diff --git a/mindarmour/adv_robustness/evaluations/attack_evaluation.py b/mindarmour/adv_robustness/evaluations/attack_evaluation.py
index c327510..48e3914 100644
--- a/mindarmour/adv_robustness/evaluations/attack_evaluation.py
+++ b/mindarmour/adv_robustness/evaluations/attack_evaluation.py
@@ -136,10 +136,10 @@ class AttackEvaluate:
             - float, return average l0, l2, or linf distance of all success
               adversarial examples, return value includes following cases.
 
-                - If return value :math:`>=` 0, average lp distance. The lower,
-                  the more successful the attack is.
+            - If return value :math:`>=` 0, average lp distance. The lower,
+              the more successful the attack is.
 
-                - If return value is -1, there is no success adversarial examples.
+            - If return value is -1, there is no success adversarial examples.
         """
         idxes = self._success_idxes
         success_num = idxes.shape[0]
@@ -164,10 +164,10 @@ class AttackEvaluate:
         Returns:
             - float, average structural similarity.
 
-                - If return value ranges between (0, 1), the higher, the more
-                  successful the attack is.
+            - If return value ranges between (0, 1), the higher, the more
+              successful the attack is.
 
-                - If return value is -1: there is no success adversarial examples.
+            - If return value is -1: there is no success adversarial examples.
         """
         success_num = self._success_idxes.shape[0]
         if success_num == 0:
diff --git a/mindarmour/privacy/diff_privacy/mechanisms/mechanisms.py b/mindarmour/privacy/diff_privacy/mechanisms/mechanisms.py
index 9419c87..c3a9e8e 100644
--- a/mindarmour/privacy/diff_privacy/mechanisms/mechanisms.py
+++ b/mindarmour/privacy/diff_privacy/mechanisms/mechanisms.py
@@ -183,8 +183,8 @@ class NoiseGaussianRandom(_Mechanisms):
         initial_noise_multiplier(float): Ratio of the standard deviation of
             Gaussian noise divided by the norm_bound, which will be used to
             calculate privacy spent. Default: 1.0.
-        seed(int): Original random seed, if seed=0 random normal will use secure
-            random number. IF seed!=0 random normal will generate values using
+        seed(int): Original random seed, if seed=0, random normal will use secure
+            random number. If seed!=0, random normal will generate values using
             given seed. Default: 0.
         decay_policy(str): Mechanisms parameters update policy. Default: None.
 
diff --git a/mindarmour/privacy/evaluation/membership_inference.py b/mindarmour/privacy/evaluation/membership_inference.py
index 4e5d00d..df9e492 100644
--- a/mindarmour/privacy/evaluation/membership_inference.py
+++ b/mindarmour/privacy/evaluation/membership_inference.py
@@ -95,7 +95,7 @@ def _softmax_cross_entropy(logits, labels):
 class MembershipInference:
     """
     Evaluation proposed by Shokri, Stronati, Song and Shmatikov is a grey-box attack.
-    The attack requires obtain loss or logits results of training samples.
+    The attack requires loss or logits results of training samples.
 
     References: `Reza Shokri, Marco Stronati, Congzheng Song, Vitaly Shmatikov.
     Membership Inference Attacks against Machine Learning Models. 2017.