local_exmp1 function and test are added

theano/tensor/opt.py

@@ -3584,34 +3584,29 @@ def local_join_make_vector(node):
 @register_canonicalize
 @gof.local_optimizer([T.Elemwise])
 def local_expm1(node):
-    """ 
+    """
     This optimization detects exp(a)-1 and converts this to expm1(a).
     """
     if (isinstance(node.op, T.Elemwise) and
             isinstance(node.op.scalar_op, theano.scalar.basic.Sub)):
         in1, in2 = node.inputs
         out = node.outputs[0]
-        #import pdb; pdb.set_trace()
 
-        if (in1.owner and isinstance(in1.owner.op, T.Elemwise) and isinstance(in1.owner.op.scalar_op, theano.scalar.basic.Exp) and 
+        if (in1.owner and isinstance(in1.owner.op, T.Elemwise) and isinstance(in1.owner.op.scalar_op, theano.scalar.basic.Exp) and
                 T.extract_constant(in2) == 1):
             in11 = in1.owner.inputs[0]
             new_out = T.expm1(in11)
-            
-           # import pdb; pdb.set_trace()
+
             if new_out.dtype != out.dtype:
-                  new_out = T.cast(new_out, dtype=out.dtype)
-            # The ones could have forced a specific length
+                new_out = T.cast(new_out, dtype=out.dtype)
             if new_out.type != out.type:
-                new_out = broadcast_like(new_out, out, node.fgraph)
-
+                return
             return [new_out]
 
 
 ###############
 # Switch opts #
 ###############
-
 @register_canonicalize('fast_compile', 'local_remove_switch_const_cond')
 @register_specialize
 @gof.local_optimizer([T.Elemwise])

theano/tensor/tests/test_opt.py

@@ -6132,25 +6132,27 @@ def test_local_expm1():
     t = T.exp(x) - x
     s = T.exp(u) - numpy.ones((4, 3)).astype(config.floatX)
 
-    #f = function([x], y)
-    #g = function([x], z)
-    #h = function([x], t)
-    r = function([u],s)
+    f = function([x], y)
+    g = function([x], z)
+    h = function([x], t)
+    r = function([u], s)
     x_val = numpy.random.rand(4, 3).astype(config.floatX)
-    #f_val = f(x_val)
-    #g_val = g(x_val)
-    #h_val = h(x_val)
-    u_val = r(1.0)
-  
+    f_val = f(x_val)
+    f_test = function([x], T.expm1(x))
 
-    #assert any(isinstance(n.op, T.Elemwise) and isinstance(n.op.scalar_op, theano.scalar.basic.Expm1)
-     #          for n in f.maker.fgraph.toposort())
+    assert numpy.all(f_val == f_test(x_val))
 
-    #assert not any(isinstance(n.op, T.Elemwise) and isinstance(n.op.scalar_op, theano.scalar.basic.Expm1)
-     #              for n in g.maker.fgraph.toposort())
+    assert any(isinstance(n.op, T.Elemwise) and isinstance(n.op.scalar_op, theano.scalar.basic.Expm1)
+               for n in f.maker.fgraph.toposort())
 
-    #assert not any(isinstance(n.op, T.Elemwise) and isinstance(n.op.scalar_op, theano.scalar.basic.Expm1)
-     #              for n in h.maker.fgraph.toposort())
+    assert not any(isinstance(n.op, T.Elemwise) and isinstance(n.op.scalar_op, theano.scalar.basic.Expm1)
+                   for n in g.maker.fgraph.toposort())
+
+    assert not any(isinstance(n.op, T.Elemwise) and isinstance(n.op.scalar_op, theano.scalar.basic.Expm1)
+                   for n in h.maker.fgraph.toposort())
+
+    assert not any(isinstance(n.op, T.Elemwise) and isinstance(n.op.scalar_op, theano.scalar.basic.Expm1)
+                   for n in r.maker.fgraph.toposort())
 
 
 def test_local_merge_alloc():