Added --return-best-dev to mert-moses.pl

Copies the run*.moses.ini with the highest development BLEU to
moses.ini, instead of copying the weights from the last run of
optimization.
Recommended when using kbmira (and probably pro).

1 files changed, 28 insertions, 3 deletions

diff --git a/scripts/training/mert-moses.pl b/scripts/training/mert-moses.pl
index 194910750..f6df313ea 100755
--- a/scripts/training/mert-moses.pl
+++ b/scripts/training/mert-moses.pl
@@ -108,6 +108,7 @@ my $___START_WITH_HISTORIC_BESTS = 0; # use best settings from all previous iter
 my $___RANDOM_DIRECTIONS = 0; # search in random directions only
 my $___NUM_RANDOM_DIRECTIONS = 0; # number of random directions, also works with default optimizer [Cer&al.,2008]
 my $___RANDOM_RESTARTS = 20;
+my $___RETURN_BEST_DEV = 0; # return the best weights according to dev, not the last
 
 # Flags related to PRO (Hopkins & May, 2011)
 my $___PAIRWISE_RANKED_OPTIMIZER = 0; # flag to enable PRO.
@@ -208,6 +209,7 @@ GetOptions(
   "random-directions" => \$___RANDOM_DIRECTIONS, # search only in random directions
   "number-of-random-directions=i" => \$___NUM_RANDOM_DIRECTIONS, # number of random directions
   "random-restarts=i" => \$___RANDOM_RESTARTS, # number of random restarts
+  "return-best-dev" => \$___RETURN_BEST_DEV, # return the best weights according to dev, not the last
   "activate-features=s" => \$___ACTIVATE_FEATURES, #comma-separated (or blank-separated) list of features to work on (others are fixed to the starting values)
   "range=s@" => \$___RANGES,
   "prev-aggregate-nbestlist=i" => \$prev_aggregate_nbl_size, #number of previous step to consider when loading data (default =-1, i.e. all previous)
@@ -295,6 +297,8 @@ Options:
                                      N means this and N previous iterations
 
   --maximum-iterations=ITERS ... Maximum number of iterations. Default: $maximum_iterations
+  --return-best-dev          ... Return the weights according to dev bleu, instead of returning
+                                 the last iteration
   --random-directions               ... search only in random directions
   --number-of-random-directions=int ... number of random directions
                                         (also works with regular optimizer, default: 0)
@@ -340,11 +344,13 @@ my $mert_extract_cmd = File::Spec->catfile($mertdir, "extractor");
 my $mert_mert_cmd    = File::Spec->catfile($mertdir, "mert");
 my $mert_pro_cmd     = File::Spec->catfile($mertdir, "pro");
 my $mert_mira_cmd    = File::Spec->catfile($mertdir, "kbmira");
+my $mert_eval_cmd    = File::Spec->catfile($mertdir, "evaluator");
 
 die "Not executable: $mert_extract_cmd" if ! -x $mert_extract_cmd;
 die "Not executable: $mert_mert_cmd"    if ! -x $mert_mert_cmd;
 die "Not executable: $mert_pro_cmd"     if ! -x $mert_pro_cmd;
 die "Not executable: $mert_mira_cmd"    if ! -x $mert_mira_cmd;
+die "Not executable: $mert_eval_cmd"    if ! -x $mert_eval_cmd;
 
 my $pro_optimizer = File::Spec->catfile($mertdir, "megam_i686.opt");  # or set to your installation
 
@@ -914,7 +920,6 @@ while (1) {
   print "loading data from $prev_score_file\n"   if defined($prev_score_file);
   print "loading data from $prev_init_file\n"    if defined($prev_init_file);
 }
-print "Training finished at " . `date`;
 
 if (defined $allsorted) {
     safesystem ("\\rm -f $allsorted") or die;
@@ -923,14 +928,34 @@ if (defined $allsorted) {
 safesystem("\\cp -f $weights_in_file run$run.$weights_in_file") or die;
 safesystem("\\cp -f $mert_logfile run$run.$mert_logfile") or die;
 
-create_config($___CONFIG_ORIG, "./moses.ini", $featlist, $run, $devbleu, $sparse_weights_file);
+if($___RETURN_BEST_DEV) {
+  my $bestit=1;
+  my $bestbleu=0;
+  my $evalout = "eval.out";
+  for (my $i = 1; $i < $run; $i++) {
+    safesystem("$mert_eval_cmd --reference " . join(",", @references) . " --candidate run$i.out 2> /dev/null 1> $evalout");
+    open my $fh, '<', $evalout or die "Can't read $evalout : $!";
+    my $bleu = <$fh>;
+    chomp $bleu;
+    if($bleu > $bestbleu) {
+      $bestbleu = $bleu;
+      $bestit = $i;
+    }
+    close $fh;
+  }
+  print "copying weights from best iteration ($bestit, bleu=$bestbleu) to moses.ini\n";
+  safesystem("\\cp -f run$bestit.moses.ini moses.ini") or die;
+}
+else {
+  create_config($___CONFIG_ORIG, "./moses.ini", $featlist, $run, $devbleu, $sparse_weights_file);
+}
 
 # just to be sure that we have the really last finished step marked
 &save_finished_step($finished_step_file, $run);
 
 #chdir back to the original directory # useless, just to remind we were not there
 chdir($cwd);
-
+print "Training finished at " . `date`;
 } # end of local scope
 
 sub get_weights_from_mert {