跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2609.38691 · v3

No Corners Cut: State-Grounded Transitions for Mid-Stream Prompt Switches in Video Generation

首次提交 2026-09-30 08:21
版本更新 2026-11-18 14:18
雷达观测 2026-11-19 19:08
arXiv ↗PDF ↗

摘要

Streaming video generators allow users to dynamically modulate video synthesis via mid-stream prompt switching. Existing streaming methods can respond to the updated instruction while still cutting corners, prematurely realizing goals or taking heuristic shortcuts that bypass necessary intermediate state changes needed for a plausible transition. In this study, we present SEGUE, a novel framework that makes this process explicit and trains the generator to execute these transitions faithfully. At each switch, a training-free planner parses the latest frame and prompts, writes a few segue prompts with roles and durations, and then hands control back to the user's prompt. Furthermore, to address the inherent difficulty of training causal models on short-lived temporal schedules without corrupting preparatory supervision, we introduce SpanDMD, which evaluates each active prompt using the full rollout as temporal context while retaining its DMD residual only within the prompt's assigned span. On OpenTrans-360, a benchmark of 1,800 switches that scores how the old state exits and the new one begins, SEGUE ranks first on all eight transition metrics and raises the overall score over the strongest baseline from 0.866 to 0.887. It leads on four of six StreamAV-Bench instruction-response metrics and achieves the highest CLIP score in all five post-switch segments of VBench-Long evaluation, while the planner transfers to frozen autoregressive generators without retraining.

作者备注:Page: https://jayjrao.github.io/SEGUE/

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v3,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v3 · 2026-11-18 14:18 · 最新观测版本

No Corners Cut: State-Grounded Transitions for Mid-Stream Prompt Switches in Video Generation

Streaming video generators allow users to dynamically modulate video synthesis via mid-stream prompt switching. Existing streaming methods can respond to the updated instruction while still cutting corners, prematurely realizing goals or taking heuristic shortcuts that bypass necessary intermediate state changes needed for a plausible transition. In this study, we present SEGUE, a novel framework that makes this process explicit and trains the generator to execute these transitions faithfully. At each switch, a training-free planner parses the latest frame and prompts, writes a few segue prompts with roles and durations, and then hands control back to the user's prompt. Furthermore, to address the inherent difficulty of training causal models on short-lived temporal schedules without corrupting preparatory supervision, we introduce SpanDMD, which evaluates each active prompt using the full rollout as temporal context while retaining its DMD residual only within the prompt's assigned span. On OpenTrans-360, a benchmark of 1,800 switches that scores how the old state exits and the new one begins, SEGUE ranks first on all eight transition metrics and raises the overall score over the strongest baseline from 0.866 to 0.887. It leads on four of six StreamAV-Bench instruction-response metrics and achieves the highest CLIP score in all five post-switch segments of VBench-Long evaluation, while the planner transfers to frozen autoregressive generators without retraining.

查看此版本解读 →查看此版本原文 ↗