diff --git a/docs/api/api_python/ops/mindspore.ops.Gcd.rst b/docs/api/api_python/ops/mindspore.ops.Gcd.rst
index 2c8a837bb9b..5e491261378 100644
--- a/docs/api/api_python/ops/mindspore.ops.Gcd.rst
+++ b/docs/api/api_python/ops/mindspore.ops.Gcd.rst
@@ -3,7 +3,7 @@ mindspore.ops.Gcd
 
 .. py:class:: mindspore.ops.Gcd
 
-    逐元素计算输入输入Tensor的最大公约数。两个输入的shape需要能进行广播操作，并且数据类型必须为：int32、int64。
+    逐元素计算输入Tensor的最大公约数。两个输入的shape需要能进行广播操作，并且数据类型必须为：int32、int64。
 
     输入：
         - **x1** (Tensor) - 第一个输入Tensor。
diff --git a/mindspore/python/mindspore/ops/function/array_func.py b/mindspore/python/mindspore/ops/function/array_func.py
index b989f3bab22..0a52d4deaa2 100644
--- a/mindspore/python/mindspore/ops/function/array_func.py
+++ b/mindspore/python/mindspore/ops/function/array_func.py
@@ -5581,7 +5581,7 @@ def diagonal(input, offset=0, dim1=0, dim2=1):
         ValueError: if the input tensor has less than two dimensions.
 
     Supported Platforms:
-        ``Ascend`` ``CPU`` ``GPU``
+        ``Ascend`` ``GPU`` ``CPU``
 
     Examples:
         >>> x = Tensor([[0, 1], [2, 3]], mstype.float32)
diff --git a/mindspore/python/mindspore/ops/operations/nn_ops.py b/mindspore/python/mindspore/ops/operations/nn_ops.py
index 7a33305420f..a14a3fb217c 100644
--- a/mindspore/python/mindspore/ops/operations/nn_ops.py
+++ b/mindspore/python/mindspore/ops/operations/nn_ops.py
@@ -9595,7 +9595,7 @@ class GridSampler2D(Primitive):
         ValueError: If `padding_mode` is not "zeros", "border", "reflection" or a string value.
 
     Supported Platforms:
-        ``Ascend`` ``CPU`` ``GPU``
+        ``Ascend`` ``GPU`` ``CPU``
 
     Examples:
         >>> gridsampler = ops.GridSampler2D(interpolation_mode='bilinear', padding_mode='zeros', align_corners=True)
@@ -9715,7 +9715,7 @@ class UpsampleNearest3D(Primitive):
         ValueError: If size of `output_size` is not equal 3 when `output_size` is specified.
 
     Supported Platforms:
-        ``GPU`` ``Ascend`` ``CPU``
+        ``Ascend`` ``GPU`` ``CPU``
 
     Examples:
         >>> x = Tensor(np.array([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16])