diff --git a/13-16章.docx b/13-16章.docx
new file mode 100644
index 0000000..e8d031e
Binary files /dev/null and b/13-16章.docx differ
diff --git a/ch17_模型压缩、加速及移动端部署/第十七章_模型压缩、加速及移动端部署.md b/ch17_模型压缩、加速及移动端部署/第十七章_模型压缩、加速及移动端部署.md
index 198db52..6b5c334 100644
--- a/ch17_模型压缩、加速及移动端部署/第十七章_模型压缩、加速及移动端部署.md
+++ b/ch17_模型压缩、加速及移动端部署/第十七章_模型压缩、加速及移动端部署.md
@@ -292,7 +292,6 @@ Group convolution是将输入特征的维度c1分成g份，每个group对应的c
 Group convolution计算方式如下：  
 ![image](./img/ch17/2.png)
 Group convolution运算量为：  
-
 $$
 B = H * W * h1 * w1 * c1/g * c2/g * g
 $$
@@ -322,7 +321,6 @@ $$
 PW卷积的计算方式如下：
 ![image](./img/ch17/5.png)
 
-
 $$
 B_PW = H_m * W_m * 1 * 1 * c_1 * c_2
 $$