Merge pull request #101 from wangguibao/doc

Fix documentaion

doc/INSTALL.md

@@ -16,10 +16,12 @@ openssl & openssl-devel
 
 ## 编译
 
-推荐使用Docker编译Paddle Serving, [Docker编译使用说明](./DOCKER.md)
+推荐使用Docker准备Paddle Serving编译环境. [Docker编译使用说明](./DOCKER.md)
+
+以下命令将会下载Paddle Serving最新代码，并执行编译
 
 ```shell
-$ git clone https://github.com/PaddlePaddle/serving.git
+$ git clone https://github.com/PaddlePaddle/Serving.git
 $ cd serving
 $ mkdir build
 $ cd build
@@ -28,39 +30,31 @@ $ make -j4
 $ make install
 ```
 
-`make install`将把目标产出放在/path/to/paddle-serving/build/output/目录下，目录结构：
+`make install`将把目标产出放在/path/to/Paddle-Serving/build/output/目录下，目录结构：
 
 ```
 .
-|-- bin                             # Paddle Serving protobuf编译插件pdcodegen所在目录
+|-- bin                             # Paddle Serving工具和protobuf编译插件pdcodegen所在目录
+|-- conf
 |-- demo                            # demo总目录
-|   |-- client
+|   |-- client                      # Demo client端
+|   |   |-- bert                    # bert模型客户端
+|   |   |-- ctr_prediction          # CTR prediction模型客户端
 |   |   |-- dense_format            # dense_format客户端
-|   |   |   |-- bin                 # bin/dense_format是dense_format客户端bin
-|   |   |   `-- conf
-|   |   |-- echo                    # echo服务客户端
-|   |   |   |-- bin                 # bin/echo是echo客户端bin
-|   |   |   \-- conf
-|   |   |-- image_classification    # image_classification服务客户端
-|   |   |   |-- bin                 # bin/ximage是image_classification客户端bin
-|   |   |   |-- conf
-|   |   |   |-- data
-|   |   |      `-- images
-|   |   |-- int64tensor_format      # int64tensor_format服务客户端
-|   |   |   |-- bin                 # bin/int64tensor_format是客户端bin
-|   |   |   `-- conf
-|   |   `-- sparse_format           # sparse_format客户端
-|   |       |-- bin                 # bin/sparse_format是客户端bin
-|   |       `-- conf
-|   `-- serving                     # serving端，同时提供echo/dense_format/sparse_format/int64tensor_format/image_class等5种服务
-|       |-- bin                     # bin/serving是serving端可执行bin
-|       |-- conf                    # 配置文件目录
-|       |-- data
-|       |   `-- model
-|       |       `-- paddle
-|       |           `-- fluid
-|       |               `-- SE_ResNeXt50_32x4d  # image_classification模型
-`-- lib                             # Paddle Serving产出的静态库文件: libpdseving.a, libsdk-cpp.a, libconfigure.a, libfluid_cpu_engine.a
+|   |   |-- echo                    # 最简单的echo service客户端
+|   |   |-- echo_kvdb               # local KV读取demo客户端
+|   |   |-- image_classification    # 图像分类任务客户端
+|   |   |-- int64tensor_format      # int64tensor_format示例客户端
+|   |   |-- sparse_format           # sparse_format示例客户端
+|   |   `-- text_classification     # 文本分类任务示例客户端
+|   |-- db_func
+|   |-- db_thread
+|   |-- kvdb_test
+|   `-- serving                     # Demo serving端；该serving可同时响应所有demo client请求
+|-- include                         # Paddle Serving发布的头文件
+|-- lib                             # Paddle Serving发布的libs
+`-- tool                            # Paddle Serving发布的工具目录
+
 ```
 
 如要编写新的预测服务，请参考[从零开始写一个预测服务](CREATING.md)