EventGPT: Event Stream Understanding with Multimodal Large Language Models

¹Shaoyu Liu, ²Jianing Li, ¹Guanghui Zhao, ³Yunjian Zhang,
⁴Xin Meng, ⁵Fei Richard Yu, ²Xiangyang Ji, ⁵Ming Li

¹Xidian University, ²Tsinghua University, ³UCAS, ⁴Peking University
⁵Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ)

EventGPT is an innovative multimodal large language model(MLLM) that integrates event streams and text, founded on spatio-temporal representations and advanced language modeling methodologies.

🏠 About

Video demo of EventGPT.

The EventGPT model, along with the N-ImageNet-Chat and Event-Chat datasets, will be released after the acceptance of our paper.

🔥 News

• [2024-12-03] 🎥 The video demo of EventGPT is live! See it in action!

• [2024-12-02] 🌐 The project page is now online. Explore EventGPT in depth.

• [2024-12-01] 📄 Check out our paper on arXiv and discover the details of EventGPT! 🎉