moses-smt/mosesdecoder
1
1
Fork 0
mirror of https://github.com/moses-smt/mosesdecoder.git synced 2024-12-30 15:34:01 +03:00
Code Issues Packages Projects Releases Wiki Activity
13d21428ed
mosesdecoder/contrib/promix
History
Barry Haddow 13d21428ed stub out README
2013-04-17 21:57:09 +01:00
..
test_data Initial checkin. Missing some test data 2013-02-21 17:59:53 +00:00
bleu.py Initial checkin. Missing some test data 2013-02-21 17:59:53 +00:00
coll.py Initial checkin. Missing some test data 2013-02-21 17:59:53 +00:00
main.py remove restriction to 2 tables 2013-03-22 10:35:19 +00:00
nbest.py Load all options. Use relative path. 2013-04-17 21:34:20 +01:00
README.md stub out README 2013-04-17 21:57:09 +01:00
sampler.py Initial checkin. Missing some test data 2013-02-21 17:59:53 +00:00
test_bleu.py Initial checkin. Missing some test data 2013-02-21 17:59:53 +00:00
test_main.py remove absolute path default 2013-04-17 21:35:32 +01:00
test_nbest.py Update for change of python interface 2013-04-15 07:25:31 +01:00
test_sampler.py Initial checkin. Missing some test data 2013-02-21 17:59:53 +00:00
test_train.py Initial checkin. Missing some test data 2013-02-21 17:59:53 +00:00
test.py Initial checkin. Missing some test data 2013-02-21 17:59:53 +00:00
train.py Remove debug, update python interface 2013-04-12 18:00:22 +01:00
util.py Initial checkin. Missing some test data 2013-02-21 17:59:53 +00:00

README.md

promix - for training translation model interpolation weights using PRO

Author: Barry Haddow <bhaddow [AT] inf.ed.ac.uk>

ABOUT

The code here provides the "inner loop" for a batch tuning algorithm (like MERT) which optimises phrase table interpolation weights at the same time as the standard linear model weights. Interpolation of the phrase tables uses the "naive" method of tmcombine.

Currently it only works on interpolations of two phrase tables.

REQUIREMENTS

The scripts require the Moses Python interface (in contrib/python) They also require scipy and numpy. They have been tested with the following versions: Python 2.7 Scipy 0.11.0 Numpy 1.6.2

USAGE

REFERENCES