diff --git a/develop/doc/api/v2/fluid/layers.html b/develop/doc/api/v2/fluid/layers.html
index 3c06ea26f7893a9d1c8ac5f899364605cc296297..98cb20d5346b58b9ac9434c0c05ef145399af210 100644
--- a/develop/doc/api/v2/fluid/layers.html
+++ b/develop/doc/api/v2/fluid/layers.html
@@ -717,25 +717,19 @@ Duplicable: False  Optional: False</li>
 <code class="descclassname">paddle.v2.fluid.layers.</code><code class="descname">transpose</code><span class="sig-paren">(</span><em>**kwargs</em><span class="sig-paren">)</span></dt>
 <dd><p>Transpose Operator.</p>
 <p>The input tensor will be permuted according to the axis values given.
-The op functions similar to how numpy.transpose works in python.
-For example:</p>
-<blockquote>
-<div><p>&gt;&gt; input = numpy.arange(6).reshape((2,3))
-&gt;&gt; input
+The op functions is similar to how numpy.transpose works in python.</p>
+<p>For example: input = numpy.arange(6).reshape((2,3))
+the input is:
 array([[0, 1, 2],</p>
 <blockquote>
 <div>[3, 4, 5]])</div></blockquote>
-<p>&gt;&gt; axis = [1, 0]
-&gt;&gt; output = input.transpose(axis)
-&gt;&gt; output
+<p>given axis is: [1, 0]</p>
+<p>output = input.transpose(axis)
+then the output is:
 array([[0, 3],</p>
 <blockquote>
-<div><dl class="docutils">
-<dt>[1, 4],</dt>
-<dd>[2, 5]])</dd>
-</dl>
-</div></blockquote>
-</div></blockquote>
+<div>[1, 4],
+[2, 5]])</div></blockquote>
 <p>So, given a input tensor of shape(N, C, H, W) and the axis is {0, 2, 3, 1},
 the output tensor shape will be (N, H, W, C)</p>
 <table class="docutils field-list" frame="void" rules="none">
diff --git a/develop/doc/operators.json b/develop/doc/operators.json
index 6228db289ea27e1f6e51ef172f58d1c8791e000b..7bd500314c58ddf78beeda7a74b96bf0bcf904a9 100644
--- a/develop/doc/operators.json
+++ b/develop/doc/operators.json
@@ -2170,7 +2170,7 @@
  } ] 
 },{
  "type" : "unpool",
- "comment" : "\n\"Input shape: $(N, C_{in}, H_{in}, W_{in})$, \nOutput shape: $(N, C_{out}, H_{out}, W_{out})$\nWhere\n$$\nH_{out} = (H_{in}−1) * strides[0] − 2 * paddings[0] + ksize[0] \\\\\nW_{out} = (W_{in}−1) * strides[1] − 2 * paddings[1] + ksize[1]\n$$\nPaper: http://www.matthewzeiler.com/wp-content/uploads/2017/07/iccv2011.pdf\n",
+ "comment" : "\nInput shape is: $(N, C_{in}, H_{in}, W_{in})$, Output shape is:\n$(N, C_{out}, H_{out}, W_{out})$, where\n$$\nH_{out} = (H_{in}−1) * strides[0] − 2 * paddings[0] + ksize[0] \\\\\nW_{out} = (W_{in}−1) * strides[1] − 2 * paddings[1] + ksize[1]\n$$\nPaper: http://www.matthewzeiler.com/wp-content/uploads/2017/07/iccv2011.pdf\n",
  "inputs" : [ 
  { 
    "name" : "X",
@@ -2214,7 +2214,7 @@
  } ] 
 },{
  "type" : "transpose",
- "comment" : "\nTranspose Operator.\n\nThe input tensor will be permuted according to the axis values given.\nThe op functions similar to how numpy.transpose works in python.\nFor example:\n >> input = numpy.arange(6).reshape((2,3))\n >> input\n array([[0, 1, 2],\n        [3, 4, 5]])\n >> axis = [1, 0]\n >> output = input.transpose(axis)\n >> output\n array([[0, 3],\n        [1, 4],\n\t\t[2, 5]])\nSo, given a input tensor of shape(N, C, H, W) and the axis is {0, 2, 3, 1},\nthe output tensor shape will be (N, H, W, C)\n\n",
+ "comment" : "\nTranspose Operator.\n\nThe input tensor will be permuted according to the axis values given.\nThe op functions is similar to how numpy.transpose works in python.\n\nFor example: input = numpy.arange(6).reshape((2,3))\nthe input is:\narray([[0, 1, 2],\n      [3, 4, 5]])\ngiven axis is: [1, 0]\n\noutput = input.transpose(axis)\nthen the output is:\narray([[0, 3],\n       [1, 4],\n       [2, 5]])\nSo, given a input tensor of shape(N, C, H, W) and the axis is {0, 2, 3, 1},\nthe output tensor shape will be (N, H, W, C)\n\n",
  "inputs" : [ 
  { 
    "name" : "X",
diff --git a/develop/doc_cn/api/v2/fluid/layers.html b/develop/doc_cn/api/v2/fluid/layers.html
index cdecea77502f0fd0a7db9ced43a0f658a8c8f061..f12ca61f45b4b1744cdabd52998ae115e88fcf82 100644
--- a/develop/doc_cn/api/v2/fluid/layers.html
+++ b/develop/doc_cn/api/v2/fluid/layers.html
@@ -730,25 +730,19 @@ Duplicable: False  Optional: False</li>
 <code class="descclassname">paddle.v2.fluid.layers.</code><code class="descname">transpose</code><span class="sig-paren">(</span><em>**kwargs</em><span class="sig-paren">)</span></dt>
 <dd><p>Transpose Operator.</p>
 <p>The input tensor will be permuted according to the axis values given.
-The op functions similar to how numpy.transpose works in python.
-For example:</p>
-<blockquote>
-<div><p>&gt;&gt; input = numpy.arange(6).reshape((2,3))
-&gt;&gt; input
+The op functions is similar to how numpy.transpose works in python.</p>
+<p>For example: input = numpy.arange(6).reshape((2,3))
+the input is:
 array([[0, 1, 2],</p>
 <blockquote>
 <div>[3, 4, 5]])</div></blockquote>
-<p>&gt;&gt; axis = [1, 0]
-&gt;&gt; output = input.transpose(axis)
-&gt;&gt; output
+<p>given axis is: [1, 0]</p>
+<p>output = input.transpose(axis)
+then the output is:
 array([[0, 3],</p>
 <blockquote>
-<div><dl class="docutils">
-<dt>[1, 4],</dt>
-<dd>[2, 5]])</dd>
-</dl>
-</div></blockquote>
-</div></blockquote>
+<div>[1, 4],
+[2, 5]])</div></blockquote>
 <p>So, given a input tensor of shape(N, C, H, W) and the axis is {0, 2, 3, 1},
 the output tensor shape will be (N, H, W, C)</p>
 <table class="docutils field-list" frame="void" rules="none">