--version

Displays version information and exits.

-h, --help

Displays usage information and exits.