diff --git a/docs/videos/beales.gif b/docs/videos/beales.gif
deleted file mode 100644
index 7c3a6f1..0000000
Binary files a/docs/videos/beales.gif and /dev/null differ
diff --git a/docs/videos/beales_20171117_00-02-20_2d.gif b/docs/videos/beales_20171117_00-02-20_2d.gif
new file mode 100644
index 0000000..08674bd
Binary files /dev/null and b/docs/videos/beales_20171117_00-02-20_2d.gif differ
diff --git a/docs/videos/beales_CyclicLR_20171117_04-01-14_2d.gif b/docs/videos/beales_CyclicLR_20171117_04-01-14_2d.gif
new file mode 100644
index 0000000..7ac96c2
Binary files /dev/null and b/docs/videos/beales_CyclicLR_20171117_04-01-14_2d.gif differ
diff --git a/readme.md b/readme.md
index 2b52c28..fca0181 100644
--- a/readme.md
+++ b/readme.md
@@ -4,18 +4,32 @@ This project generates animations of pytorch optimizers solving toy problems. Ex
 
 # Examples
 
-![](docs/videos/beales.gif)
+Please note that these don't have the same learning rate as simple optimizers perform better on low dimensional problems. So with a constant learning rate SGD gets to the minima too fast to see while Adam crawls along. So I used differen't learning rate per optimizer to show them in the same video.
 
+## With cyclic annealing:
+
+![](docs/videos/beales_CyclicLR_20171117_04-01-14_2d.gif)
+
+## Constant learning rate
+
+### Beales function
+![](docs/videos/beales_20171117_00-02-20_2d.gif)
+![](docs/videos/beales_20171115_07-18-03_3d.gif)
+
+### Six humped camel function
 ![](docs/videos/six_humped_camel_back_20171115_09-38-57.gif)
 ![](docs/videos/six_humped_camel_back_20171115_09-38-57_3d.gif)
 ![](docs/videos/six_humped_camel_back_20171115_09-38-57_loss.gif)
 
+### Madsen function
+
 ![](docs/videos/madsen_20171115_10-04-15_3d.gif)
 ![](docs/videos/madsen_20171115_10-04-15_loss.gif)
 
+### Rosenbrock function
+
 ![](docs/videos/rosenbrock_20171115_09-47-52.gif)
 
-![](docs/videos/beales_20171115_07-18-03_3d.gif) 
 
 
 # Usage: