Remove the if indentation.

theano/sandbox/gpuarray/dnn.py

@@ -1487,11 +1487,9 @@ err%(name)s = cudnnSoftmaxBackward(
         """
 
 
-# Intentation for history
-if False:
-    # @register_opt('cudnn')  # this optimizer is registered in opt.py instead.
-    @local_optimizer([GpuConv])
-    def local_conv_dnn(node):
+# @register_opt('cudnn')  # this optimizer is registered in opt.py instead.
+@local_optimizer([GpuConv])
+def local_conv_dnn(node):
     if not dnn_available():
         return
     if isinstance(node.op, GpuConv):
@@ -1509,11 +1507,11 @@ if False:
                 rval, node.outputs[0].type.broadcastable)
         return [rval]
 
-    # This optimizer is registered in opt.py as part of the meta-optimizer.
-    # It tries exactly the opposite code path of what local_conv_dnn() uses,
-    # because for some input/kernel shape configurations, this is faster.
-    @local_optimizer([GpuConv])
-    def local_conv_dnn_alternative(node):
+# This optimizer is registered in opt.py as part of the meta-optimizer.
+# It tries exactly the opposite code path of what local_conv_dnn() uses,
+# because for some input/kernel shape configurations, this is faster.
+@local_optimizer([GpuConv])
+def local_conv_dnn_alternative(node):
     if not dnn_available():
         return
     if isinstance(node.op, GpuConv):
@@ -1542,8 +1540,8 @@ if False:
                 rval, node.outputs[0].type.broadcastable)
         return [rval]
 
-    @local_optimizer([GpuDnnConv], inplace=True)
-    def local_dnn_conv_inplace(node):
+@local_optimizer([GpuDnnConv], inplace=True)
+def local_dnn_conv_inplace(node):
     if type(node.op) != GpuDnnConv or node.op.inplace:
         return
     inputs = list(node.inputs)
@@ -1554,8 +1552,8 @@ if False:
         inputs[2] = GpuAllocEmpty(dest.owner.op.dtype)(*dest.owner.inputs)
     return [GpuDnnConv(workmem=node.op.workmem, inplace=True)(*inputs)]
 
-    @local_optimizer([GpuDnnConvGradW], inplace=True)
-    def local_dnn_convgw_inplace(node):
+@local_optimizer([GpuDnnConvGradW], inplace=True)
+def local_dnn_convgw_inplace(node):
     if type(node.op) != GpuDnnConvGradW or node.op.inplace:
         return
     inputs = list(node.inputs)
@@ -1566,8 +1564,8 @@ if False:
         inputs[2] = GpuAllocEmpty(dest.owner.op.dtype)(*dest.owner.inputs)
     return [GpuDnnConvGradW(inplace=True)(*inputs)]
 
-    @local_optimizer([GpuDnnConvGradI], inplace=True)
-    def local_dnn_convgi_inplace(node):
+@local_optimizer([GpuDnnConvGradI], inplace=True)
+def local_dnn_convgi_inplace(node):
     if type(node.op) != GpuDnnConvGradI or node.op.inplace:
         return
     inputs = list(node.inputs)
@@ -1578,55 +1576,55 @@ if False:
         inputs[2] = GpuAllocEmpty(dest.owner.op.dtype)(*dest.owner.inputs)
     return [GpuDnnConvGradI(inplace=True)(*inputs)]
 
-    optdb.register('local_dnn_conv_inplace',
+optdb.register('local_dnn_conv_inplace',
                tensor.opt.in2out(local_dnn_conv_inplace,
                                  local_dnn_convgw_inplace,
                                  local_dnn_convgi_inplace,
                                  name="local_dnn_conv_inplace"),
                70.0, 'fast_run', 'inplace', 'gpu', 'cudnn')
 
-    @register_opt('cudnn')
-    @alpha_merge(GpuDnnConv, alpha_in=4, beta_in=5, nd=4)
-    def local_dnn_conv_alpha_merge(node, *inputs):
+@register_opt('cudnn')
+@alpha_merge(GpuDnnConv, alpha_in=4, beta_in=5, nd=4)
+def local_dnn_conv_alpha_merge(node, *inputs):
     if not dnn_available() or version() == -1:
         return None
     return [GpuDnnConv(workmem=node.op.workmem)(*inputs)]
 
-    @register_opt('cudnn')
-    @alpha_merge(GpuDnnConvGradW, alpha_in=4, beta_in=5, nd=4)
-    def local_dnn_convw_alpha_merge(node, *inputs):
+@register_opt('cudnn')
+@alpha_merge(GpuDnnConvGradW, alpha_in=4, beta_in=5, nd=4)
+def local_dnn_convw_alpha_merge(node, *inputs):
     if not dnn_available() or version() == -1:
         return None
     return [GpuDnnConvGradW()(*inputs)]
 
-    @register_opt('cudnn')
-    @alpha_merge(GpuDnnConvGradI, alpha_in=4, beta_in=5, nd=4)
-    def local_dnn_convi_alpha_merge(node, *inputs):
+@register_opt('cudnn')
+@alpha_merge(GpuDnnConvGradI, alpha_in=4, beta_in=5, nd=4)
+def local_dnn_convi_alpha_merge(node, *inputs):
     if not dnn_available() or version() == -1:
         return None
     return [GpuDnnConvGradI()(*inputs)]
 
-    @register_opt('cudnn')
-    @output_merge(GpuDnnConv, alpha_in=4, beta_in=5, out_in=2, nd=4)
-    def local_dnn_conv_output_merge(node, *inputs):
+@register_opt('cudnn')
+@output_merge(GpuDnnConv, alpha_in=4, beta_in=5, out_in=2, nd=4)
+def local_dnn_conv_output_merge(node, *inputs):
     inputs = inputs[0:2] + (gpu_contiguous(inputs[2]),) + inputs[3:]
     return [GpuDnnConv(workmem=node.op.workmem)(*inputs)]
 
-    @register_opt('cudnn')
-    @output_merge(GpuDnnConvGradW, alpha_in=4, beta_in=5, out_in=2, nd=4)
-    def local_dnn_convw_output_merge(node, *inputs):
+@register_opt('cudnn')
+@output_merge(GpuDnnConvGradW, alpha_in=4, beta_in=5, out_in=2, nd=4)
+def local_dnn_convw_output_merge(node, *inputs):
     inputs = inputs[0:2] + (gpu_contiguous(inputs[2]),) + inputs[3:]
     return [GpuDnnConvGradW()(*inputs)]
 
-    @register_opt('cudnn')
-    @output_merge(GpuDnnConvGradI, alpha_in=4, beta_in=5, out_in=2, nd=4)
-    def local_dnn_convi_output_merge(node, *inputs):
+@register_opt('cudnn')
+@output_merge(GpuDnnConvGradI, alpha_in=4, beta_in=5, out_in=2, nd=4)
+def local_dnn_convi_output_merge(node, *inputs):
     inputs = inputs[0:2] + (gpu_contiguous(inputs[2]),) + inputs[3:]
     return [GpuDnnConvGradI()(*inputs)]
 
-    @register_opt('cudnn')
-    @local_optimizer([GpuDownsampleFactorMax])
-    def local_pool_dnn(node):
+@register_opt('cudnn')
+@local_optimizer([GpuDownsampleFactorMax])
+def local_pool_dnn(node):
     if not dnn_available():
         return
     if isinstance(node.op, GpuDownsampleFactorMax):
@@ -1636,9 +1634,9 @@ if False:
         ds = node.op.ds
         return [dnn_pool(gpu_contiguous(img), ds, ds)]
 
-    @register_opt('cudnn')
-    @local_optimizer([DownsampleFactorMax])
-    def local_pool_dnn_alternative(node):
+@register_opt('cudnn')
+@local_optimizer([DownsampleFactorMax])
+def local_pool_dnn_alternative(node):
     if not dnn_available():
         return
     if isinstance(node.op, DownsampleFactorMax):
@@ -1654,9 +1652,9 @@ if False:
                            ds, stride=stride, pad=pad, mode=mode)
             return [host_from_gpu(ret)]
 
-    @register_opt('cudnn')
-    @local_optimizer([GpuDownsampleFactorMaxGrad])
-    def local_pool_dnn_grad(node):
+@register_opt('cudnn')
+@local_optimizer([GpuDownsampleFactorMaxGrad])
+def local_pool_dnn_grad(node):
     if not dnn_available():
         return
     if isinstance(node.op, GpuDownsampleFactorMaxGrad):
@@ -1671,9 +1669,9 @@ if False:
                                  gpu_contiguous(inp_grad),
                                  desc)]
 
-    @register_opt('cudnn')
-    @local_optimizer([DownsampleFactorMaxGrad])
-    def local_pool_dnn_grad_stride(node):
+@register_opt('cudnn')
+@local_optimizer([DownsampleFactorMaxGrad])
+def local_pool_dnn_grad_stride(node):
     if not dnn_available():
         return
     if isinstance(node.op, DownsampleFactorMaxGrad):
@@ -1696,9 +1694,9 @@ if False:
                                    desc)
             return [host_from_gpu(ret)]
 
-    @register_opt('cudnn')
-    @local_optimizer([GpuSoftmax])
-    def local_softmax_dnn(node):
+@register_opt('cudnn')
+@local_optimizer([GpuSoftmax])
+def local_softmax_dnn(node):
     if not dnn_available():
         return
     if isinstance(node.op, GpuSoftmax):
@@ -1708,7 +1706,7 @@ if False:
         out = as_gpuarray_variable(out.dimshuffle(0, 1))
         return [out]
 
-    class NoCuDNNRaise(Optimizer):
+class NoCuDNNRaise(Optimizer):
     def apply(self, fgraph):
         """ Raise a RuntimeError if cudnn can't be used"""
         if not dnn_available():
@@ -1718,11 +1716,11 @@ if False:
                 "cuDNN optimization was enabled, but Theano was not able"
                 " to use it. We got this error: \n" +
                 dnn_available.msg)
-    gpu_seqopt.register("NoCuDNNRaise", NoCuDNNRaise(), 0, 'cudnn')
+gpu_seqopt.register("NoCuDNNRaise", NoCuDNNRaise(), 0, 'cudnn')
 
-    @register_opt('cudnn')
-    @local_optimizer([SoftmaxGrad])
-    def local_softmax_dnn_grad(node):
+@register_opt('cudnn')
+@local_optimizer([SoftmaxGrad])
+def local_softmax_dnn_grad(node):
     if (isinstance(node.op, SoftmaxGrad) and
         ((node.inputs[0].owner and
           isinstance(node.inputs[0].owner.op, HostFromGpu)) or
@@ -1738,12 +1736,8 @@ if False:
                 return
             ins.append(n.dimshuffle(0, 1, 'x', 'x'))
 
-            out = GpuDnnSoftmaxGrad(
-                'bc01',
+        out = GpuDnnSoftmaxGrad('bc01',
                                 'accurate',
-                'channel'
-            )(
-                gpu_contiguous(ins[0]),
-                gpu_contiguous(ins[1])
-            )
+                                'channel')(gpu_contiguous(ins[0]),
+                                           gpu_contiguous(ins[1]))
         return [out.dimshuffle(0, 1)]