diff --git a/README.md b/README.md
index 57ec263f8f5c579457b62cf9cdb73c65557e60bf..d576c961a3978e12a148a2b1d45bc6cabe6188e8 100644
--- a/README.md
+++ b/README.md
@@ -91,15 +91,14 @@ python train.py
 From the left
 `I`, `Itegt`, `Ite`, `Ite_`, `Msgt`, `Ms`, `Ms_`
 
-* Epoch 2
-  ![](./doc/epoch1.png)
-
-* Epoch 5
-  ![](./doc/epoch5.png)
-* Epoch 10
-  ![](./doc/epoch10.png)
-* Epoch 30
-  ![](./doc/epoch30.png)
+* Epoch 2<br>
+  ![](./doc/epoch1.PNG)
+* Epoch 5<br>
+  ![](./doc/epoch5.PNG)
+* Epoch 10<br>
+  ![](./doc/epoch10.PNG)
+* Epoch 30<br>
+  ![](./doc/epoch30.PNG)
 
 These are not good enough for real task. I think the reason is lack of dataset and simplicity. 
 But, it was a good experience for me to implement the paper.