# HG changeset patch
# User James Bergstra <bergstrj@iro.umontreal.ca>
# Date 1225213069 14400
# Node ID a419edf4e06c2d9e7275c24af18c0b6fd1b004aa
# Parent  2be795cc5c3abeeee172fc78a7e5cf437ba4dc74
removed unpicklable nested classes in logistic regression

diff -r 2be795cc5c3a -r a419edf4e06c algorithms/logistic_regression.py
--- a/algorithms/logistic_regression.py	Tue Oct 28 12:25:15 2008 -0400
+++ b/algorithms/logistic_regression.py	Tue Oct 28 12:57:49 2008 -0400
@@ -7,21 +7,22 @@
 
 import numpy as N
 
+class LogRegInstanceType(module.FancyModuleInstance):
+    def initialize(self, n_in, n_out=1, rng=N.random):
+        #self.component is the LogisticRegressionTemplate instance that built this guy.
+
+        self.w = N.zeros((n_in, n_out))
+        self.b = N.zeros(n_out)
+        self.lr = 0.01
+        self.__hide__ = ['params']
 
 class Module_Nclass(module.FancyModule):
-    class InstanceType(module.FancyModuleInstance):
-        def initialize(self, n_in, n_out, rng=N.random):
-            #self.component is the LogisticRegressionTemplate instance that built this guy.
+    InstanceType = LogRegInstanceType
 
-            self.w = N.zeros((n_in, n_out))
-            self.b = N.zeros(n_out)
-            self.lr = 0.01
-            self.__hide__ = ['params']
-
-    def __init__(self, input=None, targ=None, w=None, b=None, lr=None, regularize=False):
+    def __init__(self, x=None, targ=None, w=None, b=None, lr=None, regularize=False):
         super(Module_Nclass, self).__init__() #boilerplate
 
-        self.input = input if input is not None else T.matrix('input')
+        self.x = x if x is not None else T.matrix('input')
         self.targ = targ if targ is not None else T.lvector()
 
         self.w = w if w is not None else module.Member(T.dmatrix())
@@ -31,12 +32,16 @@
         self.params = [p for p in [self.w, self.b] if p.owner is None]
 
         xent, output = nnet.crossentropy_softmax_1hot(
-                T.dot(self.input, self.w) + self.b, self.targ)
+                T.dot(self.x, self.w) + self.b, self.targ)
         sum_xent = T.sum(xent)
 
         self.output = output
         self.sum_xent = sum_xent
+
+        #compatibility with current implementation of stacker/daa or something
+        #TODO: remove this, make a wrapper
         self.cost = sum_xent
+        self.input = self.x
 
         #define the apply method
         self.pred = T.argmax(T.dot(self.input, self.w) + self.b, axis=1)
@@ -49,14 +54,7 @@
                     updates = dict((p, p - self.lr * g) for p, g in zip(self.params, gparams)))
 
 class Module(module.FancyModule):
-    class InstanceType(module.FancyModuleInstance):
-        def initialize(self, n_in):
-            #self.component is the LogisticRegressionTemplate instance that built this guy.
-
-            self.w = N.random.randn(n_in,1)
-            self.b = N.random.randn(1)
-            self.lr = 0.01
-            self.__hide__ = ['params']
+    InstanceType = LogRegInstanceType
 
     def __init__(self, input=None, targ=None, w=None, b=None, lr=None, regularize=False):
         super(Module, self).__init__() #boilerplate
@@ -89,8 +87,6 @@
             self.update = module.Method([self.input, self.targ], sum_xent,
                                         updates = dict((p, p - self.lr * g) for p, g in zip(self.params, gparams)))
 
-
-
 class Learner(object):
     """TODO: Encapsulate the algorithm for finding an optimal regularization coefficient"""
     pass