[doc] Add performance comparison

Author: Banyc

docs/_media/csharp_vs_python_speed_memory.jpg

@@ -12,6 +12,10 @@ SciSharp's philosophy allows a large number of machine learning code written in
 
 > No surprise, .NET version (left window) is faster than python version (right window) binding. 10K iterates for linear regression in TensorFlow SGD. (CPU)
 
+![python vs csharp on speed and memory](../_media/csharp_vs_python_speed_memory.jpg)
+
+> It is 2x faster and 1/4 memory occupation of training time in eager mode than python binding. (TensorFlow.NET 0.20-preview2)
+
 In comparison to other projects, like for instance [TensorFlowSharp](https://www.nuget.org/packages/TensorFlowSharp/) which only provide Tensorflow's low-level C++ API and can only run models that were built using Python, Tensorflow.NET also implements Tensorflow's high level API where all the magic happens. This computation graph building layer is still under active development. Once it is completely implemented you can build new Machine Learning models in C# or F#. 
 
 | TensorFlow                 | tf native1.14 | tf native 1.15 | tf native 2.3 |

docs/essentials/introduction.md

@@ -12,6 +12,10 @@
 
 > 毫无悬念， .NET 版本 (左边那个) 完爆 python 版本 (右边那个). 👆 用 TensorFlow SGD 跑 1 万次循环的线性回归. (CPU)
 
+![python vs csharp on speed and memory](../../_media/csharp_vs_python_speed_memory.jpg)
+
+> TensorFlow.NET 有 2x 速度和 1/4 空间占用相比较 python 版. (TensorFlow.NET 0.20-preview2)
+
 举个例子， [TensorFlowSharp](https://www.nuget.org/packages/TensorFlowSharp/) 需要在 python 层之上跑 C# 代码，不够高效。但是 [TF.NET](https://github.com/SciSharp/TensorFlow.NET) 就直接跟 Tensorflow 的 C++ API 打交道，够快！
 
 | TensorFlow                 | tf native1.14 | tf native 1.15 | tf native 2.3 |