6 papers
INTRA: Interleaved Non-contiguous Token spaRse Attention
Yuxuan Xia, Wenbo Lu, Shaoyi Zheng +1
Transformers achieve strong performance across modalities but are bottlenecked by the quadratic cost of attention. Sparse attention can reduce this cost, but effective sparsity mus…
Smoothing Dark Areas in Molecular Latent Diffusion
Xi Wang, Jiahan Li, Yuxuan Xia +3
Latent diffusion is a promising framework for scalable 3D molecular generation, but it requires a latent space that remains smooth, valid, and navigable beyond posterior samples. E…
Parallel Prefix Verification for Speculative Generation
Yuncheng Yao, Yuxuan Xia, Shengjie Wang +1
We introduce PARSE (PArallel pRefix Speculative Engine), a speculative generation framework that accelerates large language model (LLM) inference by parallelizing prefix verificati…
HilbertA: Hilbert Attention for Image Generation with Diffusion Models
Shaoyi Zheng, Wenbo Lu, Yuxuan Xia +2
Designing sparse attention for diffusion transformers requires reconciling two-dimensional spatial locality with GPU efficiency, a trade-off that current methods struggle to achiev…
ToMA: Token Merge with Attention for Diffusion Models
Wenbo Lu, Shaoyi Zheng, Yuxuan Xia +1
Diffusion models excel in high-fidelity image generation but face scalability limits due to transformers' quadratic attention complexity. Plug-and-play token reduction methods like…
Boosting Few-Shot Segmentation via Instance-Aware Data Augmentation and Local Consensus Guided Cross Attention
Li Guo, Haoming Liu, Yuxuan Xia +2
Few-shot segmentation aims to train a segmentation model that can fast adapt to a novel task for which only a few annotated images are provided. Most recent models have adopted a p…