publications

2025

1. A closer look at time steps is worthy of triple speed-up for diffusion model training

   Kai Wang, Mingjia Shi, Yukun Zhou, and 6 more authors

   In Proceedings of the Computer Vision and Pattern Recognition Conference, 2025

   Bib

   @inproceedings{wang2025closer,
     title = {A closer look at time steps is worthy of triple speed-up for diffusion model training},
     author = {Wang, Kai and Shi, Mingjia and Zhou, Yukun and Li, Zekai and Yuan, Zhihang and Shang, Yuzhang and Peng, Xiaojiang and Zhang, Hanwang and You, Yang},
     booktitle = {Proceedings of the Computer Vision and Pattern Recognition Conference},
     pages = {12934--12944},
     year = {2025}
   }

2. Dataset Growth

   Ziheng Qin, Zhaopan Xu, Yukun Zhou, and 8 more authors

   In European Conference on Computer Vision, 2025

   Bib

   @inproceedings{qin2025dataset,
     title = {Dataset Growth},
     author = {Qin, Ziheng and Xu, Zhaopan and Zhou, Yukun and Zheng, Zangwei and Cheng, Zebang and Tang, Hao and Shang, Lei and Sun, Baigui and Peng, Xiaojiang and Timofte, Radu and others},
     booktitle = {European Conference on Computer Vision},
     pages = {58--75},
     year = {2025},
     organization = {Springer}
   }

3. HumanDreamer: Generating Controllable Human-Motion Videos via Decoupled Generation

   Boyuan Wang, Xiaofeng Wang, Chaojun Ni, and 8 more authors

   In Proceedings of the Computer Vision and Pattern Recognition Conference, 2025
4. GigaVideo-1: Advancing Video Generation via Automatic Feedback with 4 GPU-Hours Fine-Tuning

   Xiaoyi Bao, Jindi Lv, Xiaofeng Wang, and 6 more authors

   arXiv preprint arXiv:2506.10639, 2025

2024

1. Neural network diffusion

   Kai Wang, Zhaopan Xu, Yukun Zhou, and 4 more authors

   arXiv preprint arXiv:2402.13144, 2024
2. Conditional lora parameter generation

   Xiaolong Jin, Kai Wang, Dongwen Tang, and 4 more authors

   arXiv preprint arXiv:2408.01415, 2024
3. Open-sora: Democratizing efficient video production for all

   Zangwei Zheng, Xiangyu Peng, Tianji Yang, and 6 more authors

   arXiv preprint arXiv:2412.20404, 2024