diff --git a/doc/fluid/advanced_usage/paddle_slim/paddle_slim_en.md b/doc/fluid/advanced_usage/paddle_slim/paddle_slim_en.md
index 2716146fc637858475276cc3b5242e77e047d3d5..1e8a265b084a2a7f606d0e69408e45777aa81567 100644
--- a/doc/fluid/advanced_usage/paddle_slim/paddle_slim_en.md
+++ b/doc/fluid/advanced_usage/paddle_slim/paddle_slim_en.md
@@ -54,7 +54,7 @@ Refer to [Performance Data and ModelZoo](https://github.com/PaddlePaddle/models/
 
 Refer to [Documentation](https://github.com/PaddlePaddle/models/blob/develop/PaddleSlim/docs/usage.md) for more details.
 
-## Introduction to Framework
+## An overview of the architecture
 
 The overall principle of the model compression tool is briefly introduced here to make you understand the use process.
 **figure 1** shows the architecture of the model compression tool, which is a API-relying relation from top to bottom. The distillation, quantization and pruning modules all rely on the bottom paddle framework indirectly. At present, the model compression tool serves as a part of PaddlePaddle framework, so users who have installed paddle of ordinary version have to download and install paddle supporting model compression again for the compression function.