fix speedup with CUDA (#2947)

src/sdk/pynni/nni/compression/torch/speedup/infer_shape.py

@@ -573,7 +573,7 @@ def view_inshape(module_masks, mask, shape):
     step_size = shape['in_shape'][2] * shape['in_shape'][3]
     for loc in mask.mask_index[1]:
         index.extend([loc * step_size + i for i in range(step_size)])
-    output_cmask.add_index_mask(dim=1, index=torch.tensor(index))  # pylint: disable=not-callable
+    output_cmask.add_index_mask(dim=1, index=torch.tensor(index).to(mask.mask_index[1].device))  # pylint: disable=not-callable
     module_masks.set_output_mask(output_cmask)
     return output_cmask
 
@@ -609,7 +609,7 @@ def view_outshape(module_masks, mask, shape):
     step_size = shape['in_shape'][2] * shape['in_shape'][3]
     for loc in mask.mask_index[1]:
         index.extend([loc * step_size + i for i in range(step_size)])
-    input_cmask.add_index_mask(dim=1, index=torch.tensor(index))  # pylint: disable=not-callable
+    input_cmask.add_index_mask(dim=1, index=torch.tensor(index).to(mask.mask_index[1].device))  # pylint: disable=not-callable
     module_masks.set_input_mask(input_cmask)
 
     return input_cmask
@@ -870,7 +870,7 @@ def conv2d_mask(module_masks, mask):
         if index is None:
             return None, None, None
         else:
-            index = torch.LongTensor(index).to(weight_mask.device)
+            index = index.long().to(weight_mask.device)
             weight_cmask = CoarseMask(num_dim=4)
             weight_cmask.add_index_mask(dim=dim, index=index)
             bias_cmask = None