Add precision param to dnnConv (gpua backend)

theano/sandbox/gpuarray/dnn.py

@@ -740,7 +740,7 @@ class GpuDnnConvGradI(DnnBase):
 
 def dnn_conv(img, kerns, border_mode='valid', subsample=(1, 1),
              conv_mode='conv', direction_hint=None, workmem=None,
-             algo=None):
+             algo=None, precision=None):
     """
     GPU convolution using cuDNN from NVIDIA.
 
@@ -774,6 +774,10 @@ def dnn_conv(img, kerns, border_mode='valid', subsample=(1, 1),
         Convolution implementation to use. Some of its values may
         require certain versions of CuDNN to be installed. Default is
         the value of :attr:`config.dnn.conv.algo_fwd`.
+    precision : {'float16', 'float32', 'float64', 'floatX'}
+        Description of the dtype in which the computation of the convolution
+        should be done. Default is the value of
+        :attr:`config.dnn.conv.precision`.
 
     .. warning:: The cuDNN library only works with GPUs that have a compute
         capability of 3.0 or higer. This means that older GPUs will not
@@ -803,7 +807,7 @@ def dnn_conv(img, kerns, border_mode='valid', subsample=(1, 1),
             shape_i(kerns, 1, fgraph),
             shape_i(img, 1, fgraph), shape2, shape3)
         desc = GpuDnnConvDesc(border_mode='valid', subsample=(1, 1),
-                              conv_mode='cross')(out.shape)
+                              conv_mode='cross', precision=precision)(out.shape)
         conv = GpuDnnConvGradW()(img, kerns, out, desc)
         return as_gpuarray_variable(conv.dimshuffle(1, 0, 2, 3), ctx_name)
 
@@ -821,7 +825,7 @@ def dnn_conv(img, kerns, border_mode='valid', subsample=(1, 1),
                                                  shape_i(kerns, 1, fgraph),
                                                  shape2, shape3)
         desc = GpuDnnConvDesc(border_mode='valid', subsample=(1, 1),
-                              conv_mode=conv_mode)(kerns.shape)
+                              conv_mode=conv_mode, precision=precision)(kerns.shape)
         return GpuDnnConvGradI()(kerns, img, out, desc)
 
     # Standard case: We use GpuDnnConv with suitable padding.
@@ -830,7 +834,7 @@ def dnn_conv(img, kerns, border_mode='valid', subsample=(1, 1),
     img = gpu_contiguous(img)
     kerns = gpu_contiguous(kerns)
     desc = GpuDnnConvDesc(border_mode=border_mode, subsample=subsample,
-                          conv_mode=conv_mode)(kerns.shape)
+                          conv_mode=conv_mode, precision=precision)(kerns.shape)
     desc_op = desc.owner.op
     out_shp = GpuDnnConv.get_out_shape(img.shape, kerns.shape,
                                        desc_op.border_mode,