diff --git a/tools/python/transform/transformer.py b/tools/python/transform/transformer.py
index 9b26d487cb3055864f36f1af94cdc02736d1a40d..a10b8701c36d34d72dbf4776d4087dec0a5ca6e1 100644
--- a/tools/python/transform/transformer.py
+++ b/tools/python/transform/transformer.py
@@ -963,7 +963,7 @@ class Transformer(base_converter.ConverterInterface):
                 if consumer_op.type == MaceOp.Activation.name \
                         and ConverterUtil.get_arg(
                             consumer_op,
-                            MaceKeyword.mace_activation_type_str).s != 'PRELU':
+                            MaceKeyword.mace_activation_type_str).s != b'PRELU':  # noqa
                     print("Fold activation: %s(%s)" % (op.name, op.type))
                     op.name = consumer_op.name
                     op.output[0] = consumer_op.output[0]
diff --git a/tools/validate.py b/tools/validate.py
index 9980a0a92c74a19744f9e0fa4d55722c1ff3849c..8b8a9e82c8eccd10243ba1232984c080c873900d 100644
--- a/tools/validate.py
+++ b/tools/validate.py
@@ -64,7 +64,16 @@ def calculate_similarity(u, v, data_type=np.float64):
         u = u.astype(data_type)
     if v.dtype is not data_type:
         v = v.astype(data_type)
-    return np.dot(u, v) / (np.linalg.norm(u) * np.linalg.norm(v))
+    u_norm = np.linalg.norm(u)
+    v_norm = np.linalg.norm(v)
+    norm = u_norm * v_norm
+    if norm == 0:
+        if u_norm == 0 and v_norm == 0:
+            return 1
+        else:
+            return 0
+    else:
+        return np.dot(u, v) / norm
 
 
 def calculate_pixel_accuracy(out_value, mace_out_value):