# HG changeset patch
# User Yoshua Bengio <bengioy@iro.umontreal.ca>
# Date 1210354734 14400
# Node ID b4657441dd65d298bfd6f27c5af8a35fc2fadfbc
# Parent  f6505ec32dc3c57f1212b87997ccae3b8489fc6f
Corrected typos

diff -r f6505ec32dc3 -r b4657441dd65 learner.py
--- a/learner.py	Thu May 08 00:54:14 2008 -0400
+++ b/learner.py	Fri May 09 13:38:54 2008 -0400
@@ -47,7 +47,7 @@
         and return the learned function.
         """
         self.forget()
-        return self.update(learning_task,train_stats_collector)
+        return self.update(training_set,train_stats_collector)
 
     def use(self,input_dataset,output_fieldnames=None,
             test_stats_collector=None,copy_inputs=True,
@@ -254,7 +254,7 @@
         Private helper function that maps a list of attribute names to a list
         of corresponding Op Results (with the same name but with a '_' prefix).
         """
-        return [self.__getattribute__('_'+name).data for name in names]
+        return [self.__getattribute__('_'+name) for name in names]
 
 
 class MinibatchUpdatesTLearner(TLearner):
@@ -311,7 +311,8 @@
     def parameterAttributes(self):
         raise AbstractFunction()
 
-        def updateStart(self): pass
+    def updateStart(self,training_set):
+        pass
 
     def updateEnd(self):
         self.setAttributes(self.updateEndOutputAttributes(),
@@ -343,12 +344,15 @@
         """
         self.updateStart(training_set)
         stop=False
+        if hasattr(self,'_minibatch_size') and self._minibatch_size:
+            minibatch_size=self._minibatch_size
+        else:
+            minibatch_size=min(100,len(training_set))
         while not stop:
             if train_stats_collector:
                 train_stats_collector.forget() # restart stats collectin at the beginning of each epoch
-            for minibatch in training_set.minibatches(self.training_set_input_fields,
-                                                      minibatch_size=self.minibatch_size):
-                self.update_minibatch(minibatch)
+            for minibatch in training_set.minibatches(minibatch_size=minibatch_size):
+                self.updateMinibatch(minibatch)
                 if train_stats_collector:
                     minibatch_set = minibatch.examples()
                     minibatch_set.setAttributes(self.attributeNames(),self.attributes())
@@ -390,7 +394,7 @@
         return self.parameterAttributes()
     
     def updateMinibatchOutputAttributes(self):
-        return ["_new"+name for name in self.parameterAttributes()]
+        return ["new_"+name for name in self.parameterAttributes()]
     
     def updateEndInputAttributes(self):
         return self.parameterAttributes()
diff -r f6505ec32dc3 -r b4657441dd65 lookup_list.py
--- a/lookup_list.py	Thu May 08 00:54:14 2008 -0400
+++ b/lookup_list.py	Fri May 09 13:38:54 2008 -0400
@@ -95,10 +95,10 @@
     def __ne__(self, other):
         return not self.__eq__(other)
 
-    def __hash__():
+    def __hash__(self):
         raise NotImplementedError()
 
-    def __call__(*names):
+    def __call__(self,*names):
         """
         Return a list of values associated with the given names (which must all be keys of the lookup list).
         """
diff -r f6505ec32dc3 -r b4657441dd65 mlp.py
--- a/mlp.py	Thu May 08 00:54:14 2008 -0400
+++ b/mlp.py	Fri May 09 13:38:54 2008 -0400
@@ -8,6 +8,7 @@
 from learner import *
 from theano import tensor as t
 from nnet_ops import *
+import math
 
 
 class OneHiddenLayerNNetClassifier(OnlineGradientTLearner):
@@ -67,10 +68,13 @@
 
     """
 
-    def __init__(self,n_hidden,n_classes,learning_rate,init_range=1.):
+    def __init__(self,n_hidden,n_classes,learning_rate,max_n_epochs,init_range=1.,n_inputs=None,minibatch_size=None):
+        self._n_inputs = n_inputs
         self._n_outputs = n_classes
         self._n_hidden = n_hidden
         self._init_range = init_range
+        self._max_n_epochs = max_n_epochs
+        self._minibatch_size = minibatch_size
         self.learning_rate = learning_rate # this is the float
         self._learning_rate = t.scalar('learning_rate') # this is the symbol
         self._input = t.matrix('input') # n_examples x n_inputs
@@ -139,7 +143,11 @@
                                            size=(self._n_outputs,self._n_hidden))
             self.b1[:]=0
             self.b2[:]=0
+            self._n_epochs=0
 
+    def isLastEpoch(self):
+        self._n_epochs +=1
+        return self._n_epochs>=self._max_n_epochs
 
 class MLP(MinibatchUpdatesTLearner):
     """
diff -r f6505ec32dc3 -r b4657441dd65 test_mlp.py
--- a/test_mlp.py	Thu May 08 00:54:14 2008 -0400
+++ b/test_mlp.py	Fri May 09 13:38:54 2008 -0400
@@ -1,9 +1,17 @@
 
 from mlp import *
+import dataset
 
 def test0():
-    nnet = OneHiddenLayerNNetClassifier(10,3,.1)
+    nnet = OneHiddenLayerNNetClassifier(10,3,.1,1000)
+    training_set = dataset.ArrayDataSet(numpy.array([[0, 0, 0],
+                                                     [0, 1, 1],
+                                                     [1, 0, 1],
+                                                     [1, 1, 1]]),
+                                        {'input':slice(2),'target':2})
+    fprop=nnet(training_set)
 
+    print fprop(training_set)
 
 test0()