diff --git a/lmdeploy/model.py b/lmdeploy/model.py
index 7abbb6295..311ab2eef 100644
--- a/lmdeploy/model.py
+++ b/lmdeploy/model.py
@@ -156,7 +156,7 @@ class Llama2(BaseModel):
     """Chat template of LLaMA2 model."""
 
     def __init__(self):
-
+        super().__init__()
         B_INST, E_INST = '[INST]', '[/INST]'
         B_SYS, E_SYS = '<<SYS>>\n', '\n<</SYS>>\n\n'
 
diff --git a/lmdeploy/serve/turbomind/chatbot.py b/lmdeploy/serve/turbomind/chatbot.py
index d44dfb4ba..e84157eb0 100644
--- a/lmdeploy/serve/turbomind/chatbot.py
+++ b/lmdeploy/serve/turbomind/chatbot.py
@@ -547,7 +547,6 @@ def stream_consumer(postprocess, res_queue, session, n_input_token,
             except Exception as e:
                 logger.error(f'catch exception: {e}')
 
-        session.response = session.response[len(session.prompt):]
         # put session back to queue so that `_stream_infer` can update it in
         # `self.sessions`
         while not res_queue.empty():