Import theano.tensor.nnet.conv instead of theano.sandbox.conv (disable warning)

theano/sandbox/cuda/opt.py

@@ -337,7 +337,7 @@ def local_gpu_softmax(node):
     return False
 
 #### Convolution, maxpooling
-import theano.sandbox.conv
+from theano.tensor.nnet import conv
 @register_opt()
 @local_optimizer([])
 def local_gpu_conv(node):
@@ -367,12 +367,12 @@ def local_gpu_conv(node):
     if node.op == gpu_from_host:
         #gpu_from_host(conv) -> gpu_conv(gpu_from_host)
         host_input = node.inputs[0]
-        if host_input.owner and isinstance(host_input.owner.op, theano.sandbox.conv.ConvOp):
+        if host_input.owner and isinstance(host_input.owner.op, conv.ConvOp):
             gpu_conv = GpuConvOp_from_ConvOp(host_input.owner.op)
             img, kern = host_input.owner.inputs
             return [gpu_conv(gpu_from_host(img), gpu_from_host(kern))]
 
-    if isinstance(node.op, theano.sandbox.conv.ConvOp):
+    if isinstance(node.op, conv.ConvOp):
         #conv(host_from_gpu) -> host_from_gpu(gpu_conv)
         img, kern = node.inputs
         img_on_gpu = (img.owner and img.owner.op == host_from_gpu)

theano/sandbox/cuda/tests/test_nnet.py

@@ -6,7 +6,7 @@ from theano.compile.pfunc import pfunc
 from theano import tensor
 import theano.tensor.nnet
 
-import theano.sandbox.conv
+import theano.tensor.nnet.conv as conv
 import theano.tensor.signal.downsample as downsample
 
 import numpy
@@ -132,7 +132,7 @@ def run_conv_nnet1(use_gpu):
     y = tensor.fmatrix('y')
     lr = tensor.fscalar('lr')
 
-    conv_op = theano.sandbox.conv.ConvOp(shape_img[2:], shape_kern[2:], n_kern, n_batch, 1, 1)
+    conv_op = conv.ConvOp(shape_img[2:], shape_kern[2:], n_kern, n_batch, 1, 1)
     conv_op.set_flops()
 
     hid = tensor.tanh(conv_op(x, w)+b.dimshuffle((0,'x','x')))
@@ -215,8 +215,8 @@ def run_conv_nnet2(use_gpu): # pretend we are training LeNet for MNIST
     y = tensor.fmatrix('y')
     lr = tensor.fscalar('lr')
 
-    conv_op = theano.sandbox.conv.ConvOp(shape_img[2:], shape_kern[2:], n_kern, n_batch, 1, 1)
-    conv_op1 = theano.sandbox.conv.ConvOp((n_kern,logical_hid_shape[0]/2, logical_hid_shape[1]/2), shape_kern1[2:], n_kern1, n_batch, 1, 1)
+    conv_op = conv.ConvOp(shape_img[2:], shape_kern[2:], n_kern, n_batch, 1, 1)
+    conv_op1 = conv.ConvOp((n_kern,logical_hid_shape[0]/2, logical_hid_shape[1]/2), shape_kern1[2:], n_kern1, n_batch, 1, 1)
     conv_op.set_flops()
     conv_op1.set_flops()
 
@@ -299,9 +299,9 @@ def run_conv_nnet2_classif(use_gpu, isize, ksize, n_batch, n_iter,
     y = tensor.fmatrix('y')
     lr = tensor.fscalar('lr')
 
-    conv_op = theano.sandbox.conv.ConvOp(shape_img[2:], shape_kern[2:], n_kern,
-                                         n_batch, 1, 1, verbose=verbose, version=version)
-    conv_op1 = theano.sandbox.conv.ConvOp(
+    conv_op = conv.ConvOp(shape_img[2:], shape_kern[2:], n_kern,
+                          n_batch, 1, 1, verbose=verbose, version=version)
+    conv_op1 = conv.ConvOp(
         (n_kern,logical_hid_shape[0]/2, logical_hid_shape[1]/2),
         shape_kern1[2:], n_kern1, n_batch, 1, 1,verbose=verbose, version=version)
     conv_op.set_flops()