From 1b0d40959f529b67db3b9d10dbf93101e0c65c7c Mon Sep 17 00:00:00 2001
From: Patrick Simianer <p@simianer.de>
Date: Mon, 27 Jan 2014 10:40:14 +0100
Subject: verbose parameter for mira (thanks Felix!)

---
 training/mira/kbest_cut_mira.cc | 9 ++++++---
 training/mira/mira.py           | 4 ++++
 2 files changed, 10 insertions(+), 3 deletions(-)

(limited to 'training/mira')

diff --git a/training/mira/kbest_cut_mira.cc b/training/mira/kbest_cut_mira.cc
index 990609d7..9415909e 100644
--- a/training/mira/kbest_cut_mira.cc
+++ b/training/mira/kbest_cut_mira.cc
@@ -95,7 +95,8 @@ bool InitCommandLine(int argc, char** argv, po::variables_map* conf) {
     ("stream,t", "Stream mode (used for realtime)")
     ("weights_output,O",po::value<string>(),"Directory to write weights to")
     ("output_dir,D",po::value<string>(),"Directory to place output in")
-    ("decoder_config,c",po::value<string>(),"Decoder configuration file");
+    ("decoder_config,c",po::value<string>(),"Decoder configuration file")
+	("verbose,v",po::value<bool>()->zero_tokens(),"verbose stderr output");
   po::options_description clo("Command line options");
   clo.add_options()
     ("config", po::value<string>(), "Configuration file")
@@ -627,6 +628,7 @@ int main(int argc, char** argv) {
   
   vector<string> corpus;
 
+  const bool VERBOSE = conf.count("verbose");
   const string metric_name = conf["mt_metric"].as<string>();
   optimizer = conf["optimizer"].as<int>();
   fear_select = conf["fear"].as<int>();
@@ -790,7 +792,8 @@ int main(int argc, char** argv) {
 	  double margin = cur_bad.features.dot(dense_weights) - cur_good.features.dot(dense_weights);
 	  double mt_loss = (cur_good.mt_metric - cur_bad.mt_metric);
 	  const double loss = margin +  mt_loss;
-	  cerr << "LOSS: " << loss << " Margin:" << margin << " BLEUL:" << mt_loss << " " << cur_bad.features.dot(dense_weights) << " " << cur_good.features.dot(dense_weights) <<endl;
+	  cerr << "LOSS: " << loss << " Margin:" << margin << " BLEUL:" << mt_loss << endl;
+	  if (VERBOSE) cerr << cur_bad.features.dot(dense_weights) << " " << cur_good.features.dot(dense_weights) << endl;
 	  if (loss > 0.0 || !checkloss) {
 	    SparseVector<double> diff = cur_good.features;
 	    diff -= cur_bad.features;	    
@@ -928,7 +931,7 @@ int main(int argc, char** argv) {
 
 			lambdas += (cur_pair[1]->features) * step_size;
 			lambdas -= (cur_pair[0]->features) * step_size;
-			cerr << " Lambdas " << lambdas << endl;
+			if (VERBOSE) cerr << " Lambdas " << lambdas << endl;
 			//reload weights based on update
 
 			dense_weights.clear();
diff --git a/training/mira/mira.py b/training/mira/mira.py
index d5a1d9f8..1555cbb4 100755
--- a/training/mira/mira.py
+++ b/training/mira/mira.py
@@ -143,6 +143,8 @@ def main():
   parser.add_argument('--pass-suffix', 
                       help='multipass decoding iteration. see documentation '
                            'at www.cdec-decoder.org for more information')
+  parser.add_argument('-v', '--verbose',
+                      help='more verbose mira optimizers')
   args = parser.parse_args()
 
   args.metric = args.metric.upper()
@@ -352,6 +354,8 @@ def optimize(args, script_dir, dev_size):
       decoder_cmd += ' -a'
     if not args.no_pseudo:
       decoder_cmd += ' -e'
+    if args.verbose:
+      decoder_cmd += ' -v'
     
     #always use fork 
     parallel_cmd = '{0} --use-fork -e {1} -j {2} --'.format(
-- 
cgit v1.2.3


From 7201ef7d46e39f42923d89688a1cb82c757abe61 Mon Sep 17 00:00:00 2001
From: Patrick Simianer <p@simianer.de>
Date: Tue, 22 Jul 2014 11:40:59 +0200
Subject: mira qsub

---
 training/mira/mira.py | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

(limited to 'training/mira')

diff --git a/training/mira/mira.py b/training/mira/mira.py
index 691a62a6..ec9c2d64 100755
--- a/training/mira/mira.py
+++ b/training/mira/mira.py
@@ -143,6 +143,10 @@ def main():
   parser.add_argument('--pass-suffix', 
                       help='multipass decoding iteration. see documentation '
                            'at www.cdec-decoder.org for more information')
+  parser.add_argument('--qsub',
+                      help='use qsub', action='store_true')
+  parser.add_argument('--pmem',
+                      help='memory for qsub', type=str, default='5G')
   parser.add_argument('-v', '--verbose',
                       help='more verbose mira optimizers')
   args = parser.parse_args()
@@ -317,6 +321,8 @@ def split_devset(dev, outdir):
 
 def optimize(args, script_dir, dev_size):
   parallelize = script_dir+'/../utils/parallelize.pl'
+  if args.qsub:
+    parallelize += " -p %s"%args.pmem
   decoder = script_dir+'/kbest_cut_mira'
   (source, refs) = split_devset(args.devset, args.output_dir)
   port = random.randint(15000,50000)
@@ -358,9 +364,12 @@ def optimize(args, script_dir, dev_size):
     if args.verbose:
       decoder_cmd += ' -v'
     
-    #always use fork 
-    parallel_cmd = '{0} --use-fork -e {1} -j {2} --'.format(
-                    parallelize, logdir, args.jobs)
+    if args.qsub:
+      parallel_cmd = '{0} -e {1} -j {2} --'.format(
+                      parallelize, logdir, args.jobs)
+    else:
+      parallel_cmd = '{0} --use-fork -e {1} -j {2} --'.format(
+                      parallelize, logdir, args.jobs)
     
     cmd = parallel_cmd + ' ' + decoder_cmd
     logging.info('OPTIMIZATION COMMAND: {}'.format(cmd))
-- 
cgit v1.2.3