Scroll to navigation

m2m-aligner(1) General Commands Manual m2m-aligner(1)


m2m-aligner - Parallel corpora aligner.


m2m-aligner [OPTIONS]



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.


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

Keep unaligned item in the output file (default false)

Cut-off sum prior probability (default 0.5)

Initial mapping (model) filename (default null)

Generate n-best alignments (default n=1)

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

Separated in-character used (default :)

Separated character used (default |)

Null character used (default _)

Specify prefix output files

Report score of each alignment (default false)

Training threshold (default 0.01)

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

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

Allow deletion of substring y (default false)

Allow deletion of substring x (default false)

Maximum length of substring y (default 2)

Maximum length of substring x (default 2)

Aligner model input filename

Aligner model output filename

Output filename

(required) Input filename

--, --ignore_rest
Ignores the rest of the labeled arguments following this flag.

Displays version information and exits.

Displays usage information and exits.

February, 2012