Commit Graph

275 Commits

Author SHA1 Message Date
ehasler
1da0544d77 enable 1best history for perceptron/analytical update
git-svn-id: http://svn.statmt.org/repository/mira@3894 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:37 +02:00
ehasler
86dba9f743 default: stop when weights converge
git-svn-id: http://svn.statmt.org/repository/mira@3892 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:37 +02:00
ehasler
a1d2252529 enable perceptron update with only 1 hope and fear translation each
git-svn-id: http://svn.statmt.org/repository/mira@3891 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:37 +02:00
ehasler
a58dd23b03 remove stop-optimal parameter, default for precision is 1.0, fix flexible slack, inable analytical update
git-svn-id: http://svn.statmt.org/repository/mira@3890 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:37 +02:00
ehasler
bc2a2078ea add 1best decoding after optimization for --updates-per-epoch option
git-svn-id: http://svn.statmt.org/repository/mira@3889 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
46cf77fa79 include precision option for constraint checking, fix some logs, ignore update when constraint change == 0 and error increases
git-svn-id: http://svn.statmt.org/repository/mira@3888 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
921735593d print out epoch for ignored updates
git-svn-id: http://svn.statmt.org/repository/mira@3887 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
ecfb8465d4 introduce parameter --nbest-first
git-svn-id: http://svn.statmt.org/repository/mira@3886 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
1f6f8b4abb update training script
git-svn-id: http://svn.statmt.org/repository/mira@3885 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
479ba8d160 fix bug in options --past-and-current-constraints, --accumulate-most-violated-constraints
git-svn-id: http://svn.statmt.org/repository/mira@3884 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
bea8af99ff refactor handling of accumulated constraints, constraint checking, introduce burn-in for bleu history, sentence-level bleu, bleu score weight
git-svn-id: http://svn.statmt.org/repository/mira@3883 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
1964eaf98a refactor handling of accumulated constraints, constraint checking, introduce burn-in for bleu history, sentence-level bleu, bleu score weight
git-svn-id: http://svn.statmt.org/repository/mira@3882 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
53cbedfd28 update hildreth code for linear classifier to work
git-svn-id: http://svn.statmt.org/repository/mira@3879 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
10abc4fd62 update weighted loss, adapt accumulate-most-violated-constraints and --past-and-current-constraints to batching, accumulate 1 constraint per sentence in batch, additional option for accumulating only 1 constraint per batch --one-per-batch
git-svn-id: http://svn.statmt.org/repository/mira@3878 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
bc7f8655d6 implement option --one-constraint that uses only the best model translation and the oracle and computes the update analytically
git-svn-id: http://svn.statmt.org/repository/mira@3877 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
e784b23eac fix case of mixing and weight dump frequency being 0
git-svn-id: http://svn.statmt.org/repository/mira@3876 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
1338ccb718 monitor changes in bleu feature history
git-svn-id: http://svn.statmt.org/repository/mira@3875 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
b23fcf8143 fix comparison of old vs. new model bleu for batches
git-svn-id: http://svn.statmt.org/repository/mira@3874 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
bhaddow
e664b4a4b3 Merge 3791-3842 from trunk
git-svn-id: http://svn.statmt.org/repository/mira@3873 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
ec6272b3fc bleu before and after opt. for batch, print error of constraints
git-svn-id: http://svn.statmt.org/repository/mira@3872 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
f29d58e3f3 bleu before and after opt. for batch, print error of constraints
git-svn-id: http://svn.statmt.org/repository/mira@3871 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
6b4a0b442b track the number of violated constraints and constraint changes
git-svn-id: http://svn.statmt.org/repository/mira@3870 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
16edd0b9bc track the number of violated constraints and constraint changes
git-svn-id: http://svn.statmt.org/repository/mira@3869 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
ac307d9e8c fix weight dumping for batching
git-svn-id: http://svn.statmt.org/repository/mira@3868 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
03c5aa5c36 sanity check for updates-per-epoch, stopping criterion 'sum of updates this epoch'
git-svn-id: http://svn.statmt.org/repository/mira@3867 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:36 +02:00
ehasler
864ade99a3 sanity check for updates-per-epoch, stopping criterion 'sum of updates this epoch'
git-svn-id: http://svn.statmt.org/repository/mira@3866 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
58eb0d0052 sanity check for updates-per-epoch, stopping criterion 'sum of updates this epoch'
git-svn-id: http://svn.statmt.org/repository/mira@3865 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
4187e0f691 rename parameter --average-weights
git-svn-id: http://svn.statmt.org/repository/mira@3864 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
0cbc5b06f4 forgot to check in .h file
git-svn-id: http://svn.statmt.org/repository/mira@3863 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
cefddaea88 change default values: --control-updates --> true, --accumulate-weights --> false, --distinct-nbest --> true
git-svn-id: http://svn.statmt.org/repository/mira@3862 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
9ee3edef82 add Hildreth unit tests, fix consstraint comparison in MiraOptimiser, some code clean-up
git-svn-id: http://svn.statmt.org/repository/mira@3861 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
0198fc5e41 add Hildreth unit tests, fix consstraint comparison in MiraOptimiser, some code clean-up
git-svn-id: http://svn.statmt.org/repository/mira@3860 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
ac8b86f0c8 show cache usage and weights
git-svn-id: http://svn.statmt.org/repository/mira@3859 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
56e97a851f change default value for averaging weights to false
git-svn-id: http://svn.statmt.org/repository/mira@3858 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
61a5d7f511 compare best model score and bleu before and after weight change
git-svn-id: http://svn.statmt.org/repository/mira@3857 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
2dfe7163b8 fix average moses weights
git-svn-id: http://svn.statmt.org/repository/mira@3856 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
3b760065ec update
git-svn-id: http://svn.statmt.org/repository/mira@3855 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
fa661939dd update training script
git-svn-id: http://svn.statmt.org/repository/mira@3853 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
269f1018c3 introduce option to use average weights for pruning
git-svn-id: http://svn.statmt.org/repository/mira@3852 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
f6483df41c average approx bleu of best model score translation
git-svn-id: http://svn.statmt.org/repository/mira@3851 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
62b5fb7f4d introduce parameter --update-after
git-svn-id: http://svn.statmt.org/repository/mira@3848 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
1630603dce code cleanup
git-svn-id: http://svn.statmt.org/repository/mira@3847 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
dbd2146188 introduce parameter --update-after-epoch
git-svn-id: http://svn.statmt.org/repository/mira@3846 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
d04066ee0e introduce parameter --update-after-epoch
git-svn-id: http://svn.statmt.org/repository/mira@3845 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
41d2d28408 update training script to compute BLEU of dev set, update computation of dev bleu (on best model score translation), check best model scores after weight update, new
git-svn-id: http://svn.statmt.org/repository/mira@3844 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
f95a9d5d26 update training script to compute BLEU of dev set, update computation of dev bleu (on best model score translation), check best model scores after weight update
git-svn-id: http://svn.statmt.org/repository/mira@3843 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:35 +02:00
ehasler
9f057c039f adapt BPfactor by 0.05, introduce params for minimum learning rate and minimum sentence update
git-svn-id: http://svn.statmt.org/repository/mira@3841 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
c41202f2ff fix case when learning rate/max sentence update is 0
git-svn-id: http://svn.statmt.org/repository/mira@3838 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
b7971689cb adapt BP factor
git-svn-id: http://svn.statmt.org/repository/mira@3836 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
8139f70df7 introduce parameter --decr-sentence-update
git-svn-id: http://svn.statmt.org/repository/mira@3835 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
6642a0437c fix reduce..
git-svn-id: http://svn.statmt.org/repository/mira@3834 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
ad1812f63a fix weight dumping
git-svn-id: http://svn.statmt.org/repository/mira@3833 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
7e181ffefd fix weight dumping
git-svn-id: http://svn.statmt.org/repository/mira@3832 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
8d95753d0c back to dumping cumulative weights, fix learning rate and max-sentence-update
git-svn-id: http://svn.statmt.org/repository/mira@3831 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
aae83df416 correct file names
git-svn-id: http://svn.statmt.org/repository/mira@3830 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
6f3506b46b rename parameter --mixing-frequency
git-svn-id: http://svn.statmt.org/repository/mira@3829 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
9eb2cdee12 rename dev set oracles
git-svn-id: http://svn.statmt.org/repository/mira@3828 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
377bd27b36 use total weights for convergence check, dump total weights after every epoch
git-svn-id: http://svn.statmt.org/repository/mira@3827 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
3906510611 closing bracket..
git-svn-id: http://svn.statmt.org/repository/mira@3826 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
fecfe64d52 Dump weights always after mixing, dump final weights after averaging cumulative weights, stopping criterion related to weights after every epoch
git-svn-id: http://svn.statmt.org/repository/mira@3825 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
8bfb25f703 change default for --accumulate-weights to true
git-svn-id: http://svn.statmt.org/repository/mira@3824 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
ffcd03dd51 change max-sentence-update to scale weights instead of clipping them, fix mpi commands
git-svn-id: http://svn.statmt.org/repository/mira@3823 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
505eba736e introduce parameters --stop-dev-bleu, --stop-approx-dev-bleu
git-svn-id: http://svn.statmt.org/repository/mira@3822 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
a26a38533b code cleanup
git-svn-id: http://svn.statmt.org/repository/mira@3821 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
512674dacb change default value for slack to 0.01
git-svn-id: http://svn.statmt.org/repository/mira@3820 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
75565119d2 introduce parameters --max-sentence-update, --check-updates, --decr-learning-rate, change some default values, code cleanup
git-svn-id: http://svn.statmt.org/repository/mira@3819 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
691d1bcb48 update training script not to do normalisation
git-svn-id: http://svn.statmt.org/repository/mira@3818 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:34 +02:00
ehasler
e81a7e6da1 optional normalisation at start
git-svn-id: http://svn.statmt.org/repository/mira@3817 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
a049734fd6 print oracles for rank0 when stopping
git-svn-id: http://svn.statmt.org/repository/mira@3816 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
ca7b28155c update of printing dev oracles, parameter for printing feature values
git-svn-id: http://svn.statmt.org/repository/mira@3815 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
b44f1447cd write oracle translation to file
git-svn-id: http://svn.statmt.org/repository/mira@3814 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
0722b78c0d introduce parameter --normalise
git-svn-id: http://svn.statmt.org/repository/mira@3813 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
9b3c9a6fd6 fix handling of large memory jobs
git-svn-id: http://svn.statmt.org/repository/mira@3811 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
eba945b61c print rank with BLEU
git-svn-id: http://svn.statmt.org/repository/mira@3810 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
eeba3547cd change default: --dev-bleu 1
git-svn-id: http://svn.statmt.org/repository/mira@3807 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
fbcc9e1401 update script to pass on decoder settings
git-svn-id: http://svn.statmt.org/repository/mira@3806 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
5312e8fc22 compute bleu on oracles translations of dev set, introduce params --min-weight-change, --msf-step, --msf-min, --slack-step, --slack-max, --decoder-settings
git-svn-id: http://svn.statmt.org/repository/mira@3804 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
c4acd339fd enable second language model
git-svn-id: http://svn.statmt.org/repository/mira@3803 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
c911b44ac3 comment printing of weights, change mpi quitting, several changes to training script
git-svn-id: http://svn.statmt.org/repository/mira@3800 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
391c6806a3 print information separately for each process ('rank 0: ...')
git-svn-id: http://svn.statmt.org/repository/mira@3799 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
606355deba Enable taking logs of feature values
git-svn-id: http://svn.statmt.org/repository/mira@3798 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
8309fa8dd7 update scripts training-expt.perl, extractInformation.perl, mapWeightsToIni.perl
git-svn-id: http://svn.statmt.org/repository/mira@3797 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:59:33 +02:00
ehasler
804033bd2c Revert to revision 3774 of public svn + add 0 to weight files < 10 and write 'stoppinDg' file
git-svn-id: http://svn.statmt.org/repository/mira@3795 cc96ff50-19ce-11e0-b349-13d7f0bd23df
2011-08-18 12:55:38 +02:00
evahasler
b8a4100ecd introduce parameters --log-feature-values and --base-of-log
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3785 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-16 20:01:36 +00:00
evahasler
64ccd4bb1a clipping, ignore certain updates
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3784 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-15 19:51:21 +00:00
evahasler
dd976678d3 clipping, ignore certain updates
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3783 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-15 19:48:45 +00:00
evahasler
3f8a00f0e4 revert clipping
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3780 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-15 12:58:32 +00:00
evahasler
6e7c6d530e revert clipping
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3779 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-15 12:54:19 +00:00
evahasler
aa2e988132 revert clipping
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3778 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-15 12:53:49 +00:00
evahasler
4024b5ef79 fix unknown word penalty parameter
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3776 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-14 18:27:25 +00:00
evahasler
4839fee8cb add parameter --ignore-weird-updates
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3774 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-14 15:31:20 +00:00
evahasler
54824d132f add parameter --ignore-u-weight
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3773 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-14 14:51:04 +00:00
evahasler
fbc8910102 reintroduce clipping
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3772 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-14 12:17:17 +00:00
evahasler
da95cb2560 compute error only for rank 0
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3771 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-13 18:10:17 +00:00
evahasler
2f1b959302 change bleu smoothing, change handling of multiple oracles, parameter for increasing BP
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3770 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-10 16:34:43 +00:00
evahasler
cd62ffc021 mpi abort, compute error over fixed set of ex., add params --suppress-convergence and --past-and-current-constraints
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3766 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-09 18:55:52 +00:00
evahasler
7c238b6b44 add parameters --accumulate-oracles and --accumulate-most-violated-constraints
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3765 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-09 11:49:52 +00:00
evahasler
0c70c1a4ef introduce weighted loss function
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3761 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-06 19:32:59 +00:00
evahasler
47d91e8b93 Introduce stopping criterion based on changes in average total weights
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3760 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-06 15:28:51 +00:00
evahasler
87141f9f89 debug code
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3758 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-02 18:27:21 +00:00
evahasler
b852fb6763 add scripts
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3757 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-02 10:22:50 +00:00
evahasler
5c96b2d07d test
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3756 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-02 10:22:12 +00:00
evahasler
fc64a41495 update training script
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3755 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-01 18:51:19 +00:00
evahasler
2ee9e26104 typo
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3754 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-01 18:21:58 +00:00
evahasler
cd12923603 typo
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3753 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-01 18:13:52 +00:00
evahasler
406f55584e introduce parameter --history-smoothing, add mpi_finalize statement
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3752 1f5c12ca-751b-0410-a591-d2e778427230
2010-12-01 18:09:49 +00:00
evahasler
da712f0eeb change return type from size_t to int
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3751 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-30 17:26:34 +00:00
evahasler
859483640f fix batching
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3747 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-29 19:09:34 +00:00
evahasler
628c2380a9 add parameter --regularise-hildreth-updates
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3746 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-29 18:42:18 +00:00
evahasler
e442c513dd sanity check number of violated constraint/overall change in objective
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3743 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-29 15:07:38 +00:00
evahasler
4af767af60 Introduce parameter --increase-BP
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3742 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-29 14:40:15 +00:00
evahasler
2feabd94fc Introduce parameter --distinct-nbest
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3741 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-29 14:13:45 +00:00
evahasler
360c03b3bf Introduce parameter --distinct-nbest
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3740 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-29 14:11:19 +00:00
evahasler
e7d3ddb358 disable mpi finalize (not working)
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3735 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-24 17:28:08 +00:00
evahasler
1de7aa4f81 enable batch optimisation, batch history updating, fix bug in Decoder.cpp (new: staticData.ReLoadBleuScoreFeatureParameter())
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3734 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-24 17:06:54 +00:00
evahasler
97b7e1401e delete fear translations
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3733 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-23 16:20:26 +00:00
evahasler
8cd6d2c34a delete fear translations
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3732 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-23 16:17:35 +00:00
evahasler
4a6027f7c7 introduce parameter --scale-by-input-length
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3731 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-23 15:59:36 +00:00
evahasler
8de8773b39 fix averaging of cumulative weights
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3730 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-23 15:11:32 +00:00
evahasler
904329798d introduce arguments --accumulate-weights and --use-scaled-reference
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3727 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-22 20:36:14 +00:00
evahasler
a25bbecd76 update smo
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3722 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-19 11:35:16 +00:00
evahasler
53b61425d5 add fixed clipping for smo
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3721 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-18 16:24:51 +00:00
evahasler
93ff481caa add parameter for size of nbest lists and for adding constraints
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3717 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-16 18:52:08 +00:00
evahasler
81088c8f30 Introduce margin scale factor, change printing of feature vectors
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3716 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-16 12:11:00 +00:00
evahasler
57b0e64cda Introduce margin scale factor, change printing of feature vectors
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3715 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-16 11:44:44 +00:00
evahasler
ffc37e64b3 normalize weights after mixing
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3707 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-12 15:40:53 +00:00
evahasler
590ad17713 normalize weights after mixing
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3706 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-12 15:35:10 +00:00
evahasler
7bc6318270 normalize weights after mixing
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3705 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-12 15:22:28 +00:00
evahasler
74cd65c24a update
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3693 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-09 17:37:10 +00:00
evahasler
7af072ace0 remove clipping from updateWeights(), update Main
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3692 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-09 17:31:30 +00:00
evahasler
9697b152be all to cerr
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3684 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-01 16:56:10 +00:00
evahasler
f17a6c6809 fix hildreth
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3683 1f5c12ca-751b-0410-a591-d2e778427230
2010-11-01 12:30:49 +00:00
evahasler
acc6ac21dc fix loop
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3681 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-29 17:36:16 +00:00
evahasler
08981b87f0 add Hildreth algorithm
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3680 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-29 15:42:50 +00:00
evahasler
8f00469259 add Hildreth algorithm
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3679 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-29 15:41:37 +00:00
bhaddow
179572c83a experiment script
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3678 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-28 16:20:57 +00:00
bhaddow
68d5f4469d weight dumping
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3677 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-28 12:41:33 +00:00
evahasler
17358fd834 sanity check for margin
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3676 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-28 10:00:35 +00:00
evahasler
144d750e33 Fix memory leak in Main.cpp
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3654 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-25 15:16:34 +00:00
evahasler
85a71793a6 Update mira optimatization code and merge Main.cpp
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3652 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-25 12:22:35 +00:00
evahasler
1bd263c4ec Return Bleu scores instead of mixed scores (to be used in loss function)
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3642 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-22 16:50:42 +00:00
bhaddow
7e72ceea22 Goodbye ScoreIndexManager.
Compiles ok, but haven't dared to run regression yet.


git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3608 1f5c12ca-751b-0410-a591-d2e778427230
2010-10-06 22:06:49 +00:00
bhaddow
a4b4ba15d7 Enable parallelisation
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3584 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-28 15:13:50 +00:00
hieuhoang1972
03419ac02a it runs\!\!
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3542 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-17 20:27:26 +00:00
starsplash
1e8d54a0d2 Add the mira criterion
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3541 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-17 17:52:33 +00:00
hieuhoang1972
c28168538a bugs
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3540 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-17 17:38:25 +00:00
hieuhoang1972
104c6a34fd bug
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3539 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-17 17:25:51 +00:00
starsplash
15810d7899 again my fault
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3538 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-17 17:07:03 +00:00
starsplash
b020c5945b sry for the typo
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3537 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-17 16:38:28 +00:00
bhaddow
44f14c776d history update
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/branches/mira-mtm5@3536 1f5c12ca-751b-0410-a591-d2e778427230
2010-09-17 16:34:48 +00:00