Merge branch 'fix_multi_inputs_converter' into 'master'

Fix pad and conv dilation issues

See merge request !527

mace/python/tools/converter_tool/tensorflow_converter.py

@@ -294,7 +294,12 @@ class TensorflowConverter(base_converter.ConverterInterface):
         if op.type != MaceOp.Deconv2D.name:
             dilation_arg = op.arg.add()
             dilation_arg.name = MaceKeyword.mace_dilations_str
-            dilation_arg.ints.extend(tf_op.get_attr(tf_dilations_str)[1:3])
+            try:
+                dilation_val = tf_op.get_attr(tf_dilations_str)[1:3]
+            except ValueError:
+                dilation_val = [1, 1]
+
+            dilation_arg.ints.extend(dilation_val)
 
     def convert_elementwise(self, tf_op):
         op = self.convert_general_op(tf_op)

mace/python/tools/converter_tool/transformer.py

@@ -713,18 +713,21 @@ class Transformer(base_converter.ConverterInterface):
             # transpose args
             if op.type == MaceOp.Pad.name:
                 for arg in op.arg:
-                    if arg.name == MaceKeyword.mace_paddings_str and len(
-                            arg.ints) == 4:
+                    if arg.name == MaceKeyword.mace_paddings_str:
+                        mace_check(len(arg.ints) == 8,
+                                   "pad dim rank should be 8.")
                         if ConverterUtil.data_format(op) == DataFormat.NHWC \
                                 and self._target_data_format == DataFormat.NCHW:  # noqa
                             print("Transpose pad args: %s(%s)"
                                   % (op.name, op.type))
-                            self.transpose_shape(arg.ints, [0, 3, 1, 2])
+                            self.transpose_shape(arg.ints,
+                                                 [0, 1, 6, 7, 2, 3, 4, 5])
                         elif ConverterUtil.data_format(op) == DataFormat.NCHW \
                                 and self._target_data_format == DataFormat.NHWC:  # noqa
                             print("Transpose pad args: %s(%s)"
                                   % (op.name, op.type))
-                            self.transpose_shape(arg.ints, [0, 2, 3, 1])
+                            self.transpose_shape(arg.ints,
+                                                 [0, 1, 4, 5, 6, 7, 2, 3])
             elif op.type == MaceOp.Concat.name or op.type == MaceOp.Slice.name:
                 for arg in op.arg:
                     if arg.name == MaceKeyword.mace_axis_str: