Merge pull request lisa-lab#13 from nouiz/master

Try to make travis pass.

Author: nouiz

.travis.yml

@@ -1,23 +1,47 @@
 # After changing this file, check it on:
 # http://lint.travis-ci.org/
 
-language: python
-python:
-  - "2.5"
+#We can't get scipy installed with the python language
+#So we will use the system python from the c language.
+language: c
+#language: python
+#python:
+#  - "2.5"
 #  - "2.7"
 #  - "3.2"
 # command to install dependencies
 before_install:
 #zlib1g-dev is needed to allow PIL to uncompress the dataset.
-  - sudo apt-get install -qq libatlas3gf-base libatlas-dev zlib1g-dev zip unzip zlibc libzip-dev libjpeg8 libjpeg62-dev libfreetype6 libfreetype6-dev
+  - sudo apt-get install -qq libatlas3gf-base libatlas-dev zlib1g-dev zip unzip zlibc libzip-dev libjpeg8 libjpeg62-dev libfreetype6 libfreetype6-dev python-numpy python-scipy python-pip python-nose python-yaml pyflakes python-imaging
 
 install:
-  - "pip install -q numpy --use-mirrors"
+#  - "pip install -q numpy --use-mirrors"
 # Use Pillow instead of PIL as it is better packaged
-  - "pip install -q Pillow --use-mirrors"
+#  - "pip install -q Pillow --use-mirrors"
 #If we don't install numpy before SciPy 0.10.1, the SciPy installations fails.
 #  - "pip install -q scipy --use-mirrors"
-  - "pip install --no-deps git+git://github.com/Theano/Theano.git"
+  - "sudo pip install --no-deps git+git://github.com/Theano/Theano.git"
+  - "sudo pip install hg+http://hg.assembla.com/pylearn"
+
+env:
+  - PART="test.py:test_logistic_sgd test.py:test_logistic_cg test.py:test_mlp"
+  - PART="test.py:test_convolutional_mlp test.py:test_dA"
+  - PART="test.py:test_SdA"
+  - PART="test.py:test_dbn"
+  - PART="test.py:test_rbm"
+  - PART="-e test.py"
+
+#569.882s   #10     code.test.test_rbm OK
+#298.992s   #9      code.test.test_dbn OK
+#268.901s   #8      code.test.test_SdA OK
+#67.292s    #7      code.test.test_dA OK
+#27.485s    #5      code.test.test_mlp OK
+#26.204s    #6      code.test.test_convolutional_mlp OK
+#14.676s    #4      code.test.test_logistic_cg OK
+#10.66s     #3      code.test.test_logistic_sgd OK
+#5.795s     #1      code.mcrbm.test_hmc.test_hmc OK
+#0.0s       #2      code.mcrbm.test_mcrbm.test_reproduce_ranzato_hinton_2010 FAILED TEST
+
 script:
   - cd data
   - ./download.sh
@@ -27,5 +51,5 @@ script:
   - ls
   - export THEANO_FLAGS=warn.ignore_bug_before=all,on_opt_error=raise,on_shape_error=raise
   - python --version
-  - nosetests test.py
+  - nosetests $PART
 

code/mcrbm/__init__.py

@@ -62,7 +62,7 @@ def __init__(self, rng, input, n_in, n_out, W=None, b=None,
 
         :type activation: theano.Op or function
         :param activation: Non linearity to be applied in the hidden
-                              layer
+                           layer
         """
         self.input = input
 
@@ -174,7 +174,7 @@ def __init__(self, rng, input, n_in, n_hidden, n_out):
 
 
 def test_mlp(learning_rate=0.01, L1_reg=0.00, L2_reg=0.0001, n_epochs=1000,
-             dataset='../data/mnist.pkl.gz', batch_size=20):
+             dataset='../data/mnist.pkl.gz', batch_size=20, n_hidden=500):
     """
     Demonstrate stochastic gradient descent optimization for a multilayer
     perceptron
@@ -219,15 +219,16 @@ def test_mlp(learning_rate=0.01, L1_reg=0.00, L2_reg=0.0001, n_epochs=1000,
     print '... building the model'
 
     # allocate symbolic variables for the data
-    index = T.lscalar()    # index to a [mini]batch
+    index = T.lscalar()  # index to a [mini]batch
     x = T.matrix('x')  # the data is presented as rasterized images
     y = T.ivector('y')  # the labels are presented as 1D vector of
                         # [int] labels
 
     rng = numpy.random.RandomState(1234)
 
     # construct the MLP class
-    classifier = MLP(rng=rng, input=x, n_in=28 * 28, n_hidden=500, n_out=10)
+    classifier = MLP(rng=rng, input=x, n_in=28 * 28,
+                     n_hidden=n_hidden, n_out=10)
 
     # the cost we minimize during training is the negative log likelihood of
     # the model plus the regularization terms (L1 and L2); cost is expressed
@@ -259,10 +260,10 @@ def test_mlp(learning_rate=0.01, L1_reg=0.00, L2_reg=0.0001, n_epochs=1000,
 
     # specify how to update the parameters of the model as a dictionary
     updates = {}
-    # given two list the zip A = [ a1,a2,a3,a4] and B = [b1,b2,b3,b4] of
+    # given two list the zip A = [a1, a2, a3, a4] and B = [b1, b2, b3, b4] of
     # same length, zip generates a list C of same size, where each element
     # is a pair formed from the two lists :
-    #    C = [ (a1,b1), (a2,b2), (a3,b3) , (a4,b4) ]
+    #    C = [(a1, b1), (a2, b2), (a3, b3), (a4, b4)]
     for param, gparam in zip(classifier.params, gparams):
         updates[param] = param - learning_rate * gparam