Update app.py

app.py

@@ -11,6 +11,8 @@ llm = Llama(model_path= hf_hub_download(repo_id="TheBloke/OpenAssistant-Llama2-1
 
 history = []
 
+<|prompter|>{prompt}<|endoftext|><|assistant|>
+
 pre_prompt = " The user and the AI are having a conversation : "
 
 def generate_text(input_text, history):
@@ -18,12 +20,12 @@ def generate_text(input_text, history):
     print("input ", input_text)
     temp =""
     if history == []:
-        input_text_with_history = f"{pre_prompt}"+ "\n" + f"Q: {input_text} " + "\n" +" A:"
+        input_text_with_history = f"{pre_prompt}"+ "\n" + f"<|prompter|> {input_text} " + "\n" +" <|assistant|>"
     else:
         input_text_with_history = f"{history[-1][1]}"+ "\n"
-        input_text_with_history += f"Q: {input_text}" + "\n" +" A:"
+        input_text_with_history += f"<|prompter|> {input_text}" + "\n" +" <|assistant|>"
     print("new input", input_text_with_history)
-    output = llm(input_text_with_history, max_tokens=1024, stop=["Q:", "\n"], stream=True)
+    output = llm(input_text_with_history, max_tokens=1024, stop=["<|prompter|>", "<|endoftext|>", "\n"], stream=True)
     for out in output:
      stream = copy.deepcopy(out)
      print(stream["choices"][0]["text"])