Speculative Forcing is a framework for efficient autoregressive video generation that treats intermediate clean estimates as drafts and predicts when individual tokens can be safely accepted early. Tokens that are accepted are frozen and no longer undergo further denoising, while their cached keys and values remain available for context. Two verifier designs—one training‑free based on similarity and stability, and one learned that uses intermediate features—enable this selective refinement, preserving full‑step quality while reducing computation across short‑video, long‑video, and world generation tasks.
arXiv:2609.39096v1 Announce Type: new Abstract: Autoregressive video diffusion supports streaming generation and interactive control, but its KV cache grows with the generated history. Existing compr...
arXiv:2609.32540v2 Announce Type: replace-cross Abstract: Few-step autoregressive video diffusion generates a long video by splitting the video into temporal chunks and generating chunk-by-chunk, eac...
The paper introduces RECAP, a method that improves speculative decoding for vision‑language models by recycling a target model’s own predictions. RECAP stores predictions using one‑, two‑, and three‑token keys, allowing contextual memories that are request‑specific and token‑level memories that persist across requests. By fusing retrieved predictions based on entropy‑confidence weighting and using best‑first search to build a candidate tree, RECAP achieves significant speedups over autoregressive decoding without requiring training or a draft model.
arXiv:2610.08777v1 Announce Type: new Abstract: Interactive video world models need to generate each video chunk efficiently while responding faithfully to user controls. Many systems use chunk-wise...
In-Distribution Forcing for Long Video Generation at Test Time introduces a test‑time framework that aligns KV caching and conditioning with training configurations to prevent out‑of‑distribution key‑value entries during long‑horizon video generation. The core technique, self‑caching, ensures each chunk is cached without attending to prior KV entries, keeping the rolling window exactly in‑distribution. This approach extends short‑horizon autoregressive video diffusion models to minute‑scale generation, outperforming prior methods in mitigating drifting as shown by drift metrics and a user study.
Interactive video world models need to generate each video chunk efficiently while responding faithfully to user controls. Many systems use chunk-wise autoregressive generation with few-step denoising...
arXiv:2610.00686v1 Announce Type: new Abstract: Recent video-based world models pair the scalability of autoregressive (AR) prediction with the visual quality of diffusion models. The choice of scene...
arXiv:2609.38114v1 Announce Type: new Abstract: Autoregressive video diffusion enables interactive streaming generation, but suffers from error accumulation over long rollouts. Self-rollout training...
arXiv:2606. 13035v1 Announce Type: cross Abstract: Autoregressive video diffusion models provide a natural formulation for streaming and variable-length video generation by conditioning newly generated frames on previously generated content.
Memory Forcing is a new few‑step streaming method for autoregressive video diffusion that keeps a diverse mid‑horizon history without enlarging the KV cache. It partitions the cache into Archive, Working, and sink banks, and uses Bank‑aware RoPE to keep each bank distinguishable even as absolute temporal indices drift. Experiments show that at 1.3B parameters it outperforms other methods on longer clips, maintains subjects and scenes when they leave and re‑enter the window, and scales to a 5B model that produces more realistic, dynamic videos.
arXiv:2609.36590v1 Announce Type: cross Abstract: Self-speculative decoding accelerates large language model (LLM) inference by drafting tokens from the target model itself, but faces a sharp tradeof...
Autoregressive video diffusion models provide a natural formulation for streaming and variable-length video generation by conditioning newly generated frames on previously generated content. However, extending these models to minute-level generation remains challenging: the limited KV-cache budget prevents the model from retaining the full history, while repeatedly conditioning on self-generated frames induces a context distribution shift that accumulates over time, leading to visual artifacts, quality degradation, and temporal drift.