ViRDM is a new post‑training method for few‑step causal video generation that eliminates the need for a large teacher model and an online critic. By applying representation distribution matching (RDM) with a precomputed target distribution, a lightweight VAE decoder, and staged vector–Jacobian products, ViRDM overcomes memory, optimization, and temporal dynamics challenges. The approach reduces GPU memory usage and training time, achieving state‑of‑the‑art VBench performance with only 20 generator updates and 16 A100 GPU‑hours.
By Zichong Meng, Chongjian Ge, Chun-Hao P. Huang, Yang Zhou, Huaizu Jiang
The paper introduces PreSTP, a training‑free method for speeding up few‑step autoregressive video diffusion models by pruning redundant tokens. It identifies two redundancy patterns—cross‑frame similarity from static content and cross‑step convergence during denoising—and uses early clean estimates from the model’s own trajectory to build a token‑wise redundancy mask. PreSTP achieves significant speedups (up to 3.20× for 5‑second videos and 1.96× for 30‑second videos) while maintaining or improving fidelity compared to existing caching baselines.
By Jia Li, Xurui Peng, Xing Wang, Hayden Kwok-Hay So
Few-step autoregressive (AR) video diffusion enables low-latency streaming generation, but existing post-training methods predominantly rely on Distribution Matching Distillation (DMD), requiring both...
arXiv:2609.40037v1 Announce Type: new
Abstract: Few-step autoregressive video generation enables efficient streaming synthesis, but errors introduced in early temporal blocks are reused as context an...
By Fangyu Lin, Xingtong Ge, Lunjie Zhu, Yi Zhang, Zhening Liu, Tianhang Wang, Mengfei Li, Yumeng Zhang, Guanglu Song, Yu Liu, Jun Zhang
LiveVVT introduces a rolling streaming diffusion framework for video virtual try‑on that maintains high visual fidelity while enabling real‑time performance. It preserves bounded bidirectional modeling within a fixed‑size window, emits clean video chunks iteratively, and uses two memory modules—a bounded temporal memory and a persistent global appearance memory—to sustain long‑term consistency. A progressive distillation process further aligns teacher‑based bidirectional learning with causal few‑step inference, resulting in superior generation quality with 26× lower latency and 11× higher throughput compared to comparable models.
By Yushe Cao, Shikun Feng, Ruxiang Duan, Liyong Wang, Dianxi Shi, Chun Yu, Junliang Xing
arXiv:2601. 09881v2 Announce Type: replace-cross Abstract: Large video diffusion and flow models have achieved remarkable success in high-quality video generation, but their use in real-time interactive applications remains limited due to their inefficient multi-step sampling process.
By Weili Nie, Julius Berner, Nanye Ma, Chao Liu, Saining Xie, Arash Vahdat
arXiv:2607. 06631v1 Announce Type: cross Abstract: Video Diffusion Models (VDMs) have demonstrated superior generation quality but suffer from prohibitive computational costs.
By Yu Cheng, Siyue Yao, Zhongang Qi, Shanyan Guan, Wei Li, Fajie Yuan
arXiv:2610.20220v1 Announce Type: new
Abstract: Autoregressive generation underlies real-time, causal video synthesis and is the backbone of video world models. Today's dominant approach distills a b...
By Zhuobai Dong, Junchao Yi, Hu Jian Guo, Linjie Li, Rui Zhao, Alex Jinpeng Wang
arXiv:2610.10524v1 Announce Type: new
Abstract: Highly compressed video autoencoders offer an effective way to accelerate video diffusion models, as the Diffusion Transformer (DiT) operates on far fe...
By Jiyoung Kim, Paul Hyunbin Cho, Jisu Nam, Donghoon Lee, Hyunsung Go, Yeonkyeong Lee, Hansaem Kim, Seungryong Kim
arXiv:2609.23153v1 Announce Type: new
Abstract: Video diffusion transformers are expensive because attention dominates long spatiotemporal token sequences. We identify the \emph{high-sparsity trap}:...
By Yuxi Liu, Haoyu Li, Zekun Zhang, Tengxu Sun, Yixiang Cai, Jiayong Li, Yifei Xia, Tianle Liu, Baole Ai, Ang Wang, Jiamang Wang, Lin Qu, Kai Zhang, Kun Yuan, Bin Cui
The paper introduces TRACK, a training‑free trajectory routing method that accelerates video diffusion by selectively switching between large and small models during denoising steps. A calibration process generates a disagreement score map, guiding the selection of the appropriate model at each step to maintain quality while reducing computational cost. Experiments on Wan 2.1, Cosmos 3, TurboDiffusion, and FastVideo show speedups ranging from 1.95× to 2.73× with comparable quality and diversity.
By Mustafa Munir, Huy Vu, Shreyas Misra, Rohit Jena, Sajad Norouzi, Ali Taghibakhshi, Anis Ahmad, Anjul Patney, Pavlo Molchanov, Nima Tajbakhsh
Avatar‑Forever introduces a decoupled parallel training framework that separates efficient generation from long‑horizon robustness, allowing each capability to be trained independently. The method employs full‑parameter distillation for high‑quality generation and a lightweight long‑horizon adapter trained via Recovery‑oriented Rollout Training (RRT) to mitigate autoregressive errors. Additionally, a ForeverCache feature‑caching mechanism reduces redundant computation during streaming inference, enabling real‑time, high‑resolution avatar generation at 27.2 FPS on a single H100 GPU.
By Ruibin Li, Zhiyuan Ma, Fangzhou Ai, Lei Zhang