Add documenation on how to use GpuDnnConv.

doc/library/tensor/nnet/conv.txt

@@ -94,6 +94,13 @@ TODO: Give examples on how to use these things! They are pretty complicated.
 
           f = theano.function(..., mode=mode)
 
+    - :func:`GpuDnnConv <theano.sandbox.cuda.dnn.GpuDnnConv>` GPU-only
+      convolution using NVIDIA's cuDNN library.  To enable it (and
+      other cudnn-acclerated ops), set
+      ``THEANO_FLAGS=optimizer_including=cudnn`` in your environment.
+      This requires that you have cuDNN installed and available.  It
+      also requires a GPU with compute capability 3.0 or more.
+
     - :func:`conv3D <theano.tensor.nnet.Conv3D.conv3D>`
       3D Convolution applying multi-channel 3D filters to batches of
       multi-channel 3D images.

theano/sandbox/cuda/__init__.py

@@ -263,8 +263,8 @@ if cuda_available:
 
     shared_constructor = float32_shared_constructor
 
-    import basic_ops
-    from basic_ops import (
+    from . import basic_ops
+    from .basic_ops import (
             GpuFromHost, HostFromGpu, GpuElemwise,
             GpuDimShuffle, GpuCAReduce, GpuReshape, GpuContiguous,
             GpuSubtensor, GpuIncSubtensor,
@@ -274,11 +274,11 @@ if cuda_available:
             ftensor3, ftensor4,
             scalar, vector, matrix, row, col,
             tensor3, tensor4)
-    from basic_ops import (host_from_gpu, gpu_from_host,
+    from .basic_ops import (host_from_gpu, gpu_from_host,
             as_cuda_array, as_cuda_ndarray_variable)
-    import opt
     import cuda_ndarray
-    from rng_curand import CURAND_RandomStreams
+    from . import opt, dnn
+    from .rng_curand import CURAND_RandomStreams
 
 
 def use(device,