fix some bugs, support python_out_sig

python/paddle/fluid/tests/unittests/op_test.py

@@ -731,11 +731,12 @@ def parse_attri_value(name, op_inputs, op_attrs):
                 if name in op_proto_attrs:
                     return op_proto_attrs[name]
                 elif name in op_inputs:
-                    if len(op_inputs[name]) == 1: 
+                    if len(op_inputs[name]) == 1:
                         # why don't use numpy().item() : if the Tensor is float64, we will change it to python.float32, where we loss accuracy: [allclose_op]
                         # why we reconstruct a tensor: because we want the tensor in cpu. 
-                        return paddle.to_tensor(op_inputs[name][0].numpy(), place='cpu')
-                    else: 
+                        return paddle.to_tensor(
+                            op_inputs[name][0].numpy(), place='cpu')
+                    else:
                         # if this is a list (test_unsqueeze2_op): we just pass it into the python api.
                         return op_inputs[name]
                 else:
@@ -828,7 +829,9 @@ def _get_kernel_signature(eager_tensor_inputs, eager_tensor_outputs,
                 """ we think the kernel_sig is missing.
                 """
                 kernel_sig = None
-                print ("[Warning: op_test.py] Kernel Signature is not found for %s, fall back to intermediate state." % self.op_type)
+                print(
+                    "[Warning: op_test.py] Kernel Signature is not found for %s, fall back to intermediate state."
+                    % self.op_type)
             return kernel_sig
 
         def cal_python_api(python_api, args, kernel_sig):
@@ -1946,16 +1949,17 @@ def _get_dygraph_grad(self,
                         attrs_outputs[attrs_name] = self.attrs[attrs_name]
 
             if check_eager:
-                eager_outputs = self._calc_python_api_output(place, inputs, outputs)
+                eager_outputs = self._calc_python_api_output(place, inputs,
+                                                             outputs)
             # if outputs is None, kernel sig is empty or other error is happens.
-            if not check_eager or outputs is None:
+            if not check_eager or eager_outputs is None:
                 block.append_op(
                     type=self.op_type,
                     inputs=inputs,
                     outputs=outputs,
                     attrs=attrs_outputs if hasattr(self, "attrs") else None)
-            else: 
-                output = eager_outputs
+            else:
+                outputs = eager_outputs
 
             if self.dtype == np.uint16:
                 cast_inputs = self._find_var_in_dygraph(outputs,