Update app.py

app.py

@@ -40,44 +40,20 @@ print(top_results)
 
 client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
-# def respond(message, history):
-#     response = ""
-
-#     messages = [{"role": "system", "content": "You are a chatbot that helps users create characters for role playing games."}]
-
-#     if history:
-#         messages.extend(history)
-
-#     messages.append({"role": "user", "content": message})
-
-#     stream = client.chat_completion(
-#         messages,
-#         max_tokens=100,
-#         temperature=1.2,
-#         stream=True
-#     )
-
-#     for message in stream:
-#         token = message.choices[0].delta.content
-
-#         if token is not None:
-#             response += token
-#             yield response
 
 def respond(message, history):
     response = ""
 
-    # Retrieve top chunks based on the current user message
     top_chunks = get_top_chunks(message)
     context = "\n".join(top_chunks)
 
-    # Add knowledge as part of system instructions
     messages = [
         {
             "role": "system",
             "content": (
                 "You are a chatbot that helps users create characters for role-playing games. "
                 "Use the following knowledge to inform your answers:\n\n" + context
+                "Make sure to finish all your sentences before you run out of tokens."
             )
         }
     ]
@@ -89,7 +65,7 @@ def respond(message, history):
 
     stream = client.chat_completion(
         messages,
-        max_tokens=100,
+        max_tokens=500,
         temperature=1.2,
         stream=True
     )