[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85643-en":3,"doc-seo-85643-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},85643,4398048949847,"Eliana","https://ap-avatar.wpscdn.com/avatar/400002536579ef2da7f?_k=1778318612642679267",8,"Research & Report","NaviCache Test-Time Self-Calibration Caching for Video Generation","Video Diffusion Models (VDMs) incur extreme computational costs due to many iterative denoising steps. Existing acceleration strategies either depend on offline calibration—introducing calibration overhead, distribution-shift sensitivity, and fitting ambiguity—or rely on calibration-free zero-order test-time approximations that suffer from delayed updates. NaviCache introduces a plug-and-play test-time self-calibration approach, reformulating feature evolution as an Inertial Navigation System problem with dual-state estimation and uncertainty-aware measurement updates, enabling theoretically error-bounded computation skipping.","NaviCache: Test-Time Self-Calibration Caching for Video Generation  \nZheqi Lv 1 2 Zhibo Zhu 1 Jinke Wang 1 Qi Tian 3 Shengyu Zhang 1 Zhengyu Chen 4 Chengxi Zang 2  \nZhou Zhao 1 Fei Wu 1  \n[https://github.com/HelloZicky/NaviCache](https://github.com/HelloZicky/NaviCache)  \narXiv :2606 .26795v2 [ cs .CV] 11 Jul 2026  \nAbstract  \nVideo Diffusion Models (VDMs) is constrained by immense computational costs. While offline calibration-based acceleration suffers from calibration data dependency, prohibitive calibration duration, and susceptibility to distribution shifts, offline calibration-free methods eliminate these hurdles. However, since they rely on instantaneous zero-order approximations where the mapping between input and output differences varies in real-time, they are susceptible to observational noise and ignore the intrinsic momentum within the diffusion trajectory. In this paper, we propose NaviCache, a plug-and-play test-time selfcalibration method re-conceptualizing feature evolution as an Inertial Navigation System (INS) problem. NaviCache bridges the fundamental domain gap and the non-stationary nature of diffusion by modeling the relative coupling between input and output variations. We introduce a dualstate estimation architecture that adaptively tracks the feature change ratio and its latent drift, initialized via a specialized Initial Alignment phase. By integrating a time-dependent noise schedule with an uncertainty-aware Measurement Update mechanism, NaviCache provides a theoretically grounded mechanism for error-bounded computation skipping. Extensive experiments on the HunyuanVideo, Wan, and Open-Sora series demonstrate that NaviCache exhibits more accurate error judgment for computation skipping and achieves outstanding comprehensive performance.  \n1Zhejiang University, Hangzhou, China 2 Cornell University, New York, USA 3Tencent Hunyuan, Shenzhen, China 4MeiTuan LongCat, Beijing, China. Correspondence to: Shengyu Zhang \u003Csy [zhang@zju.edu.cn](zhang@zju.edu.cn) >.  \nProceedings of the 43 rd International Conference on Machine Learning, Seoul, South Korea. PMLR 306, 2026 . Copyright 2026 by the author(s) .  \n1. Introduction  \nWith the advent of video diffusion models (VDM) such as HunyuanVideo (Kong et al., 2024), Wan (Wan Team et al., 2025), and Open-Sora (Zheng et al., 2024 ; 2025), the field of video generation has witnessed rapid development. Despite this success, the iterative sampling process remains a computational tax requiring dozens of forward passes through massive architectures that hinder real-time deployment. Consequently, recent research on video generation acceleration has primarily bifurcated into: (i) Offline calibration-dependent methods. This category includes training-based methods or those relying on statistical priors obtained from curated datasets. For instance, TeaCache (Liu et al., 2025) uses polynomial fitting of residuals, while MagCache (Ma et al., 2026) utilizes a unified magnitude law. However, these methods necessitate a calibration cost and susceptible to distribution shifts between the calibration set and actual deployment scenarios. Furthermore, they are prone to multi-value fitting issues—where a single x maps to multiple y values—as depicted by the “raw points” in Figure 1. (ii) Offline calibration-free methods. To address these limitations, offline calibration-free approaches like EasyCache (Zhou et al., 2025) have emerged. They eliminate the need for calibration datasets, avoiding performance degradation caused by distribution shifts and saving preprocessing time. However, existing offline calibration-free methods use zero-order approximation at test time, which reduces performance due to the delayed update.  \nIn this work, we focus on offline calibration-free methods and offer a transformative perspective: visualizing the relationship between input variations and output responses reveals a manifold resembling a kinematic navigation track. As visualized in Figure 1 (Raw Po","cbCainWlMyhQkct8","https://ap.wps.com/l/cbCainWlMyhQkct8","pdf",5694805,1,14,"English","en",105,"# Abstract\n# Introduction","[{\"question\":\"What problem does NaviCache address in video diffusion model inference?\",\"answer\":\"NaviCache targets the high computational cost of iterative sampling in video diffusion models by enabling computation skipping while maintaining accurate error judgment.\"},{\"question\":\"How is feature evolution modeled in NaviCache?\",\"answer\":\"NaviCache re-conceptualizes feature evolution as an Inertial Navigation System (INS) problem and models relative coupling between input and output variations to handle domain gaps and non-stationary diffusion dynamics.\"},{\"question\":\"How does NaviCache estimate and control errors during sampling?\",\"answer\":\"It uses a dual-state estimation architecture that adaptively tracks feature change ratio and latent drift, initialized through a dedicated Initial Alignment phase, and integrates a time-dependent noise schedule with an uncertainty-aware Measurement Update to bound errors for computation skipping.\"}]",1784205283,35,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"navicache-test-time-self-calibration-caching-for-video-generation","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/navicache-test-time-self-calibration-caching-for-video-generation/85643/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does NaviCache address in video diffusion model inference?","Question",{"text":75,"@type":76},"NaviCache targets the high computational cost of iterative sampling in video diffusion models by enabling computation skipping while maintaining accurate error judgment.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How is feature evolution modeled in NaviCache?",{"text":80,"@type":76},"NaviCache re-conceptualizes feature evolution as an Inertial Navigation System (INS) problem and models relative coupling between input and output variations to handle domain gaps and non-stationary diffusion dynamics.",{"name":82,"@type":73,"acceptedAnswer":83},"How does NaviCache estimate and control errors during sampling?",{"text":84,"@type":76},"It uses a dual-state estimation architecture that adaptively tracks feature change ratio and latent drift, initialized through a dedicated Initial Alignment phase, and integrates a time-dependent noise schedule with an uncertainty-aware Measurement Update to bound errors for computation skipping.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]