Temporal Consistency in Video Generation

Follow this topic Ask about this topic

Momentum

30 papers in the last four weeks, up 275% on the four weeks before. 0.3% of all new papers.

Jul 13Week of Sep 28

Latest papers 224

All topics

In-Flight KV Cache with Clean Anchors for Faster Autoregressive Video Diffusion
Sep 26, 2026Yikai Wang, Xiao Han, Mengmeng Xu +9Long-Horizon Video Generation Diffusion Model Caching
Meta · S-Lab, Nanyang Technological University
The Past Frames the Future: Memory for Autoregressive Video Generation
Sep 23, 2026Harold Haodong Chen, Rongjin Guo, Disen Lan +22Long-Horizon Video Generation Memory-Augmented Video Generation
HKUST
MotionSpec: Spectral Trajectory Supervision for Motion-Consistent Video Generation
Sep 23, 2026Ziqi Ni, Rui Li, Shiqi Jiang +1Text-to-Video Generation Temporal Consistency in Video Generation
Cardiff University · University of Science and Technology of China · East China Normal University
QuantWM: Temporally Consistent 2-Bit KV Cache Quantization for Video World Models
Sep 22, 2026Jiaqi Zhao, Xiaobin Hu, Bo Yin +3KV-Cache Compression KV-Cache Quantization
Harbin Institute of Technology (Shenzhen), Shenzhen, China · National University of Singapore, Singapore
WorldCrafter: Consistent Video World Model with Implicit 3D-aware Memory
Sep 21, 2026Wangbo Yu, Kunhao Liu, Wenbo Hu +8Camera-Conditioned Video Generation Memory-Augmented Video Generation
ARC Lab, Tencent IEG · Peking University
GAE: Learning a Geometry-Native Latent Space for 3D-Consistent World Generation
Sep 21, 2026Jiahao Lu, Minghao Yin, Wenbo Hu +53D Scene Generation 3D Scene Representation
The Hong Kong University of Science and Technology · ARC Lab, Tencent IEG · The University of Hong Kong +1
Recency Forcing: Bridging the Long-Horizon Gap in Autoregressive Video Generation
Sep 17, 2026Tri Cao, Hung Nguyen, Phong Nguyen +1Long-Horizon Video Generation KV-Cache Management
Qualcomm AI Research
SignMimic: Robust High-Quality Sign Language Motion Generation via Human-Shape-Oblivious Pose Transfer Guidance
Sep 12, 2026Zhewen He, Junyi Yu, Haomian Huang +2Identity-Preserving Video Generation Pose-Conditioned Video Generation
New York University Abu Dhabi · ChatSign Technology
World in World: Explore the World with World Models
Sep 11, 2026Chenxi Song, Yanming Yang, Chi ZhangCamera-Controlled Video Generation Long-Horizon Video Generation
Westlake AGI Lab · AGI Lab, Westlake University
Temporal State Transport in Video Generation: Diagnosing and Correcting Spectral Imbalance
Sep 8, 2026Luyao Tang, Bingjun Luo, Dong Yi +5Video Generation Temporal Consistency in Video Generation
The University of Hong Kong, Hong Kong SAR, China · Tsinghua University, Beijing, China · Harbin Engineering University, Harbin, China +2
ActionSplice: In-Flight Action Editing for Interactive World Models
Sep 8, 2026Pardis Taghavi, Tingyu Guo, Jonas Lossner +2Action-Conditioned Video Generation Efficient Inference for World Action Models
Texas A&M University
Better Call CineCrew: Consistent Ultra-Long Narrative-to-Film Generation
Sep 7, 2026Jiaben Chen, Sixun Dong, Qinhong Zhou +4Text-to-Video Generation Long Video Generation
University of Massachusetts Amherst, Amherst, MA 01003, USA · Massachusetts Institute of Technology, Cambridge, MA 02139, USA
BooM-VVT: Boosting Mask-Free Video Virtual Try-On with Image-Level Pseudo Data
Sep 3, 2026Wei Zhang, Xin Li, Peishu Shi +4Virtual Try-On Temporal Consistency in Video Generation
Nanjing University of Science and Technology Nanjing, Jiangsu, China · University of Science and Technology of China Hefei, Anhui, China · National University of Singapore Singapore, Singapore +1
Do Video Generators Track the World Across Segments? A Benchmark and Method for World-State Reasoning in Video Continuation
Sep 3, 2026Yingmao Miao, Pengfei Zhang, Chaoran Xu +5Long Video Generation Video Generation
1Xi’an Jiaotong University · 2DreamX Team, Alibaba Group · †Work done during an internship at DreamX Team, Alibaba Group +1
Stabilizing Camera-Controlled Novel View Synthesis at Inference Time
Sep 3, 2026Prajwal Singh, Arjun Badola, Seema Kumari +2Camera-Controlled Video Generation Video Diffusion Models
CVIG Lab and ISLab, D3 Center · Osaka University
DramaChain Bench: An End-to-End Benchmark for Short-Drama Generation
Sep 1, 2026Haoyuan Shi, Mingtao Chen, Shuo Jiang +12Agent Evaluation Temporal Consistency in Video Generation
Hunyuan, Tencent · Beijing Film Academy · Peking University +1
FlowVVTON: Flow-Guided Mask-Free Video Virtual Try-On
Aug 31, 2026Shengyao Chen, Xianbing Sun, Liqing Zhang +1Virtual Try-On Temporal Consistency in Video Generation
Shanghai Jiao Tong University, Shanghai, China
ScaleVid: Geometry-Aware Video Object Scaling with Mesh-Free Inference
Aug 12, 2026Youze Huang, Penghui Ruan, Bojia Zi +3Physical Consistency in Video Generation Controllable Video Generation
University of Electronic Science and Technology of China · The Hong Kong Polytechnic University · The Chinese University of Hong Kong +2
Bridging Event Streams and DiT: Event-Guided Video Frame Interpolation
Aug 11, 2026Guixu Lin, Yuyang Yu, Xiang Ji +6Video Frame Interpolation Temporal Consistency in Video Generation
The University of Tokyo · South China University of Technology · Wuhan University +1
Stream Forcing: Constructing Unified Training Trajectory for Robust Streaming Video Generation
Aug 11, 2026Yueting Zhu, Yuehao Song, Kaicheng Zhang +5Video Diffusion Models Long-Horizon Video Generation
Huazhong University of Science & Technology · Anyverse Dynamics · Horizon Robotics
LogiShot: Logically Coherent Cross-Shot Video Generation
Aug 9, 2026Shuai Guo, Yuhang Yang, Zeyu Zhang +4Memory-Augmented Video Generation Multi-Shot Video Generation
University of Science and Technology of China · Li Auto Inc.
CosmosAlign: Adapting a World Foundation Model for Generative Traffic Video Forecasting
Aug 7, 2026Quang Minh Dinh, Tuan Kiet DoanLong-Horizon Video Generation Vision Foundation Model Adaptation
Simon Fraser University, Burnaby, BC, Canada · Institut Polytechnique de Paris, Palaiseau, France
Addressable Memory for Video World Models
Aug 7, 2026Xindi Wu, Sven Elflein, James Lucas +5KV Caching Long-Horizon Video Generation
1NVIDIA · 2Princeton University · University of Toronto +1
Surg-UniWorld: A Unified Surgical World Model with Multimodal Control Experts
Aug 7, 2026Rulin Zhou, Wanhao Liu, Guoheng Ma +8Controllable Video Generation Temporal Consistency in Video Generation
Department of Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR 999077, China · Shenzhen Loop Area Institute, Shen Zhen 518055, China · School of Electrical and Computer Engineering, Faculty of Engineering, the University of Sydney, Sydney, NSW 2006, Australia
EmoWorld: A Decoupled Affective Field for Controllable Emotional Video Generation
Aug 6, 2026Bingyuan Wang, Baistan Zhyldyzbekov, Kunyu Feng +1Image-to-Video Generation Video Diffusion Models
The Hong Kong University of Science and Technology (Guangzhou) · The Hong Kong University of Science and Technology
Diff-VF: Training-free High-quality Long Video Generation via Diffusion Model
Aug 6, 2026Haoning Yang, Xinyuan Chen, Yaohui Wang +1Diffusion Model Guidance Video Diffusion Models
Shanghai Jiao Tong University, Shanghai, China · Shanghai Artificial Intelligence Laboratory, Shanghai, China
Vorch-Director: Interactive World Story Model via Noise-Aware Error Rectification
Aug 6, 2026Lisai Zhang, Yidi Wu, Qi Liu +7Audio-Video Generation Long-Horizon Video Generation
1Vorch Team · 2Zhejiang University
In-Context Forcing: Uncovering Context Effects in Autoregressive Video Diffusion
Aug 5, 2026Lingxiao Yang, Liu Liu, Moran Li +4Video Diffusion Models Autoregressive Video Generation
School of Information Science and Technology, ShanghaiTech University, Shanghai 201210, China · Tencent Youtu Lab · College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China
JoyAI-Video-Edit: Real-Time Open-Ended Video Editing with Autoregressive Diffusion
Aug 4, 2026Yicheng Xiao, Wenxun Dai, Xinran Qin +22Video Diffusion Models Video Editing
Joy Future Academy, JD
LiveLight: Real-time Streaming Video Relighting with Interactive Control
Aug 3, 2026Yue Ma, Jiangming Wang, Yucheng Wang +8Video Editing Controllable Video Generation
1HKUST · University of Macau · 3THU +2

Sep 26, 2026, cs.CVJ/K move · Enter open · S save

In-Flight KV Cache with Clean Anchors for Faster Autoregressive Video Diffusion

Yikai Wang, Xiao Han, Mengmeng Xu, Juan Camilo Perez+8

Meta · S-Lab, Nanyang Technological University

Long-Horizon Video Generation Diffusion Model Caching KV-Cache Reuse Autoregressive Video Generation Diffusion Model Inference Acceleration Temporal Consistency in Video Generation

Open paper AskPDF arXiv

Few-step autoregressive video diffusion generates a long video by splitting the video into temporal chunks and generating chunk-by-chunk, each through a short sequence of denoising stages. To memorize chunks that are already generated, previous methods reconstruct a clean or less-noisy key--value (KV) cache by additional forwards to build the cache without advancing an output latent. However, every denoising forward itself already computes the in-flight KV of the current chunk. We introduce FlashForward, which directly reuses this cache to avoid the heavy cache-update-only model forwards. After the current chunk completes one denoising stage, its stage-specific cache is already available for the next chunk. Assigning one GPU to each stage therefore lets different chunks occupy different stages concurrently. This early availability has a quality cost: the resulting stage-matched history is noisy, causing appearance and motion drift among chunks. To complement it, FlashForward produces sparse auxiliary clean anchor latents before the corresponding region is generated so the generation trajectories can be stabilized by this two-sided conditioning. The two memories operate at different temporal scales: sparse clean anchor KV supplies coarse, long-range two-sided structural guidance, while dense stage-matched history preserves fine, recent evolution. With up to four GPUs, FlashForward runs

1.16

1.69\times

faster than HiAR and

1.42

2.92\times

faster than Self-Forcing for 16 FPS videos of 20 seconds or longer across 1.3B and 14B backbone scales at 480p and 720p. On VBench, for the 1.3B model at 480p, it achieves higher scores and remains stable at longer durations, demonstrating that FlashForward generates high-quality and temporally consistent videos across durations of 20s, 35s and 65s at a much faster generation speed.

Momentum

Related topics

Latest papers 224