11 papers
Anchoring Instruction Outside Mask: Exact Reference Caching for Efficient In-Context Diffusion Transformers
Yangshuai Liu, Zheming Li, Jiaao Li +4
Omnimodal generation is central to a wide range of content creation and editing applications. In-context conditioning is essential to this paradigm. It allows diffusion transformer…
TEMPO: Makespan-Aware Expert-Parallel Load Balancing Across Memory- and Compute-Bound Regimes
Jie Li, Chenxin Jia, Jinliang Shen +5
In expert-parallel (EP) MoE serving, every layer synchronizes at the slowest GPU. Dispatchers balance token counts (EPLB, LPLB, UltraEP) or activated-expert counts (METRO), assumin…
AlayaWorld: Interactive Long-Horizon World Modeling - Full Technical Report (v1.1)
AlayaWorld Team, Kaipeng Zhang, Chuanhao Li +15
This report presents an improved version of AlayaWorld. While the backbone architecture, chunk-wise autoregressive generation scheme, and training data remain unchanged from the pr…
Sekai2: From World Exploration to Interactive World Modeling
Kang He, Wenshuo Peng, Zihui Gao +3
Video world models must capture how scenes evolve over time and across viewpoints. Training them for long-horizon generation and camera control therefore benefits from long videos…
X-Stage: An Overlooked Pipeline Stage for Communication-Computation Overlap in DiT Inference
Jianwen Xian, Zhiyuan Xu, Yuchen Li +8
Fine-grained, device-initiated communication lets persistent GPU kernels in distributed diffusion transformer (DiT) inference issue remote stores and overlap data movement with Ten…
HeadCast: Casting Attention Heads for Efficient Autoregressive Video Generation
Jinliang Shen, Lianghao Su, Zheming Li +4
Autoregressive (AR) video diffusion models have become a promising paradigm for long and streaming video synthesis, but the continuously growing Key-Value (KV) cache makes attentio…