[DOCS] Improve README (#35)

* Improve README

* Improve README

README.md

@@ -52,10 +52,6 @@ The master branch works with **PyTorch 1.6+**.
 <details open>
 <summary>Major features</summary>
 
-- **Modular Design**
-
-  MMYOLO decompose the framework into different components and users can easily construct a customized model by combining different modules and training and testing strategies.
-
 - **Fair and convenient algorithm evaluation**
 
   MMYOLO unifies the modules of various YOLO algorithms and provides a unified benchmark process. Users can compare and analyze in a fair and convenient way.
@@ -64,6 +60,13 @@ The master branch works with **PyTorch 1.6+**.
 
   MMYOLO provides a series of documents from getting started, to model deployment, advanced guidelines, and algorithm analysis, making it easy for different users to get started and make extensions quickly.
 
+- **Modular Design**
+
+  MMYOLO decompose the framework into different components and users can easily construct a customized model by combining different modules and training and testing strategies.
+
+  ![BaseModule](https://user-images.githubusercontent.com/33799979/190382319-6b4e1fcb-cc3f-4fbe-9d6b-3c9c4e57472c.png)
+  The picture is provided by RangeKing@GitHub, thank you very much!
+
 </details>
 
 ## What's New

README_zh-CN.md

@@ -51,10 +51,6 @@ MMYOLO 是一个基于 PyTorch 的 YOLO 系列算法开源工具箱。它是 [Op
 <details open>
 <summary>主要特性</summary>
 
-- **模块化设计**
-
-  MMYOLO 将框架解耦成不同的模块组件，通过组合不同的模块和训练测试策略，用户可以便捷地构建自定义模型。
-
 - **公平便捷的算法评测**
 
   MMYOLO 统一各类 YOLO 算法模块, 并提供统一评测流程，用户可以公平便捷的进行对比分析。
@@ -63,6 +59,13 @@ MMYOLO 是一个基于 PyTorch 的 YOLO 系列算法开源工具箱。它是 [Op
 
   MMYOLO 提供了从入门到部署到进阶和算法解析等一系列文档，方便不同用户快速上手和扩展。
 
+- **模块化设计**
+
+  MMYOLO 将框架解耦成不同的模块组件，通过组合不同的模块和训练测试策略，用户可以便捷地构建自定义模型。
+
+  ![基类](https://user-images.githubusercontent.com/33799979/190382319-6b4e1fcb-cc3f-4fbe-9d6b-3c9c4e57472c.png)
+  图为 RangeKing@GitHub 提供，非常感谢！
+
 </details>
 
 ## 最新进展