--- a/pylearn/algorithms/tests/test_daa.py	Thu Mar 18 11:33:49 2010 -0400
+++ b/pylearn/algorithms/tests/test_daa.py	Thu Mar 18 11:34:29 2010 -0400
@@ -6,12 +6,14 @@
 import time
 
 import pylearn.algorithms.logistic_regression
-from theano.compile.mode import default_mode
-
-def test_train_daa(mode = default_mode):
+from theano import config
+from pylearn.algorithms.stacker import Stacker
+from pylearn.algorithms.daa import SigmoidXEDenoisingAA
+from pylearn.algorithms.regressor import BinRegressor
+def test_train_daa(mode = config.mode):
 
     ndaa = 3
-    daa = models.Stacker([(models.SigmoidXEDenoisingAA, 'hidden')] * ndaa + [(models.BinRegressor, 'output')],
+    daa = Stacker([(SigmoidXEDenoisingAA, 'hidden')] * ndaa + [(BinRegressor, 'output')],
                          regularize = False)
 
     model = daa.make([4, 20, 20, 20, 1],
@@ -39,7 +41,7 @@
 def test_train_daa2(mode = theano.Mode('c|py', 'fast_run')):
 
     ndaa = 3
-    daa = models.Stacker([(models.SigmoidXEDenoisingAA, 'hidden')] * ndaa + [(pylearn.algorithms.logistic_regression.Module_Nclass, 'pred')],
+    daa = Stacker([(SigmoidXEDenoisingAA, 'hidden')] * ndaa + [(pylearn.algorithms.logistic_regression.Module_Nclass, 'pred')],
                          regularize = False)
 
     model = daa.make([4] + [20] * ndaa + [10],

--- a/pylearn/sandbox/test_scan_inputs_groups.py	Thu Mar 18 11:33:49 2010 -0400
+++ b/pylearn/sandbox/test_scan_inputs_groups.py	Thu Mar 18 11:34:29 2010 -0400
@@ -9,6 +9,7 @@
 import theano.tensor as T
 from pylearn.sandbox.scan_inputs_groups import FillMissing
 import theano.compile.mode as mode_module
+import theano
 
 class TestFillMissing(unittest.TestCase):
     def setUp(self):
@@ -16,9 +17,9 @@
 
         #we need to desactivate the check for NaN value as we have them in input
         #TODO: Make an option to don't check NaN value in input only, bug check in output.
-        m=mode_module.default_mode
-        if m=="DEBUG_MODE":
-            m=copy.copy(mode_module.predefined_modes[m])
+        m=mode_module.get_default_mode()
+        if isinstance(m,theano.compile.debugmode.DebugMode):
+            m=copy.copy(m)
             m.check_isfinite=False
         self.mode = m
 

--- a/pylearn/shared/layers/kording2004.py	Thu Mar 18 11:33:49 2010 -0400
+++ b/pylearn/shared/layers/kording2004.py	Thu Mar 18 11:34:29 2010 -0400
@@ -1,7 +1,6 @@
 import numpy
 import theano.tensor
-from hpu.theano_outgoing import mean, var, cov
-
+from theano.tensor.basic import mean
 from pylearn.shared.layers.exponential_mean import ExponentialMean # exponential_mean.py
 
 import logging

--- a/pylearn/shared/layers/tests/test_kouh2008.py	Thu Mar 18 11:33:49 2010 -0400
+++ b/pylearn/shared/layers/tests/test_kouh2008.py	Thu Mar 18 11:34:29 2010 -0400
@@ -9,9 +9,9 @@
     n_out = 10
     n_terms = 3
     rng = numpy.random.RandomState(23455)
-    layer = Kouh2008.new_filters(rng, tensor.dmatrix(), n_in, n_out, n_terms, dtype='float64')
+    layer = Kouh2008.new_filters_expbounds(rng, tensor.dmatrix(), n_in, n_out, n_terms, dtype='float64')
     assert layer.output.dtype =='float64'
-    layer = Kouh2008.new_filters(rng, tensor.fmatrix(), n_in, n_out, n_terms, dtype='float32')
+    layer = Kouh2008.new_filters_expbounds(rng, tensor.fmatrix(), n_in, n_out, n_terms, dtype='float32')
     assert layer.output.dtype =='float32'
 
 def run_w_random(bsize=10, n_iter=200, n_in = 1024, n_out = 100, n_terms=2, dtype='float64'):
@@ -19,7 +19,7 @@
     y = tensor.lvector()
     rng = numpy.random.RandomState(23455)
 
-    layer = Kouh2008.new_filters(rng, x, n_in, n_out, n_terms, dtype='float64')
+    layer = Kouh2008.new_filters_expbounds(rng, x, n_in, n_out, n_terms, dtype='float64')
     out = LogisticRegression.new(layer.output, n_out, 2)
     cost = out.nll(y).sum()
 
@@ -52,7 +52,7 @@
     y = tensor.lvector()
     rng = numpy.random.RandomState(23455)
 
-    layer = Kouh2008.new_filters(rng, x, n_in, n_out, n_terms, dtype='float64')
+    layer = Kouh2008.new_filters_expbounds(rng, x, n_in, n_out, n_terms, dtype='float64')
     out = LogisticRegression.new(layer.output, n_out, 2)
     cost = out.nll(y).sum()
     #joint optimization except for one of the linear filters