Paper page - Speculative Forcing for Efficient Autoregressive Video Generation

Source: https://huggingface.co/papers/2610.22347

Abstract

Speculative Forcing accelerates few-step autoregressive video generation by accepting early token drafts and skipping unnecessary denoising steps while preserving output quality.

Few-stepautoregressive video generatorsenable streaming generation, yet they run every token through the finaldenoisingstep, regardless of how much refinement it needs. We find this uniform computation is often unnecessary: many tokens already closely match their eventual full-step outputs in the initial steps, while others require refinement until the end. We introduceSpeculative Forcing, a framework that treats intermediateclean estimatesas drafts and predicts when individual tokens can be safely accepted early. Once accepted, a token is frozen as its final output and undergoes no furtherdenoising, while its cached keys and values remain available as context for the remaining tokens. We explore two alternativeverifierdesigns: a training-free one based on similarity to clean content and stability across steps, and a learned one that reads thefrozen generator’sintermediate featuresto estimate each token’s remaining gap to its full-step output. Across short-video, long-video, and world generation,Speculative Forcingpreserves the quality of full-step sampling while substantially reducing computation. On LongLive, it matches the four-step generator with ~1.7denoisingsteps per token, running over 1.5x faster at 41.7 and 37.8 fps. On Lingbot-World 2.0, where camera motion continually changes the visible content, it maintains quality with ~2.5denoisingsteps per token and runs 1.35x faster at 24.5 fps.

View arXiv pageView PDFProject pageGitHub2Add to collection

Get this paper in your agent:

hf papers read 2610\.22347

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.22347 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.22347 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.22347 in a Space README.md to link it from this page.

Collections including this paper1