added a parameter to compute the python version of ConvOp or not.

theano/sandbox/test_conv.py

@@ -63,7 +63,6 @@ class TestConvOp(unittest.TestCase):
             # fixed parameters
             bsize = 7     # batch size
             imshp = (5,4)# image shape
-            print >> sys.stderr, "WARNING: only square shape tested"
             kshps = [(2,3)]
             nkern = 6      # nb kernel
             ssizes = [(1,1)] #step size
@@ -72,7 +71,6 @@ class TestConvOp(unittest.TestCase):
             # fixed parameters
             bsize = 7     # batch size
             imshp = (5,4)# image shape
-            print >> sys.stderr, "WARNING: only square shape tested"
             kshps = [(2,3)]
             nkern = 6      # nb kernel
             ssizes = [(1,1)] #step size
@@ -112,7 +110,7 @@ class TestConvOp(unittest.TestCase):
 
                     # now test with real values
                     img2d = 1 + N.arange(bsize*N.prod(imshp)).reshape((bsize,)+imshp)
-                    print 'img2d', img2d
+#                    print 'img2d', img2d
                     img1d = img2d.reshape(bsize,-1)
 
                     # create filters (need to be flipped to use convolve2d)
@@ -121,12 +119,12 @@ class TestConvOp(unittest.TestCase):
                     # compute with new convolve2 (no timing info)
                     output4, outshp4  = convolve2(kerns, kshp, nkern, input,\
                             imshp, bsize, (1,1), bias=bias, mode=conv_mode)
-                    print 'output4', output4
+#                    print 'output4', output4
 
                     ttime1 = time.time()
                     f = function([kerns, bias, input], output4)
                     out4 = f(filtersflipped.reshape(nkern,-1), biasvals, img1d)
-                    print 'out4', out4, img1d, filtersflipped
+#                    print 'out4', out4, img1d, filtersflipped
                     tconv2 += [time.time() - ttime1]
                     out4 = out4.reshape(bsize, nkern, outshp4[1], outshp4[2])
                     out4 = out4[:,:,0::ss[0],0::ss[1]]
@@ -228,7 +226,7 @@ class TestConvOp(unittest.TestCase):
         kerns4=dmatrix4()
         assert len(kshps)==len(nkerns)==len(kerns)
 
-        def do_test(conv_mode, ss, unroll_batch=0, unroll_kern=0, img=img, validate=True):
+        def do_test(conv_mode, ss, unroll_batch=0, unroll_kern=0, img=img, validate=True,conv_op_py=False):
             
                 # build actual input images
                 imgval = rng.rand(bsize, imshp_start[0], imshp_start[1], imshp_start[2])
@@ -309,17 +307,21 @@ class TestConvOp(unittest.TestCase):
                     hidval2 = hidval2_[:,:,0::ss[0],0::ss[1]]
                     tctot += time.time() - time1
 
+                    if conv_op_py:
                         time1 = time.time()
-#                    hidval3_ = propup3(imgval,w_flip)
-#                    hidval3 = hidval3_[:,:,0::ss[0],0::ss[1]]
+                        hidval3_ = propup3(imgval,w_flip)
+                        hidval3 = hidval3_[:,:,0::ss[0],0::ss[1]]
                         tpytot += time.time() - time1
-#                    assert (N.abs(hidval2-hidval3)<1e-5).all()
+                        assert (N.abs(hidval2-hidval3)<1e-5).all()
+                    else:
+                        tpytot += 0
 
                     if validate:
                         temp = N.abs(outval - hidval2)
                         assert (temp < 1e-5).all()
-#                    temp = N.abs(outval - hidval3)
-#                    assert (temp < 1e-5).all()
+                    if validate and conv_op_py:
+                        temp = N.abs(outval - hidval3)
+                        assert (temp < 1e-5).all()
 
                     img, imshp = hid, tuple(outshp)
                     imgval = outval.reshape(bsize,outshp[0],outshp[1],outshp[2])
@@ -332,8 +334,8 @@ class TestConvOp(unittest.TestCase):
             
             validate=False# we don't validate the result to have it much faster!
 
-            unroll_batch = [0,1,2,5,10,20]
-            unroll_kern = [0,1,2,5,10,20]
+            unroll_batch = [0,1,2,4,5,10,20]
+            unroll_kern = [0,2,4,5,10,20]
 #            unroll_batch = [0,2,5]
 #            unroll_kern = [0,2,5]
             
@@ -344,6 +346,10 @@ class TestConvOp(unittest.TestCase):
             ssizes = [(1,1),]#(1,1)]#(2,2) bugged
             convmodes = ['valid','full']
             do_theano=False
+            a=T.dmatrix()
+            kerns = [a for i in nkerns]
+
+            assert len(kshps)==len(nkerns)==len(kerns)
         
             timing = N.zeros((len(unroll_batch),len(unroll_kern),3))
             t_b_k=[]
@@ -353,7 +359,7 @@ class TestConvOp(unittest.TestCase):
                     tctot, tpytot, ntot=[],[],[]
                     for conv_mode, n_mode in zip(convmodes,range(len(convmodes))):
                         for ss, n_ss in zip(ssizes,range(len(ssizes))):
-                            tctot_, tpytot_, ntot_ = do_test(conv_mode, ss,unroll_batch=unroll_b, unroll_kern=unroll_k, validate=validate)
+                            tctot_, tpytot_, ntot_ = do_test(conv_mode, ss, unroll_batch=unroll_b, unroll_kern=unroll_k, validate=validate)
                             tctot+=[tctot_]
                             tpytot+=[tpytot_]
                             ntot+=[ntot_]
@@ -446,3 +452,10 @@ class TestConvOp(unittest.TestCase):
                     kernvals = kernvals.reshape(nkern,-1)
 
                     utt.verify_grad(testf, [imgvals, kernvals])
+
+if __name__ == '__main__':
+    t = TestConvOp("test_convolution")
+    t.test_convolution()
+#    t.test_multilayer_conv()
+#    from theano.tests import main
+#    main("test_sp")