BMTools provides an open-source framework for extending large language models with tools, inspired by ChatGPT Plugins. It enables users to easily build custom plugins using Python functions and integrate existing ChatGPT-compatible plugins, aiming to be an academic, open-source alternative.

How It Works

BMTools facilitates tool integration by converting Python functions into OpenAI-compatible plugin interfaces. It supports both single and multiple tool usage, allowing the language model to recursively call sub-APIs to solve complex queries. The framework emphasizes optimizing tool descriptions and function names for better AI model understanding and execution.

Quick Start & Requirements

• Install via pip install -r requirements.txt and python setup.py develop.
• Requires Python.
• To use local tools, clone CPM-Bee and copy its src/cpm_live to bmtools/models/.
• Official Docs: https://github.com/OpenBMB/BMTools/blob/main/docs/README.md
• Demo: https://github.com/OpenBMB/BMTools/blob/main/web_demo.py

Highlighted Details

• Supports integration of existing ChatGPT-Plugins and custom tool development.
• Released ToolBench, a large-scale tool learning benchmark, and associated training data.
• Offers functionality for single and multi-tool question answering.
• Provides guidance on optimizing tool prompts for improved AI model interaction.

Maintenance & Community

• Active development with recent updates including new tool support and benchmark releases.
• Related project XAgent for further developments.
• Citation available for research use.

Licensing & Compatibility

• The repository does not explicitly state a license in the README.

Limitations & Caveats

The README does not specify a license, which may impact commercial use or integration into closed-source projects. The setup for CPM-Bee integration is described as a manual copy operation.