Video DeltaNet: A Video-Native Hybrid Attention for Livestream Video Generation
This research provides a concrete method for improving the speed of video generation models, which is a key factor in the performance of AI tools for video creation.
Researchers present Video DeltaNet (VDN), a new approach that combines local Softmax attention with bidirectional linear memory to address computational bottlenecks in video diffusion models.
VDN is instantiated on MiniMax H3 and achieves a 14.5x speedup for generating a 14.
3-second video on eight NVIDIA B200 GPUs, compared to a 50-step dense H3 baseline.
Video diffusion models repeatedly process long spatiotemporal token sequences during denoising, making attention a major computational bottleneck. Linear attention offers an appealing alternative and has been widely adopted in recent large language models, but directly applying it to video models often fails to preserve the fine-grained interactions required for high-quality generation. We present Video DeltaNet (VDN), which combines local Softmax attention with bidirectional linear memory for long-range video context. Its linear branch introduces Video Delta Attention (VDA), which updates memory once per frame by jointly incorporating its spatial tokens. Separate output projections and learnable gates calibrate the two branches, while a staged teacher-alignment recipe progressively introduces the new pathway into pretrained models. We instantiate VDN on MiniMax H3, applying the hybrid to video-to-video interactions while retaining Softmax for interactions involving text or audio. With eight-step distillation and an optimized SGLang serving stack, VDN-H3 completes DiT denoising for a 14.3-second, 768p video in 6.70 seconds on eight NVIDIA B200 GPUs, corresponding to a 14.5x speedup over the 50-step dense H3 baseline on the same GPU count.
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Video DeltaNet, MiniMax, video generation, diffusion models.
Source attribution: curated from arXiv cs.LG (@Haocheng Xi, Yiming Xie, Hexu Zhao, Yiwen Zhang, Michael Liu, Thomas Creavin, Kurt Keutzer, Xiuyu Li, Zhaoyang Lv, Chenfeng Xu, Haiwen Feng). Cross-source corroboration: 8 independent sources have covered this development (including NVIDIA AI, Gary Marcus, Tech in Asia), confirming industry-wide signal strength.
