new released model for emnlp2013

QUICKSTART.md

@@ -53,8 +53,8 @@ Then run this command to test the default trained system on those two examples:
              @cacheserver=local \
              -Learner.maxTrainIters 0 \
              -Dataset.inPaths test:testinput \
-             -Builder.inParamsPath lib/models/2174.exec/params \
-             -Grammar.inPaths lib/models/2174.exec/grammar \
+             -Builder.inParamsPath lib/models/15.exec/params \
+             -Grammar.inPaths lib/models/15.exec/grammar \
              -Dataset.readLispTreeFormat true
 
 This run should take about a minute or two.  This will save the output to
@@ -67,7 +67,7 @@ Alternatively, you can launch an interactive shell to test out the system:
              @domain=webquestions \
              @sparqlserver=localhost:3093 \
              @cacheserver=local \
-             @load=2174 \
+             @load=15\
              @executeTopOnly=0
 
 # Training the System

release-code.files

@@ -18,6 +18,7 @@ Makefile
 README.md
 LICENSE.txt
 TUTORIAL.md
+QUICKSTART.md
 testng.xml
 download-dependencies
 

release-emnlp2013.files

@@ -32,7 +32,7 @@ lib/lucene/4.4/inexact
 
 ############################################################
 # Model
-lib/models/2174.exec
+lib/models/15.exec
 
 ############################################################
 # Grammars

sempre

@@ -233,6 +233,7 @@ def webquestions
     cachePaths('LexiconFn.cache', 'SparqlExecutor.cache'),
 
     o('EntityLexicon.inexactMatchIndex','lib/lucene/4.4/inexact/'),
+    o('EntityLexicon.maxEntries',50),
 
     # Learner reward
     o('Learner.partialReward', true),
@@ -278,7 +279,7 @@ def train
     header,
     'edu.stanford.nlp.sempre.Main',
     defaultOpts, selectDomain,
-    o('Learner.maxTrainIters', 6),
+    o('Learner.maxTrainIters', 4),
   nil)
 end
 

src/edu/stanford/nlp/sempre/LexiconFn.java

@@ -124,8 +124,6 @@ public class LexiconFn extends SemanticFn {
       features.addWithBias("basicStats", mode + ".popularity", Math.log(entry.getPopularity() + 1));
       //features.addWithBias(mode + ".distance", entry.getDistance());
     }
-    //if (FeatureExtractor.opts.features.contains("weikum")) 
-    //features.addWithBias(mode + ".distance", entry.getDistance());
 
     if (mode.equals("entity")) {
       // Entities