fix conv3d import problem.

theano/sandbox/cuda/GpuConv3D.py

+import numpy
+
 import theano
 import theano.tensor as T
-import numpy as N
-from theano.sandbox.cuda import cuda_available, cuda_enabled
-from theano.sandbox.cuda.basic_ops import *
-#from util import strutil
-from .Conv3D import Conv3D
-
-if cuda_available:
-    from theano.sandbox.cuda import CudaNdarrayType, float32_shared_constructor
+from theano.gof import local_optimizer
+from theano.sandbox.cuda.basic_ops import as_cuda_ndarray_variable, host_from_gpu, HostFromGpu
+from theano.misc import strutil
+from theano.tensor.nnet.Conv3D import Conv3D
+from theano.sandbox.cuda.opt import register_opt
+from theano.sandbox.cuda import CudaNdarrayType
 
 class GpuConv3D(theano.Op):
     """ GPU implementation of Conv3D """
@@ -282,8 +282,8 @@ conv_rows_stack( float* img, float* kern, float* bias, float* out,
 
 gpu_convd = GpuConv3D()
 
-@theano.sandbox.cuda.opt.register_opt()
-@theano.gof.opt.local_optimizer([])
+@register_opt()
+@local_optimizer([])
 def local_gpu_conv3d(node):
     if isinstance(node.op, Conv3D):
         if numpy.any([i.owner and isinstance(i.owner.op, HostFromGpu) for i in node.inputs]):

theano/sandbox/cuda/GpuConvGrad3D.py

+import numpy
+
 import theano
 import theano.tensor as T
-import numpy as N
-from theano.sandbox.cuda import cuda_available, cuda_enabled
-from theano.sandbox.cuda.basic_ops import *
-#from util import strutil
-from .ConvGrad3D import ConvGrad3D
+from theano.gof import local_optimizer
+from theano.sandbox.cuda.basic_ops import as_cuda_ndarray_variable
+from theano.misc import strutil
+
+from theano.tensor.nnet.ConvGrad3D import ConvGrad3D
+from theano.sandbox.cuda.opt import register_opt
+from theano.sandbox.cuda import CudaNdarrayType, HostFromGpu, host_from_gpu
+
 
 
 class GpuConvGrad3D(theano.Op):
@@ -43,7 +48,7 @@ class GpuConvGrad3D(theano.Op):
         inputDur = V.shape[4]
         dr, dc, dt = d
 
-        dCdW = N.zeros(WShape, dtype=V.dtype)
+        dCdW = numpy.zeros(WShape, dtype=V.dtype)
 
         #block
         for j in xrange(0,WShape[0]):
@@ -259,7 +264,7 @@ if(!work_complete){
             ///////////// < /code generated by GpuConvGrad3D >
         """
 
-        return strutls.render_string(codeSource,locals())
+        return strutil.render_string(codeSource,locals())
 
     def c_support_code_apply(self, node, nodename):
         # This code is not sensitive to the ignore_border flag.
@@ -333,8 +338,8 @@ convgrad_rows_stack( float* img, float* dCdH, float* dCdW,
 
 gpu_conv_grad3d = GpuConvGrad3D()
 
-@theano.sandbox.cuda.opt.register_opt()
-@theano.gof.opt.local_optimizer([])
+@register_opt()
+@local_optimizer([])
 def local_gpu_conv_gradd(node):
     if isinstance(node.op, ConvGrad3D):
         if numpy.any([i.owner and isinstance(i.owner.op, HostFromGpu) for i in node.inputs]):

theano/sandbox/cuda/GpuConvTransp3D.py

-import numpy as N
+import numpy
+
 import theano.tensor as T
-#from util import strutil
+from theano.misc import strutil
 import theano
-from .ConvTransp3D import ConvTransp3D
 
-from theano.sandbox.cuda import cuda_available, cuda_enabled
-from theano.sandbox.cuda.basic_ops import *
-if cuda_available:
-    from theano.sandbox.cuda import CudaNdarrayType, float32_shared_constructor
+from theano.tensor.nnet.ConvTransp3D import ConvTransp3D
+from theano.gof import local_optimizer
+
+from theano.sandbox.cuda.basic_ops import as_cuda_ndarray_variable
+from theano.sandbox.cuda.opt import register_opt
+from theano.sandbox.cuda import CudaNdarrayType, HostFromGpu, host_from_gpu
 
 
 class GpuConvTransp3D(theano.Op):
@@ -259,7 +261,7 @@ if(!work_complete){
 }}}}}} // for fail
 			///////////// < /code generated by GpuConvTransp3D >
 		"""
-		return renderString(codeSource,locals())
+		return strutil.renderString(codeSource,locals())
 
     def c_support_code_apply(self, node, nodename):
         # This code is not sensitive to the ignore_border flag.
@@ -317,7 +319,7 @@ conv_transp_rows_stack( float* H, float* kern, float* bias, float* R,
     			int tk = t - tc * dt;
     			if(tk < 0)
     				break;
-    			//R[i,j,r,c,t] += N.dot(W[:,j,rk,ck,tk], H[i,:,rc,cc,tc] )
+    			//R[i,j,r,c,t] += numpy.dot(W[:,j,rk,ck,tk], H[i,:,rc,cc,tc] )
                         for(int q=0;q<nkern;q++){
                           sum += kern[q*kern_stride_nkern+stack_id*kern_stride_stack+rk*kern_stride_row+ck*kern_stride_col+tk*kern_stride_frame]*
                                  H[batch_id*H_stride_batch+q*H_stride_stack+rc*H_stride_row+cc*H_stride_col+tc*H_stride_frame];
@@ -343,8 +345,8 @@ conv_transp_rows_stack( float* H, float* kern, float* bias, float* R,
 
 gpu_conv_transpd = GpuConvTransp3D()
 
-@theano.sandbox.cuda.opt.register_opt()
-@theano.gof.opt.local_optimizer([])
+@register_opt()
+@local_optimizer([])
 def local_gpu_conv_transpd(node):
     if isinstance(node.op, ConvTransp3D):
         if numpy.any([i.owner and isinstance(i.owner.op, HostFromGpu) for i in node.inputs]):
@@ -390,7 +392,7 @@ def computeR(W,b,d,H,Rshape = None):
 
 		#print "video size: "+str((videoHeight, videoWidth, videoDur))
 
-		R =  N.zeros( (batchSize, inputChannels, videoHeight,
+		R =  numpy.zeros( (batchSize, inputChannels, videoHeight,
 			videoWidth, videoDur ) , dtype=H.dtype)
 
 		#R[i,j,r,c,t] = b_j + sum_{rc,rk | d \circ rc + rk = r} sum_{cc,ck | ...} sum_{tc,tk | ...} sum_k W[k, j, rk, ck, tk] * H[i,k,rc,cc,tc]
@@ -404,10 +406,10 @@ def computeR(W,b,d,H,Rshape = None):
 						for t in xrange(0,videoDur):
 							R[i,j,r,c,t] = b[j]
 
-							ftc = max([0, int(N.ceil(float(t-filterDur +1  )/float(dt))) ])
-							fcc = max([0, int(N.ceil(float(c-filterWidth +1)/float(dc))) ])
+							ftc = max([0, int(numpy.ceil(float(t-filterDur +1  )/float(dt))) ])
+							fcc = max([0, int(numpy.ceil(float(c-filterWidth +1)/float(dc))) ])
 
-							rc =  max([0, int(N.ceil(float(r-filterHeight+1)/float(dr))) ])
+							rc =  max([0, int(numpy.ceil(float(r-filterHeight+1)/float(dr))) ])
 							while rc < outputHeight:
 								rk = r - rc * dr
 								if rk < 0:
@@ -425,7 +427,7 @@ def computeR(W,b,d,H,Rshape = None):
 										if tk < 0:
 											break
 
-										R[i,j,r,c,t] += N.dot(W[:,j,rk,ck,tk], H[i,:,rc,cc,tc] )
+										R[i,j,r,c,t] += numpy.dot(W[:,j,rk,ck,tk], H[i,:,rc,cc,tc] )
 
 										tc += 1
 									"" #close loop over tc
@@ -441,7 +443,3 @@ def computeR(W,b,d,H,Rshape = None):
 		"" #close loop over i
 
 		return R
-
-
-from ops.Conv3D import *
-from ops.ConvGrad3D import *