Merge branch 'atrous_depthwise' into 'master'

Support atrous depthwise convolution converter.

See merge request !438

mace/python/tools/tf_converter_lib.py

@@ -946,7 +946,8 @@ class TFConverter(object):
     def is_atrous_conv2d(self, op):
         return op.type == 'SpaceToBatchND' and \
                len(self.tf_graph[op.name]) == 1 and \
-               self.tf_graph[op.name][0].type == 'Conv2D'
+               (self.tf_graph[op.name][0].type == 'Conv2D'
+                or self.tf_graph[op.name][0].type == 'DepthwiseConv2dNative')
 
     def convert_atrous_conv2d(self, op):
         op_def = mace_pb2.OperatorDef()
@@ -955,18 +956,25 @@ class TFConverter(object):
         arg.i = self.dt
         conv_op = self.tf_graph[op.name][0]
         op_def.name = conv_op.name
-        op_def.type = conv_op.type
+        if conv_op.type == 'DepthwiseConv2dNative':
+            op_def.type = 'DepthwiseConv2d'
+        else:
+            op_def.type = conv_op.type
 
         if self.device == 'gpu':
-            self.transpose_filter_tensor[
-                get_input_tensor(conv_op, 1).name] = (0, 1, 3, 2)
             op_def.input.extend([op.inputs[0].name])
+            if op_def.type == 'DepthwiseConv2d':
+                buffer_type = "DW_CONV2D_FILTER"
+            else:
+                self.transpose_filter_tensor[get_input_tensor(
+                    conv_op, 1).name] = (0, 1, 3, 2)
+                buffer_type = "CONV2D_FILTER"
             output_name = self.add_buffer_to_image(
-                get_input_tensor(conv_op, 1).name, "CONV2D_FILTER")
+                get_input_tensor(conv_op, 1).name, buffer_type)
             op_def.input.extend([output_name])
         else:
-            self.transpose_filter_tensor[
-                get_input_tensor(conv_op, 1).name] = (3, 2, 0, 1)
+            self.transpose_filter_tensor[get_input_tensor(
+                conv_op, 1).name] = (3, 2, 0, 1)
             op_def.input.extend([get_input_tensor(op, 0).name])
             op_def.input.extend([get_input_tensor(conv_op, 1).name])