diff --git a/README.md b/README.md
index 8385e6e..f877fb1 100644
--- a/README.md
+++ b/README.md
@@ -61,7 +61,7 @@ python -m task.text_classification.retrain --arch "xxxx"
 
 The `xxxx` should be replaced with the architectures searched. The best architecture we find is
 ```
-python -m task.text_classification.retrain --arch "[[4, 0, 1, -1, 0], [0, 0, 3, -1, 0], [1, 0, 3, 1, 1], [1, 1, 3, 1, 1], [0, 1, 4, -1, 0], [2, 5, 1, 3, 1], [3, 1, 2, 1, 1], [3, 1, 3, 2, 1], [1, 1, 1, -1, 0], [1, 1, 2, 1, 1], [1, 2, 3, 4, 1], [3, 6, 4, 0, 1], [3, 2, 0, -1, 0], [1, 0, 0, -1, 0], [1, 4, 4, 0, 1], [3, 15, 2, 1, 1], [3, 10, 1, -1, 0], [1, 14, 2, 3, 1], [3, 18, 1, 2, 1], [4, 9, 0, -1, 0], [1, 16, 1, -1, 0], [0, 12, 0, -1, 0], [4, 3, 3, 2, 1], [2, 0, 4, -1, 0]]"
+python -m task.text_classification.retrain --arch "[[5, 0, 2, -1, 0], [1, 0, 4, -1, 0], [2, 0, 4, 2, 1], [2, 1, 4, 2, 1], [1, 1, 5, -1, 0], [3, 5, 2, 4, 1], [4, 1, 3, 2, 1], [4, 1, 4, 3, 1], [2, 1, 2, -1, 0], [2, 1, 3, 2, 1], [2, 2, 4, 5, 1], [4, 6, 5, 1, 1], [4, 2, 1, -1, 0], [2, 0, 1, -1, 0], [2, 4, 5, 1, 1], [4, 15, 3, 2, 1], [4, 10, 2, -1, 0], [2, 14, 3, 4, 1], [4, 18, 2, 3, 1], [5, 9, 1, -1, 0], [2, 16, 2, -1, 0], [1, 12, 1, -1, 0], [5, 3, 4, 3, 1], [3, 0, 5, -1, 0]]"
 ```
 
 You should derive a mean test accuracy around `0.5371`