Merge pull request #2524 from nouiz/tests

doc/install_ubuntu.txt

@@ -5,14 +5,14 @@ Easy Installation of an Optimized Theano on Current Ubuntu
 
 For Ubuntu 11.10 through 14.04:
 
-.. code-block: bash
+.. code-block:: bash
 
     sudo apt-get install python-numpy python-scipy python-dev python-pip python-nose g++ libopenblas-dev git
     sudo pip install Theano
 
 For Ubuntu 11.04:
 
-.. code-block: bash
+.. code-block:: bash
 
     sudo apt-get install python-numpy python-scipy python-dev python-pip python-nose g++ git libatlas3gf-base libatlas-dev
     sudo pip install Theano
@@ -71,7 +71,7 @@ If you would like, instead, to install the bleeding edge Theano (from github)
 such that you can edit and contribute to Theano, replace the `pip install Theano` 
 command with:
 
-.. code-block: bash
+.. code-block:: bash
 
     git clone git://github.com/Theano/Theano.git
     cd Theano 
@@ -85,7 +85,7 @@ If you would like to install Theano in a VirtualEnv, you will want to pass the
 `--system-site-packages` flag when creating the VirtualEnv so that it will pick up 
 the system-provided `Numpy` and `SciPy`.
 
-.. code-block: bash
+.. code-block:: bash
     
     virtualenv --system-site-packages -p python2.7 theano-env
     source theano-env/bin/activate
@@ -166,7 +166,7 @@ Updating Bleeding Edge Installs
 
 Change to the Theano directory and run:
 
-.. code-block: bash
+.. code-block:: bash
     
     git pull
 

theano/sandbox/cuda/cuda_ndarray.cu

@@ -3065,7 +3065,8 @@ CudaNdarray_ptr_int_size(PyObject* _unused, PyObject* args)
         PyErr_SetString(PyExc_RuntimeError, "error copying data to from memory");
         return NULL;
     }
-    return Py_BuildValue("iiii", gpu_sizes[0], sizeof(float*), sizeof(int), gpu_sizes[1]);
+    return Py_BuildValue("iiii", (int) gpu_sizes[0], (int)sizeof(float*),
+                         (int)sizeof(int), (int) gpu_sizes[1]);
 }
 
 static int cublas_init();

theano/sandbox/cuda/opt.py

@@ -313,8 +313,8 @@ def local_gpu_split(node):
         if host_input.owner and isinstance(host_input.owner.op, tensor.Split):
             split_node = host_input.owner
             new_op = GpuSplit(split_node.op.len_splits)
-            return [new_op(gpu_from_host(split_node.inputs[0]),
-                           *split_node.inputs[1:])]
+            return new_op(gpu_from_host(split_node.inputs[0]),
+                          *split_node.inputs[1:])
     return False
 
 
@@ -1742,7 +1742,7 @@ def get_device_type_sizes():
         cuda_ndarray = theano.sandbox.cuda.cuda_ndarray.cuda_ndarray
         t = cuda_ndarray.ptr_int_size()
         gpu_ptr_size, cpu_ptr_size, int_size, gpu_int_size = t
-        assert int_size == gpu_int_size
+        assert int_size == gpu_int_size, (int_size, gpu_int_size)
         del gpu_int_size
         del t
     except Exception, e:
@@ -1751,7 +1751,9 @@ def get_device_type_sizes():
             "This could cause less GpuElemwise fused together.\n"
             "%s") % e)
 
-    rval = get_device_type_sizes.rval = locals()
+    rval = get_device_type_sizes.rval = dict(gpu_ptr_size=gpu_ptr_size,
+                                             cpu_ptr_size=cpu_ptr_size,
+                                             int_size=int_size)
     return rval
 
 

theano/sandbox/cuda/tests/test_fftconv.py

 import unittest
 import numpy
-import copy
 
 import theano
 from theano.tests import unittest_tools as utt
@@ -86,7 +85,6 @@ class TestConv2dFFT(unittest.TestCase):
         assert sum(isinstance(n.op, theano.sandbox.cuda.fftconv.CuFFTOp)
                    for n in topo) == 2, topo
 
-
         res_ref = f_ref()
         res_fft = f_fft()
 
@@ -130,11 +128,11 @@ class TestConv2dFFT(unittest.TestCase):
         inputs = shared(inputs_val)
         filters = shared(filters_val)
 
-        conv = theano.tensor.nnet.conv.conv2d(inputs, filters, version='no_fft')
+        conv = theano.tensor.nnet.conv.conv2d(inputs, filters,
+                                              version='no_fft')
 
         mode = mode_with_gpu.including('conv_fft_valid')
 
-        f_ref = theano.function([], conv)
         f_fft = theano.function([], conv, mode=mode)
 
         # make sure we that no CuFFTOp has been inserted
@@ -158,7 +156,6 @@ class TestConv2dFFT(unittest.TestCase):
 
         mode = mode_with_gpu.including('conv_fft_full')
 
-        f_ref = theano.function([], conv)
         f_fft = theano.function([], conv, mode=mode)
 
         # make sure we that no CuFFTOp has been inserted
@@ -178,15 +175,16 @@ class TestConv3dFFT(unittest.TestCase):
         bias = shared(numpy.zeros(filters_shape[0]).astype('float32'))
 
         # Flip filter as conv3D compute correlation
-        filters_flip = filters[:,::-1,::-1,::-1,:]
-        #filters_flip = filters
+        filters_flip = filters[:, ::-1, ::-1, ::-1, :]
+        # filters_flip = filters
         conv_ref = theano.tensor.nnet.conv3D(V=inputs, W=filters_flip,
-                                             b=bias, d=(1,1,1))
+                                             b=bias, d=(1, 1, 1))
 
-        conv_fft = theano.sandbox.cuda.fftconv.conv3d_fft(inputs.dimshuffle(0, 4, 1, 2, 3),
-                                                          filters.dimshuffle(0, 4, 1, 2, 3),
-                                                          border_mode = "valid",
-                                                          pad_last_dim = pad)
+        conv_fft = theano.sandbox.cuda.fftconv.conv3d_fft(
+            inputs.dimshuffle(0, 4, 1, 2, 3),
+            filters.dimshuffle(0, 4, 1, 2, 3),
+            border_mode="valid",
+            pad_last_dim=pad)
         conv_fft = conv_fft.dimshuffle(0, 2, 3, 4, 1)
 
         f_ref = theano.function([], conv_ref, mode="FAST_RUN")
@@ -198,8 +196,6 @@ class TestConv3dFFT(unittest.TestCase):
         res_fft = f_fft()
         utt.assert_allclose(res_ref, res_fft,  rtol=1e-05, atol=1e-05)
 
-
-
     def run_conv_full(self, inputs_shape, filters_shape, pad=False):
         inputs_val = numpy.random.random(inputs_shape).astype('float32')
         filters_val = numpy.random.random(filters_shape).astype('float32')
@@ -208,14 +204,15 @@ class TestConv3dFFT(unittest.TestCase):
         filters = shared(filters_val)
         bias = shared(numpy.zeros(filters_shape[4]).astype('float32'))
 
-        conv_ref = theano.tensor.nnet.convTransp3D(W=filters, b=bias, d=(1,1,1),
-                                                   H=inputs)
+        conv_ref = theano.tensor.nnet.convTransp3D(
+            W=filters, b=bias, d=(1, 1, 1),
+            H=inputs)
 
         filters = filters.dimshuffle(4, 0, 1, 2, 3)
         inputs = inputs.dimshuffle(0, 4, 1, 2, 3)
         conv_fft = theano.sandbox.cuda.fftconv.conv3d_fft(inputs, filters,
-                                                          border_mode = "full",
-                                                          pad_last_dim = pad)
+                                                          border_mode="full",
+                                                          pad_last_dim=pad)
         conv_fft = conv_fft.dimshuffle(0, 2, 3, 4, 1)
 
         f_ref = theano.function([], conv_ref)
@@ -233,6 +230,7 @@ class TestConv3dFFT(unittest.TestCase):
         self.run_conv_valid(inputs_shape=(16, 20, 32, 15, 1),
                             filters_shape=(10, 6, 12, 4, 1),
                             pad=True)
+
     def test_full(self):
         self.run_conv_full(inputs_shape=(16, 15, 21, 16, 10),
                            filters_shape=(10, 6, 12, 4, 1),
@@ -253,7 +251,7 @@ class TestConv3dFFT(unittest.TestCase):
         bias = shared(numpy.zeros(filters_shape[0]).astype('float32'))
 
         conv = theano.tensor.nnet.conv3D(V=inputs, W=filters,
-                                         b=bias, d=(1,1,1))
+                                         b=bias, d=(1, 1, 1))
         mode = mode_with_gpu.including('conv3d_fft')
         mode.check_py_code = False
 
@@ -265,16 +263,19 @@ class TestConv3dFFT(unittest.TestCase):
         assert sum(isinstance(n.op, theano.sandbox.cuda.fftconv.CuFFTOp)
                    for n in topo) == 2
 
-
         res_ref = f_ref()
         res_fft = f_fft()
 
         utt.assert_allclose(res_ref, res_fft)
 
     def test_opt_convgrad3d_fft(self):
-        inputs_shape = (16, 20, 32, 16, 1)
-        filters_shape = (10, 6, 12, 4, 1)
-        dCdH_shape = (16, 15, 21, 13, 10)
+        inputs_shape = (2, 17, 15, 16, 1)
+        filters_shape = (10, 6, 7, 4, 1)
+        dCdH_shape = (inputs_shape[0],
+                      inputs_shape[1] - filters_shape[1] + 1,
+                      inputs_shape[2] - filters_shape[2] + 1,
+                      inputs_shape[3] - filters_shape[3] + 1,
+                      filters_shape[0])
 
         inputs_val = numpy.random.random(inputs_shape).astype('float32')
         dCdH_val = numpy.random.random(dCdH_shape).astype('float32')
@@ -284,7 +285,7 @@ class TestConv3dFFT(unittest.TestCase):
 
         conv = theano.tensor.nnet.convGrad3D(V=inputs, dCdH=dCdH,
                                              WShape=filters_shape,
-                                             d=(1,1,1))
+                                             d=(1, 1, 1))
         mode = mode_with_gpu.including('convgrad3d_fft')
         mode.check_py_code = False
 
@@ -296,16 +297,14 @@ class TestConv3dFFT(unittest.TestCase):
         assert sum(isinstance(n.op, theano.sandbox.cuda.fftconv.CuFFTOp)
                    for n in topo) == 2
 
-
         res_ref = f_ref()
         res_fft = f_fft()
 
         utt.assert_allclose(res_ref, res_fft,  rtol=1e-04, atol=1e-04)
 
-
     def test_opt_convtransp3d_fft(self):
-        inputs_shape = (16, 15, 21, 12, 10)
-        filters_shape = (10, 6, 12, 4, 1)
+        inputs_shape = (2, 9, 16, 12, 10)
+        filters_shape = (10, 3, 8, 4, 1)
 
         inputs_val = numpy.random.random(inputs_shape).astype('float32')
         filters_val = numpy.random.random(filters_shape).astype('float32')
@@ -314,7 +313,7 @@ class TestConv3dFFT(unittest.TestCase):
         inputs = shared(inputs_val)
         filters = shared(filters_val)
 
-        conv = theano.tensor.nnet.convTransp3D(W=filters, b=bias, d=(1,1,1),
+        conv = theano.tensor.nnet.convTransp3D(W=filters, b=bias, d=(1, 1, 1),
                                                H=inputs)
         mode = mode_with_gpu.including('convtransp3d_fft')
 
@@ -326,9 +325,7 @@ class TestConv3dFFT(unittest.TestCase):
         assert sum(isinstance(n.op, theano.sandbox.cuda.fftconv.CuFFTOp)
                    for n in topo) == 2
 
-
         res_ref = f_ref()
         res_fft = f_fft()
 
         utt.assert_allclose(res_ref, res_fft, rtol=1e-04, atol=1e-04)
-