Documents cn (#85)

* Update README.cn.md

* Update README.md

* Update README.md

* Update README.cn.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

README.cn.md

@@ -103,7 +103,7 @@ ans = agent.sum(1,5) # run remotely and not comsume any local computation resour
 # 安装:
 ### 依赖
 - Python 2.7 or 3.5+. 
-- PaddlePaddle >=1.2.1 (**非必须的**，如果你只用并行部分的接口不需要安装paddle) 
+- [PaddlePaddle](https://github.com/PaddlePaddle/Paddle) >=1.2.1 (**非必须的**，如果你只用并行部分的接口不需要安装paddle) 
 
 
 ```
@@ -118,7 +118,7 @@ pip install parl
 - [IMPALA](examples/IMPALA/)
 - [A2C](examples/A2C/)
 - [GA3C](examples/GA3C/)
-- [NIPS2018强化学习假肢挑战赛冠军解决方案](examples/NeurIPS2018-AI-for-Prosthetics-Challenge/)
+- [冠军解决方案：NIPS2018强化学习假肢挑战赛](examples/NeurIPS2018-AI-for-Prosthetics-Challenge/)
 
 <img src=".github/NeurlIPS2018.gif" width = "300" height ="200" alt="NeurlIPS2018"/> <img src=".github/Half-Cheetah.gif" width = "300" height ="200" alt="Half-Cheetah"/> <img src=".github/Breakout.gif" width = "200" height ="200" alt="Breakout"/> 
 <br>

README.md

@@ -4,7 +4,7 @@
 
 English | [简体中文](./README.cn.md)
 
-> PARL is a flexible and high-efficient reinforcement learning framework based on [PaddlePaddle](https://github.com/PaddlePaddle/Paddle).
+> PARL is a flexible and high-efficient reinforcement learning framework.
 
 # Features
 **Reproducible**. We provide algorithms that stably reproduce the result of many influential reinforcement learning algorithms.
@@ -28,7 +28,7 @@ The main abstractions introduced by PARL that are used to build an agent recursi
 `Algorithm` describes the mechanism to update parameters in `Model` and often contains at least one model.
 
 ### Agent
-`Agent`, a data bridge between environment and algorithm, is responsible for data I/O with the outside environment and describes data preprocessing before feeding data into the training process.
+`Agent`, a data bridge between the environment and the algorithm, is responsible for data I/O with the outside environment and describes data preprocessing before feeding data into the training process.
 
 Here is an example of building an agent with DQN algorithm for Atari games.
 ```python
@@ -106,7 +106,7 @@ For users, they can write code in a simple way, just like writing multi-thread c
 # Install:
 ### Dependencies
 - Python 2.7 or 3.5+. 
-- PaddlePaddle >=1.2.1 (**Optional**, if you only want to use APIs related to parallelization alone)  
+- [PaddlePaddle](https://github.com/PaddlePaddle/Paddle) >=1.2.1 (**Optional**, if you only want to use APIs related to parallelization alone)  
 
 
 ```

examples/A2C/README.md

@@ -20,7 +20,7 @@ Mean episode reward in training process after 10 million sample steps.
 + [paddlepaddle>=1.3.0](https://github.com/PaddlePaddle/Paddle)
 + [parl](https://github.com/PaddlePaddle/PARL)
 + gym
-+ atari_py
++ atari-py
 
 
 ### Distributed Training

examples/DQN/README.md

@@ -20,7 +20,7 @@ Please see [here](https://gym.openai.com/envs/#atari) to know more about Atari g
 + [parl](https://github.com/PaddlePaddle/PARL)
 + gym
 + tqdm
-+ atari_py
++ atari-py
 + [ale_python_interface](https://github.com/mgbellemare/Arcade-Learning-Environment)
 
 

examples/GA3C/README.md

@@ -20,7 +20,7 @@ Results with one learner (in a P40 GPU) and 24 simulators (in 12 CPU) in 10 mill
 + [paddlepaddle>=1.3.0](https://github.com/PaddlePaddle/Paddle)
 + [parl](https://github.com/PaddlePaddle/PARL)
 + gym
-+ atari_py
++ atari-py
 
 
 ### Distributed Training

examples/IMPALA/README.md

@@ -24,7 +24,7 @@ Result with one learner (in a P40 GPU) and 32 actors (in 32 CPUs).
 + [paddlepaddle>=1.3.0](https://github.com/PaddlePaddle/Paddle)
 + [parl](https://github.com/PaddlePaddle/PARL)
 + gym
-+ atari_py
++ atari-py
 
 
 ### Distributed Training: