Update README.md

README.md

@@ -8,9 +8,10 @@ tags:
 
 ---
 
-# {MODEL_NAME}
+# RecBERT
+
+RecBERT is a Pretrain-for-Recommentation model. It could serve as an item encoder with the capability of capturing both collaborative filtering signals and text semantic Information. RecBERT is pre-trained on CMCC, an industrial video recommendation dataset collected from China Mobile.
 
-This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
 
 <!--- Describe your model here -->