fixed bug in debug mode

theano/scan.py

@@ -1469,7 +1469,7 @@ class ScanGrad(Op):
                     [i.type() for i in outputs_grad ])
 
     def perform(self, node, args, storage):
-
+        print 'perform args'
         # get scan inputs
         n_steps = args[0]
 
@@ -1488,7 +1488,12 @@ class ScanGrad(Op):
         seqs = inputs[:self.n_seqs]
         outInfo = inputs[self.n_seqs:self.n_seqs+self.n_outs]
         non_seqs = inputs[self.n_outs+self.n_seqs:]
-
+        print 'seqs'
+        print seqs
+        print 'outInfo'
+        print outInfo
+        print 'non_Seqs'
+        print non_seqs
         if (self.n_seqs == 0 ) and (not numpy.isfinite(n_steps) ):
             raise ValueError('Scan does not know how many steps it '
                 'should iterate! Either provide some input sequences from '
@@ -1546,10 +1551,16 @@ class ScanGrad(Op):
 
         g_non_seqs = [numpy.zeros_like(k) for k in non_seqs]
         # get gradient on the outputs 
-        g_outs = args[1:self.n_outs_not_shared+1]
+        g_outs = [arg.copy() for arg in args[1:self.n_outs_not_shared+1]]
 
         # get the output of the scan operation
         outs = args[1+self.n_outs_not_shared:self.n_outs_not_shared+self.n_outs+1]
+        print 'g_outs'
+        print g_outs
+        print 'outs'
+        print outs
+        print 'steps'
+        print args[0]
 
 
 
@@ -1623,8 +1634,6 @@ class ScanGrad(Op):
                   min_tap =      seqs_mins[j]
                   for tap_value in ls_taps :
                       k = _i - min_tap + tap_value
-                      print k, lower_limit, k-lower_limit
-                      print g_seqs[j].shape
                       g_seqs[j][k-lower_limit] += grads[pos]
                       pos += 1
 
@@ -1651,7 +1660,12 @@ class ScanGrad(Op):
 
 
         # return the gradient
-
+        print 'g_seqs'
+        print g_seqs
+        print 'g_outInfo'
+        print g_outInfo
+        print 'g_non_seqs'
+        print g_non_seqs
         for i,v in enumerate(g_seqs + g_outInfo+ g_non_seqs):
             storage[i][0] = v
 

theano/tests/test_scan.py

@@ -894,7 +894,7 @@ class T_Scan(unittest.TestCase):
             return x_t
 
         cost, updates = scan_project_sum(f_rnn_cmpl,u,x0,W_in, n_steps = None,
-                truncate_gradient = 40, go_backwards = False)
+                truncate_gradient = 3, go_backwards = False)
         vparams = [v_u, v_x0,vW_in]
         params = [u,x0,W_in ]
         gparams = theano.tensor.grad(cost, params)
@@ -919,7 +919,7 @@ class T_Scan(unittest.TestCase):
         analytic_grad = reset_rng_grad_fn(v_u, v_x0, vW_in)
 
 
-        assert len(analytic_grad[0]) == 40
+        assert len(analytic_grad[0]) == 3