Merge pull request #2514 from sebastien-j/faster_tests_2

theano/scan_module/tests/test_scan.py

@@ -4,6 +4,7 @@ import sys
 from tempfile import mkdtemp
 import time
 import unittest
+import copy
 
 import cPickle
 import numpy
@@ -1909,20 +1910,25 @@ class T_Scan(unittest.TestCase):
         gparams = theano.grad(cost, params)
         updates = [(param, param - gparam * learning_rate)
                    for param, gparam in zip(params, gparams)]
+        mode = copy.copy(theano.compile.get_default_mode())
+        mode.check_py_code = False
         learn_rnn_fn = theano.function(inputs=[x, t],
                                        outputs=cost,
-                                       updates=updates)
+                                       updates=updates,
+                                       mode=mode)
         eval_rnn_fn = theano.function(inputs=[x],
-                                      outputs=y)
+                                      outputs=y,
+                                      mode=mode)
 
         # artificial data
-        x_v = numpy.arange(0., 100., 0.21, dtype=theano.config.floatX)
+        x_v = numpy.arange(0., 10.49, 0.21, dtype=theano.config.floatX)
         x_v = x_v.reshape(len(x_v), 1)
         s_v = numpy.sin(x_v)
         t_v = numpy.roll(s_v, -1)[:-1]
         s_v = s_v[:-1]
         for i in xrange(100):
             cost = learn_rnn_fn(s_v, t_v)
+            print i, cost
         pred = eval_rnn_fn(s_v)
         assert cost < 0.02