| @@ -167,11 +167,11 @@ class TokenBufferMemory: | |||
| else: | |||
| prompt_messages.append(AssistantPromptMessage(content=message.answer)) | |||
| if not prompt_messages: | |||
| return [] | |||
| if not prompt_messages: | |||
| return [] | |||
| # prune the chat message if it exceeds the max token limit | |||
| curr_message_tokens = self.model_instance.get_llm_num_tokens(prompt_messages) | |||
| # prune the chat message if it exceeds the max token limit | |||
| curr_message_tokens = self.model_instance.get_llm_num_tokens(prompt_messages) | |||
| if curr_message_tokens > max_token_limit: | |||
| while curr_message_tokens > max_token_limit and len(prompt_messages) > 1: | |||