Publication: 'Token Pruning for Caching Better: 9× Acceleration on Stable Diffusion for Free', Evelyn Zhang*, Bang Xiao*, et al., arXiv preprint, 2024. Based on token prune and layer cache technology, we present a new Stable Diffusion acceleration method named dynamics-aware token pruning (DaTo). In the COCO-30k, we observed a 7× acceleration coupled with a notable FID reduction of 2.17.
Research Experience
Research Intern at GenAI Group, Microsoft Research Asia, supervised by Shaohan Huang and Tengchao Lv, May. 2025 - Oct. 2025; Research Intern at SJTU MVIG Lab, supervised by Cewu Lu and Yonglu Li, Dec. 2024 - Now; Research Intern at SJTU EPIC Lab, supervised by Linfeng Zhang, Aug. 2024 - Dec. 2024.
Education
B.S. in Computer Science, Zhiyuan Honors Program, Shanghai Jiao Tong University, Sep. 2023 - now; High School, No.1 Middle School Affiliated to Central China Normal University, Sep. 2020 - Jun. 2023.
Background
Research interests broadly span generative models, world models, reinforcement learning, and representation learning. The long-term goal is to build intelligent systems that can truly understand the physical world and are capable of reasoning and planning.