paper-with-me

홈 › Papers

DySink: Dynamic Frame Sinks for Autoregressive Long Video Generation

2026-05-20 · Bo Ye, Xinyu Cui, Jian Zhao, Tong Wei, Min-Ling Zhang arxiv

Autoregressive long video generation often adopts bounded-memory streaming for efficiency, typically combining local windows for short-term continuity with static early-frame sinks as long-range anchors. However, this fixed allocation keeps early frames cached even when the current visual state has substantially diverged from them, while discarding potentially more relevant intermediate history. As a result, the retained long-range context may become less adaptive and bias generation toward outdated cues; in severe cases, RoPE-induced phase re-alignment can homogenize inter-head attention and cause sink collapse, where content regresses toward sink frames. We propose DySink, a retrieval-based framework that maintains a compact memory bank and selects visually relevant historical frames as dynamic frame sinks. DySink couples adaptive retrieval with a sink anomaly gate, which detects excessive inter-head consensus over retrieved context and suppresses collapse-prone context. Experiments on minute-long videos show that DySink consistently improves temporal quality over strong baselines while also achieving higher dynamic degree, enabling coherent and more natural long-horizon visual evolution. The code and model weights are released at https://github.com/yebo0216best/DySink.

📄 PDF Abstract BibTeX arXiv:2605.21028

Code (0)

등록된 구현이 없습니다.

Tasks

Video Generation

Similar Papers 제목 키워드 기반

Attention Sinks in Diffusion Language Models

2025-10-17 · Maximo Eduardo Rulli, Simone Petruzzi, Edoardo Michielon, Fabrizio Silvestri 외 arxiv

Masked Diffusion Language Models (DLMs) have recently emerged as a promising alternative to traditional Autoregressive Models (ARMs). DLMs employ transformer encoders with bidirectional attention, enabling parallel token…

Attention Sinks in Diffusion Transformers: A Causal Analysis

2026-05-10 · Fangzheng Wu, Brian Summa arxiv

Attention sinks -- tokens that receive disproportionate attention mass -- are assumed to be functionally important in autoregressive language models, but their role in diffusion transformers remains unclear. We present a…

Deep Forcing: Training-Free Long Video Generation with Deep Sink and Participative Compression

2025-12-04 · Jung Yi, Wooseok Jang, Paul Hyunbin Cho, Jisu Nam 외 arxiv

Recent advances in autoregressive video diffusion have enabled real-time frame streaming, yet existing solutions still suffer from temporal repetition, drift, and motion deceleration. We find that naively applying Stream…

Video Generation

MemRoPE: Training-Free Infinite Video Generation via Evolving Memory Tokens

2026-03-12 · Youngrae Kim, Qixin Hu, C. -C. Jay Kuo, Peter A. Beerel arxiv

Autoregressive diffusion enables real-time frame streaming, yet existing sliding-window caches discard past context, causing fidelity degradation, identity drift, and motion stagnation over long horizons. Current approac…

Video Generation

Sink-Aware Pruning for Diffusion Language Models

2026-02-19 · Aidar Myrzakhan, Tianyi Li, Bowei Guo, Shengkun Tang 외 arxiv

Diffusion Language Models (DLMs) incur high inference cost due to iterative denoising, motivating efficient pruning. Existing pruning heuristics largely inherited from autoregressive (AR) LLMs, typically preserve attenti…