Merge pull request #1643 from nouiz/tests

theano/sandbox/cuda/opt.py

@@ -1050,6 +1050,21 @@ def local_gpu_conv(node):
                 return make_graph
         return ret
 
+    def values_eq_approx(a, b):
+        """This fct is needed to don't have DebugMode raise useless
+        error due to ronding error.
+
+        This happen as We reduce on the two last dimensions, so this
+        can raise the absolute error if the number of element we
+        reduce on is significant.
+
+        """
+        assert a.ndim == 4
+        atol = None
+        if a.shape[-1] * a.shape[-2] > 100:
+            atol = 3e-5
+        return tensor.TensorType.values_eq_approx(a, b, atol=atol)
+
     if node.op == gpu_from_host:
         #gpu_from_host(conv) -> gpu_conv(gpu_from_host)
         host_input = node.inputs[0]
@@ -1058,12 +1073,14 @@ def local_gpu_conv(node):
             if gpu_conv is None:
                 return
             img, kern = host_input.owner.inputs
+            out = gpu_conv(gpu_from_host(img),
+                           gpu_from_host(kern))
+            out = tensor.patternbroadcast(out,
+                                          node.outputs[0].broadcastable)
+            out.values_eq_approx = values_eq_approx
             # in some case the ConvOp broadcast the last 2 dimensions
             # differently then the gpu ConvOp
-            return [tensor.patternbroadcast(
-                gpu_conv(gpu_from_host(img),
-                         gpu_from_host(kern)),
-                         node.outputs[0].broadcastable)]
+            return [out]
 
     if isinstance(node.op, conv.ConvOp):
         #conv(host_from_gpu) -> host_from_gpu(gpu_conv)
@@ -1074,12 +1091,15 @@ def local_gpu_conv(node):
             gpu_conv = GpuConvOp_from_ConvOp(node.op)
             if gpu_conv is None:
                 return
+            out = gpu_conv(gpu_from_host(img),
+                           gpu_from_host(kern))
+            out = tensor.patternbroadcast(
+                host_from_gpu(out),
+                node.outputs[0].broadcastable)
+            out.values_eq_approx = values_eq_approx
             # in some case the ConvOp broadcast the last 2 dimensions
             # differently then the gpu ConvOp
-            return [tensor.patternbroadcast(
-                host_from_gpu(gpu_conv(gpu_from_host(img),
-                                       gpu_from_host(kern))),
-                node.outputs[0].broadcastable)]
+            return [out]
 
 import theano.tensor.signal.downsample as downsample
 

theano/sandbox/cuda/tests/test_conv_cuda_ndarray.py

@@ -19,6 +19,7 @@ except ImportError:
 import theano
 from theano import tensor
 from theano.gof.python25 import any
+from theano.tests.unittest_tools import seed_rng
 
 # Skip test if cuda_ndarray is not available.
 import theano.sandbox.cuda as cuda_ndarray
@@ -415,6 +416,7 @@ def get_valid_shapes():
 
 
 def test_valid_0_2():
+    seed_rng()
     shapes = get_valid_shapes()
     version = [0, 2]
     verbose = 0
@@ -446,6 +448,7 @@ def test_valid_0_2():
 
 
 def test_valid_1_3_11_12():
+    seed_rng()
     shapes = get_valid_shapes()
     version = [1, 3, 11, 12]
     verbose = 0
@@ -475,6 +478,7 @@ def test_valid_1_3_11_12():
 
 
 def test_valid_4():
+    seed_rng()
     shapes = get_valid_shapes()
     version = [4]
     verbose = 0
@@ -506,6 +510,7 @@ def test_valid_4():
 
 
 def test_valid_5():
+    seed_rng()
     shapes = get_valid_shapes()
     version = [5]
     verbose = 0
@@ -537,6 +542,7 @@ def test_valid_5():
 
 
 def test_valid_7_8_13():
+    seed_rng()
     shapes = get_valid_shapes()
     # This is to test the "new" lower shared memory usage.
     shapes.append(((10, 30, 60, 60), (20, 30, 40, 40),
@@ -571,6 +577,7 @@ def test_valid_7_8_13():
 
 
 def test_valid_9_10():
+    seed_rng()
     shapes = get_valid_shapes()
     version = [9, 10]
     verbose = 0
@@ -601,6 +608,7 @@ def test_valid_9_10():
 
 
 def test_valid():
+    seed_rng()
     shapes = get_valid_shapes()
 
     #shapes=shapes[400:426]
@@ -621,6 +629,7 @@ def test_valid():
 
 
 def test_full():
+    seed_rng()
     shapes = get_basic_shapes()
     shapes += get_shapes2()
     #test image stride
@@ -682,6 +691,7 @@ def test_full():
 
 
 def test_subsample():
+    seed_rng()
     # implement when
     shapes = [((1, 1, 1, 1), (1, 1, 1, 1), (1, 1), (1, 1), (1, 1)),
               ((1, 1, 1, 1), (1, 1, 1, 1), (2, 2), (1, 1), (1, 1)),
@@ -712,6 +722,7 @@ def test_subsample():
 
 class TestConv2DGPU(unittest.TestCase):
     def test_logical_shapes(self):
+        seed_rng()
         for stride in range(1, 4):
             kshp = (10, 2, 10, 10)
             featshp = (3, 10, 11, 11)
@@ -749,6 +760,7 @@ class TestConv2DGPU(unittest.TestCase):
         Tests that when the shape gived at build time is not the same as
         run time we raise an error
         """
+        seed_rng()
         verbose = 0
         random = True
         print_ = False
@@ -860,6 +872,7 @@ def benchmark():
 
 
 def test_stack_rows_segfault_070312():
+    seed_rng()
     # 07/03/2012
     # Running this unittest with cuda-memcheck exposes an illegal read.
     # THEANO_FLAGS=device=gpu cuda-memcheck nosetests \