diff --git a/README.md b/README.md
index 6fba5b9..d9f6123 100644
--- a/README.md
+++ b/README.md
@@ -8,10 +8,12 @@ Training process is pretty fast as there is no backpropagation, and is not very
 
 Here is a demonstration of the model before training and after ~3000 iterations (less than an hour of training):
 
-[Before training](/demo/flappy-lose.gif)
 [After training:](/demo/flappy-success.gif)
+
 ![after training](/demo/flappy-success.gif)
 
+[Before training](/demo/flappy-lose.gif)
+
 For each frame the bird stays alive, +1 score is given to him. For each wall he passes, +10 score is given.
 
 Try it yourself