metadata
license: apache-2.0
license: apache-2.0
Dispider: Enabling Video LLMs with Active Real-Time Interaction via Disentangled Perception, Decision, and Reaction (CVPR 2025)
This repository contains the checkpoint (ckpt) of the Dispider model, a multimodal model designed for [online VideoLLM].
Quick Start
First download the checkpoints at the folder.
Important: Modify the mm_compressor path in config.json to align with your local environment. The checkpoint for mm_compressor is located within a sub-folder of this repository.
For detailed evaluation, please refer to Github repo.
✒️ Citation
If you find our work helpful for your research, please consider giving a star ⭐ and citation 📝.
@article{qian2025dispider,
title={Dispider: Enabling Video LLMs with Active Real-Time Interaction via Disentangled Perception, Decision, and Reaction},
author={Qian, Rui and Ding, Shuangrui and Dong, Xiaoyi and Zhang, Pan and Zang, Yuhang and Cao, Yuhang and Lin, Dahua and Wang, Jiaqi},
journal={arXiv preprint arXiv:2501.03218},
year={2025}
}
@article{qian2025streaming,
title={Streaming long video understanding with large language models},
author={Qian, Rui and Dong, Xiaoyi and Zhang, Pan and Zang, Yuhang and Ding, Shuangrui and Lin, Dahua and Wang, Jiaqi},
journal={Advances in Neural Information Processing Systems},
volume={37},
pages={119336--119360},
year={2025}
}