CuriGPT: Interactive Humanoid Manipulation by Multimodal Large Language Models

Installation

System requirements

The package is only tested by the following configuration.

  1. Ubuntu 20.04
  2. Ros Neotic

Installation

Please refer to the installation manual.

Functions

Multimodal Voice Q&A

For this function, we implemented a Voice Q&A robot based on multimodal large language modelsqwen_vl_chat_v1.

This whole pipeline can be activated by calling

cd /path/to/curigpt_ros
python3 scripts/curigpt.py

The Team

CuriGPT is developed and maintained by the CLOVER Lab (Collaborative and Versatile Robots Laboratory), CUHK.

Acknowledge

We would like to acknowledge the following projects:

  1. wpr_chatgpt
  2. ros-vosk
  3. tts-ros1