Yuming Jiang
17 Papers
8 Citations
Yuming Jiang is an academic researcher. The author has contributed to research in topics: Computer science & Codebook. The author has an hindex of 5, co-authored 8 publications.
Chat about Author
Papers
LAVIE: High-Quality Video Generation with Cascaded Latent Diffusion Models
Yaohui Wang,Xinyuan Chen,Xin Ma,Shangchen Zhou,Ziqi Huang,Yian Wang,Ceyuan Yang,Yinan He,Jiashuo Yu,Pe-der Yang,Yuwei Guo,Tianxing Wu,Chenyang Si,Yuming Jiang,Cunjian Chen,Chen Change Loy,B. Z. Dai,Dahua Lin,Yu Qiao,Ziwei Liu +19 more
TL;DR: The incorporation of simple temporal self-attentions, coupled with rotary positional encoding, adequately captures the temporal correlations inherent in video data and validate that the process of joint image-video fine-tuning plays a pivotal role in producing high-quality and creative outcomes.
VBench: Comprehensive Benchmark Suite for Video Generative Models
Ziqi Huang,Yinan He,Jiashuo Yu,Fan Zhang,Chenyang Si,Yuming Jiang,Yuanhan Zhang,Tianxing Wu,Qingyang Jin,Nattapol Chanpaisit,Yaohui Wang,Xinyuan Chen,Limin Wang,Dahua Lin,Yu Qiao,Ziwei Liu +15 more
TL;DR: VBench is a comprehensive benchmark suite that dissects video generation quality into specific, hierarchical, and disentangled dimensions, each with tailored prompts and evaluation methods, and provides a dataset of human preference annotations to validate the benchmarks' alignment with human perception.
105
StyleGAN-Human: A Data-Centric Odyssey of Human Generation
Jian Yu Fu,Shikai Li,Yuming Jiang,Kwan-Yee Lin,Chen Qian,Chen Change Loy,Wayne Wu,Ziwei Liu +7 more
- 25 Apr 2022
TL;DR: This work takes a data-centric perspective and investigates multiple critical aspects in “data engineering”, which it believes would complement the current practice and improve the generation quality with rare face poses compared to the long-tailed counterpart.
95
FreeU: Free Lunch in Diffusion U-Net
Chenyang Si,Ziqi Huang,Yuming Jiang,Ziwei Liu +3 more
TL;DR: The untapped potential of diffusion U-Net is uncovered, which serves as a "free lunch" that substantially improves the generation quality on the fly, and a simple yet effective method-termed "FreeU"- is proposed that enhances generation quality without additional training or finetuning.
Text2Human
TL;DR: Extensive quantitative and qualitative evaluations demonstrate that the proposed Text2Human framework can generate more diverse and realistic human images compared to state-of-the-art methods.
64