use the new DebugMode machanism to disable error when an stability optimization…

theano/tensor/opt.py

@@ -1196,10 +1196,12 @@ def local_mul_switch_sink(node):
             if isinstance(switch.inputs[1],Constant) and get_constant_value(switch.inputs[1]) == 0.:
                 listmul = node.inputs[:idx] + node.inputs[idx+1:]
                 fct = [T.switch(switch.inputs[0],0,T.mul(*(listmul + [switch.inputs[2]])))]
+                fct[0].values_eq_approx = fct[0].type.values_eq_approx_remove_nan
                 return fct
             if isinstance(switch.inputs[2],Constant) and get_constant_value(switch.inputs[2]) == 0.:
                 listmul = node.inputs[:idx] + node.inputs[idx+1:]
                 fct = [T.switch(switch.inputs[0],T.mul(*(listmul + [switch.inputs[1]])),0)]
+                fct[0].values_eq_approx = fct[0].type.values_eq_approx_remove_nan
                 return fct
     return False
 
@@ -1223,9 +1225,11 @@ def local_div_switch_sink(node):
         switch = node.inputs[0].owner
         if isinstance(switch.inputs[1],Constant) and get_constant_value(switch.inputs[1]) == 0.:
             fct = [T.switch(switch.inputs[0],0,op(switch.inputs[2],node.inputs[1]))]
+            fct[0].values_eq_approx = fct[0].type.values_eq_approx_remove_nan
             return fct
         if isinstance(switch.inputs[2],Constant) and get_constant_value(switch.inputs[2]) == 0.:
             fct = [T.switch(switch.inputs[0],op(switch.inputs[1],node.inputs[1]),0)]
+            fct[0].values_eq_approx = fct[0].type.values_eq_approx_remove_nan
             return fct
     return False
 

theano/tensor/tests/test_opt.py

@@ -1612,13 +1612,17 @@ class T_local_switch_sink(unittest.TestCase):
                     [[numpy.ones((4,)),numpy.zeros((4,)),numpy.ones((4,)),numpy.zeros((4,))]] + \
                     [[numpy.asarray(1.0),numpy.asarray(0.0),numpy.asarray(1.0),numpy.asarray(0.0)]]
 
+        self.mode = theano.compile.mode.get_default_mode().including('canonicalize','fast_run').excluding('gpu','fusion')
+        self.mode = copy.copy(self.mode)
+        self.mode.check_isfinite = False
+
     def test_local_mul_switch_sink(self):
         c = T.dscalar()
         idx = 0
         for condition in [(T.dmatrix('cond'),self.condm),(T.dvector('cond'),self.condv),(T.dscalar('cond'),self.conds)]:
             for x in [(T.dmatrix('x'),self.xm),(T.dvector('x'),self.xv),(T.dscalar('x'),self.xs)]:
                 y = T.mul(T.switch(condition[0]>0,1.*x[0],0.*x[0]),T.switch(condition[0]>0,1.*x[0],T.log(c)*x[0]))
-                f = theano.function([condition[0],x[0],c],[y], mode='FAST_RUN')
+                f = theano.function([condition[0],x[0],c],[y], mode=self.mode)
                 if type(condition[1]) is list:
                     for i in range(len(condition[1])):
                         res= f(condition[1][i],x[1],-1)
@@ -1634,7 +1638,7 @@ class T_local_switch_sink(unittest.TestCase):
         for condition in [(T.dmatrix('cond'),self.condm),(T.dvector('cond'),self.condv),(T.dscalar('cond'),self.conds)]:
             for x in [(T.dmatrix('x'),self.xm),(T.dvector('x'),self.xv),(T.dscalar('x'),self.xs)]:
                 y = T.true_div(T.switch(condition[0]>0,1.*x[0],0.*x[0]),T.switch(condition[0]>0,1.*x[0],T.log(c)*x[0]))
-                f = theano.function([condition[0],x[0],c],[y], mode='FAST_RUN')
+                f = theano.function([condition[0],x[0],c],[y], mode=self.mode)
                 if type(condition[1]) is list:
                     for i in range(len(condition[1])):
                         res= f(condition[1][i],x[1],-1)