Efficient LLM-based driving planning is performed on edge devices, while challenging inference tasks are offloaded to the cloud, which possesses extensive resources and larger-scale parameter LLMs.
This project utilizes data collected by autonomous vehicles at the Guangzhou International Campus of South China University of Technology.
LLM performs step-by-step logical reasoning and decision making in a complex traffic environment.
We will be releasing more data and demonstrations soon. Thank you for your attention.