fix Variable-length

Author: tjujingzong

docs/design/concepts/tensor_array.md

@@ -212,7 +212,7 @@ class TensorArray:
 ```
 
 ## DenseTensor-related Supports
-The `RecurrentGradientMachine` in Paddle serves as a flexible RNN layer; it takes variance-length sequences as input, and output sequences too.
+The `RecurrentGradientMachine` in Paddle serves as a flexible RNN layer; it takes variable-length sequences as input, and output sequences too.
 
 Since each step of RNN can only take a tensor-represented batch of data as input,
 some preprocess should be taken on the inputs such as sorting the sentences by their length in descending order and cut each word and pack to new batches.
@@ -244,10 +244,10 @@ def pack(level, indices_map):
     pass
 ```
 
-With these two methods, a variance-length sentence supported RNN can be implemented like
+With these two methods, a variable-length sentence supported RNN can be implemented like
 
 ```c++
-// input is the variant-length data
+// input is the variable-length data
 LodTensor sentence_input(xxx);
 TensorArray ta;
 Tensor indice_map;
@@ -268,4 +268,4 @@ for (int step = 0; step = ta.size(); step++) {
 DenseTensor rnn_output = ta.pack(ta, indice_map);
 ```
 the code above shows that by embedding the DenseTensor-related preprocess operations into `TensorArray`,
-the implementation of a RNN that supports variant-length sentences is far more concise than `RecurrentGradientMachine` because the latter mixes all the codes together, hard to read and extend.
+the implementation of a RNN that supports variable-length sentences is far more concise than `RecurrentGradientMachine` because the latter mixes all the codes together, hard to read and extend.

docs/design/dynamic_rnn/rnn_design_en.md

@@ -1,4 +1,4 @@
-# Variant Length supported RNN Design
+# Variable Length supported RNN Design
 For the learning of variable length sequences, the existing mainstream frameworks such as tensorflow, pytorch, caffe2, mxnet and so on all use padding.
 
 Different-length sequences in a mini-batch will be padded with zeros and transformed to same length.