diff --git a/README.md b/README.md
index 51f3f37..f1bcd8d 100644
--- a/README.md
+++ b/README.md
@@ -160,8 +160,8 @@ training:
 
 For now the neural-lam config only defines few things: 1) the kind of data
 store and the path to its config, 2) the weighting of different features in
-the loss function, and 3) valid numerical range for output of each feature. 
-If you don't define the state feature weighting it will default to 
+the loss function, and 3) valid numerical range for output of each feature.
+If you don't define the state feature weighting it will default to
 weighting all features equally. The numerical range of all features default
 to $]-\infty, \infty[$.