Update app.py

app.py

@@ -65,26 +65,28 @@ for message in st.session_state.messages:
 def set_query(query):
     st.session_state.messages.append({"role": "user", "content": query})
 # Create a list of candidate questions
-candidate_questions = ["Is boiling water (100°C) an obtuse angle (larger than 90 degrees)?", "Write a quicksort code in Python.", "笼子里有好几只鸡和兔子。笼子里有72个头，200只腿。里面有多少只鸡和兔子"]
+candidate_questions = ["Is boiling water (100 degrees) an obtuse angle (larger than 90 degrees)?", "Write a quicksort code in Python.", "笼子里有好几只鸡和兔子。笼子里有72个头，200只腿。里面有多少只鸡和兔子"]
 # Display the chat interface with a list of clickable question buttons
 for question in candidate_questions:
     st.sidebar.button(label=question, on_click=set_query, args=[question])
 
 def clear_chat_history():
-    st.session_state.messages = [{"role": "assistant", "content": "How may I assist you today?"}]
+    st.session_state.messages = []
+    # st.session_state.messages = [{"role": "assistant", "content": "How may I assist you today?"}]
 st.sidebar.button('Clear Chat History', on_click=clear_chat_history)
 
-
+@torch.no_grad()
 def generate_fusechat_response():
     conversations=[]
     conversations.append({"role": "system", "content": "You are FuseChat-3.0, created by Sun Yat-sen University. You are a helpful assistant."})
     for dict_message in st.session_state.messages:
         if dict_message["role"] == "user":
-            conversations.append({"role": "user", "content": dict_message["content"]})
+            conversations.append({"role":"user", "content":dict_message["content"]})
         else:
-            conversations.append({"role": "assistant", "content": dict_message["content"]})
-    string_dialogue = tokenizer.apply_chat_template(conversations, tokenize=False, add_generation_prompt=True)
-    input_ids = tokenizer(string_dialogue, return_tensors="pt").input_ids.to('cuda')
+            conversations.append({"role":"assistant", "content":dict_message["content"]})
+    string_dialogue = tokenizer.apply_chat_template(conversations,tokenize=False,add_generation_prompt=True)
+    input_ids = tokenizer(string_dialogue,
+                          return_tensors="pt").input_ids.to('cuda')
     streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)
     generate_kwargs = dict(
         {"input_ids": input_ids},