GPT4Tools：通过自我教育教授大型语言模型使用工具

GPT4Tools: Teaching Large Language Model to Use Tools via Self-instruction

May 30, 2023

作者: Rui Yang, Lin Song, Yanwei Li, Sijie Zhao, Yixiao Ge, Xiu Li, Ying Shan

cs.AI

摘要

本文旨在有效地使大型语言模型（LLMs）能够使用多模态工具。先进的专有LLMs，如ChatGPT和GPT-4，通过复杂的提示工程展现了利用工具的巨大潜力。然而，这些模型通常依赖于高昂的计算成本和公开不可访问的数据。为了解决这些挑战，我们提出了基于自我指导的GPT4Tools，以使开源LLMs，如LLaMA和OPT，能够使用工具。它通过提示一个先进的教师以各种多模态上下文来生成一个遵循指令的数据集。通过使用低秩适应（LoRA）优化，我们的方法促进了开源LLMs解决各种视觉问题，包括视觉理解和图像生成。此外，我们提供了一个基准来评估LLMs使用工具的能力，这在零-shot和微调方式下进行。大量实验证明了我们的方法对各种语言模型的有效性，不仅显著提高了调用已见工具的准确性，还实现了对未见工具的零-shot能力。代码和演示可在https://github.com/StevenGrove/GPT4Tools找到。

English

This paper aims to efficiently enable Large Language Models (LLMs) to use multimodal tools. Advanced proprietary LLMs, such as ChatGPT and GPT-4, have shown great potential for tool usage through sophisticated prompt engineering. Nevertheless, these models typically rely on prohibitive computational costs and publicly inaccessible data. To address these challenges, we propose the GPT4Tools based on self-instruct to enable open-source LLMs, such as LLaMA and OPT, to use tools. It generates an instruction-following dataset by prompting an advanced teacher with various multi-modal contexts. By using the Low-Rank Adaptation (LoRA) optimization, our approach facilitates the open-source LLMs to solve a range of visual problems, including visual comprehension and image generation. Moreover, we provide a benchmark to evaluate the ability of LLMs to use tools, which is performed in both zero-shot and fine-tuning ways. Extensive experiments demonstrate the effectiveness of our method on various language models, which not only significantly improves the accuracy of invoking seen tools, but also enables the zero-shot capacity for unseen tools. The code and demo are available at https://github.com/StevenGrove/GPT4Tools.