separate thinking process

conversation_public.py

@@ -219,6 +219,7 @@ class Conversation:
                 if type(msg) is tuple:
                     audio_b64_str = base64.b64encode(open(msg[1], "rb").read()).decode("utf-8")
                     msg = f'<audio src="data:audio/wav;base64,{audio_b64_str}" controls autoplay></audio>'
+                msg = msg.replace('</think>', "\n# Final Response\n\n")
                 ret.append({"role": "assistant", "content": msg})
         return ret