(NeurIPS 2025) $\text{S}^2$Q-VDiT: Accurate Quantized Video Diffusion Transformer with Salient Data and Sparse Token Distillation

arXiv | BibTeX

---

This project is the official implementation of our "$\text{S}^2$Q-VDiT: Accurate Quantized Video Diffusion Transformer with Salient Data and Sparse Token Distillation".

---

Results

Comments

• Our code will be released soon!

BibTeX

If you find $\text{S}^2$Q-VDiT is useful and helpful to your work, please kindly cite this paper:

@article{feng2025s,
  title={S $\^{} 2$ Q-VDiT: Accurate Quantized Video Diffusion Transformer with Salient Data and Sparse Token Distillation},
  author={Feng, Weilun and Qin, Haotong and Yang, Chuanguang and Li, Xiangqi and Yang, Han and Li, Yuqi and An, Zhulin and Huang, Libo and Magno, Michele and Xu, Yongjun},
  journal={arXiv preprint arXiv:2508.04016},
  year={2025}
}