Merge pull request #5815 from nouiz/useless_reshape

theano/gpuarray/tests/test_opt.py

@@ -92,7 +92,8 @@ def test_flatten():
     assert GpuReshape in [type(node.op)
                           for node in f.maker.fgraph.toposort()]
 
-    f = theano.function([m], m.flatten(ndim=2), mode=mode_with_gpu)
+    f = theano.function([m], m.flatten(ndim=2),
+                        mode=mode_with_gpu.excluding("local_useless_reshape"))
     val = np.random.rand(10, 11).astype("float32")
     res = f(val)
     utt.assert_allclose(res, val)

theano/tensor/basic.py

@@ -4745,6 +4745,8 @@ class Reshape(Op):
 
     def __init__(self, ndim, name=None):
         self.ndim = ndim
+        if ndim < 0:
+            raise ValueError("The output dimensions after reshape must be 0 or greater")
         assert name is None, 'name attribute for Reshape has been deprecated'
 
     def __str__(self):

theano/tensor/opt.py

@@ -4380,6 +4380,7 @@ def local_useless_reshape(node):
         else:
             shape_feature = getattr(node.fgraph, 'shape_feature', None)
 
+        nb_m1 = 0
         shape_match = [False] * input.ndim
         for dim in xrange(input.ndim):
             outshp_i = output_shape_is[dim]
@@ -4403,11 +4404,17 @@ def local_useless_reshape(node):
                         continue
 
             # Match 1 if input.broadcastable[dim] is True
-            if (input.broadcastable[dim] and
-                    extract_constant(outshp_i, only_process_constants=1) == 1):
+            cst_outshp_i = extract_constant(outshp_i, only_process_constants=1)
+            if input.broadcastable[dim] and cst_outshp_i == 1:
                 shape_match[dim] = True
                 continue
 
+            # Match -1
+            if cst_outshp_i == -1:
+                shape_match[dim] = True
+                nb_m1 += 1
+                continue
+
             # Match shape_of[input][dim] or its constant equivalent
             if shape_feature:
                 inpshp_i = shape_feature.get_shape(input, dim)
@@ -4417,7 +4424,7 @@ def local_useless_reshape(node):
                     shape_match[dim] = True
                     continue
 
-        if all(shape_match):
+        if all(shape_match) and nb_m1 <= 1:
             return [input]
 
         # TODO later: if all the shapes except one match, we may want to

theano/tensor/tests/test_basic.py

@@ -7632,21 +7632,19 @@ class TestInferShape(utt.InferShapeTester):
         # Flatten
         atens3 = tensor3()
         atens3_val = rand(4, 5, 3)
-        self._compile_and_check([atens3],
-                                [flatten(atens3, 1)],
-                                [atens3_val], Reshape)
-
         for outdim in (3, 2, 1):
             self._compile_and_check([atens3],
                                     [flatten(atens3, outdim)],
-                                    [atens3_val], Reshape)
+                                    [atens3_val], Reshape,
+                                    excluding=['local_useless_reshape'])
 
         amat = matrix()
         amat_val = rand(4, 5)
         for outdim in (2, 1):
             self._compile_and_check([amat],
                                     [flatten(amat, outdim)],
-                                    [amat_val], Reshape)
+                                    [amat_val], Reshape,
+                                    excluding=['local_useless_reshape'])
 
         avec = vector()
         avec_val = rand(4)

theano/tensor/tests/test_opt.py

@@ -6475,6 +6475,21 @@ class Test_local_useless_reshape(unittest.TestCase):
         topo = f2.maker.fgraph.toposort()
         assert not any(isinstance(n.op, tensor.basic.Reshape) for n in topo)
 
+    def test_m1(self):
+            x = theano.tensor.matrix('x')
+            r = x.reshape((x.shape[0], -1))
+
+            m0 = theano.compile.get_default_mode()
+            m1 = m0.including('local_useless_reshape')
+            f1 = theano.function([x], r, mode=m1)
+            topo = f1.maker.fgraph.toposort()
+            assert not any(isinstance(n.op, tensor.basic.Reshape) for n in topo)
+
+            m2 = m1.excluding('ShapeOpt')
+            f2 = theano.function([x], r, mode=m2)
+            topo = f2.maker.fgraph.toposort()
+            assert not any(isinstance(n.op, tensor.basic.Reshape) for n in topo)
+
 
 class Test_local_reshape_to_dimshuffle(unittest.TestCase):
     def setUp(self):