diff --git a/README.md b/README.md
index af572db..3ce47ab 100644
--- a/README.md
+++ b/README.md
@@ -350,6 +350,10 @@ and then return text embedding in ndarray.
 
 ​	The text embedding extracted by model.
 
+***return_sentence_emb***: *bool*
+
+The flag to output sentence embedding instead of token embeddings, defaults to False.
+
 <br />
 
 ***save_model(format='pytorch', path='default')***