fix doc errors, test=develop (#2023)

doc/fluid/api_cn/layers_cn/GRUCell_cn.rst

@@ -35,7 +35,7 @@ GRUCell
 ..  code-block:: python 
 
     import paddle.fluid.layers as layers
-    cell = layers.rnn.GRUCell(hidden_size=256)
+    cell = layers.GRUCell(hidden_size=256)
 
 
 .. py:method:: call(inputs, states)

doc/fluid/api_cn/layers_cn/lstm_cn.rst

@@ -10,7 +10,7 @@ lstm
 .. note::
     该OP仅支持 GPU 设备运行
 
-该OP实现了 LSTM，即 Long-Short Term Memory（长短期记忆）运算 - `Hochreiter, S., & Schmidhuber, J. (1997) <http://deeplearning.cs.cmu.edu/pdfs/Hochreiter97_lstm.pdf>`_。
+该OP实现了 LSTM，即 Long-Short Term Memory（长短期记忆）运算 - `Hochreiter, S., & Schmidhuber, J. (1997) <https://www.bioinf.jku.at/publications/older/2604.pdf>`_。
 
 该OP的实现不包括 diagonal/peephole 连接，参见 `Gers, F. A., & Schmidhuber, J. (2000) <ftp://ftp.idsia.ch/pub/juergen/TimeCount-IJCNN2000.pdf>`_。
 如果需要使用 peephole 连接方法，请使用 :ref:`cn_api_fluid_layers_dynamic_lstm` 。

doc/fluid/api_cn/layers_cn/lstm_unit_cn.rst

@@ -48,26 +48,18 @@ Long-Short Term Memory（LSTM）循环神经网络计算单元。该OP用于完
 **代码示例**：
 
 .. code-block:: python
-
+   
     import paddle.fluid as fluid
-     
+
     dict_dim, emb_dim, hidden_dim = 128, 64, 512
-    data = fluid.layers.data(name='step_data', shape=[1], dtype='int32')
-    x = fluid.layers.embedding(input=data, size=[dict_dim, emb_dim])
-    pre_hidden = fluid.layers.data(name='pre_hidden', shape=[hidden_dim], dtype='float32')
-    pre_cell = fluid.layers.data(name='pre_cell', shape=[hidden_dim], dtype='float32')
+    data = fluid.data(name='step_data', shape=[None], dtype='int64')
+    x = fluid.embedding(input=data, size=[dict_dim, emb_dim])
+    pre_hidden = fluid.data(
+        name='pre_hidden', shape=[None, hidden_dim], dtype='float32')
+    pre_cell = fluid.data(
+        name='pre_cell', shape=[None, hidden_dim], dtype='float32')
     hidden = fluid.layers.lstm_unit(
         x_t=x,
         hidden_t_prev=pre_hidden,
         cell_t_prev=pre_cell)
-
-
-
-
-
-
-
-
-
-
-
+