Update app.py

app.py

@@ -145,9 +145,9 @@ def get_prompt(instruction, new_system_prompt=DEFAULT_SYSTEM_PROMPT ):
 
 
 instruction = """\
-You are a helpful assistant, below is a query from a user and some relevant information.
-Answer the user query from these information. first use businessknowledge data try to find answer if you not get  any relevant information then only use context data.  
-you should return only helpfull answer without telling extra things. if you not find any proper information just give output as i don't know .
+You are a helpful assistant, below is a query from a user and businessknowledge and context information.
+Answer the user query from these information . first use businessknowledge information try to find answer if you not get  any relevant information then only use context information.  
+you should return only helpfull answer with proper ending to it  without telling extra things. if you not find any proper information just give output as i don't know .
 
 businessknowledge:
 {context1}
@@ -179,7 +179,13 @@ class QueueCallback(BaseCallbackHandler):
 
     def on_llm_end(self, *args, **kwargs: Any) -> None:
         return self.q.empty()
-
+llm1 = LlamaCpp(
+        model_path="llama-2-7b-chat.Q2_K.gguf",    #  model path
+        callbacks=[QueueCallback(q)],
+        verbose=True,
+        n_ctx=4000,
+        streaming=True,
+    )
 app = FastAPI()
 
 
@@ -192,14 +198,8 @@ def stream(input_text,prompt,context1,context2) -> Generator:
 
     # Initialize the LLM we'll be using
     
-    llm = LlamaCpp(
-        model_path="llama-2-7b-chat.Q2_K.gguf",    #  model path
-        callbacks=[QueueCallback(q)],
-        verbose=True,
-        n_ctx=4000,
-        streaming=True,
-    )
-    llm_chain = LLMChain(prompt=prompt, llm=llm)
+    llm1.callbacks=[QueueCallback(q)]
+    llm_chain = LLMChain(prompt=prompt, llm=llm1)
 
     # Create a funciton to call - this will run in a thread
     def task():