Rethinking Token-wise Feature Caching: Accelerating Diffusion Transformers with Dual Feature Caching

[2412.18911] Rethinking Token-wise Feature Caching: Accelerating Diffusion Transformers with Dual Feature Caching Skip to main content Search Submit Donate Log in Search arXiv Press Enter to search · Advanced search Computer Science > Machine Learning arXiv:2412.18911 (cs) [Submitted on 25 Dec 2024 ( v1 ), last revised 15 Aug 2026 (this version, v3)] Title: Rethinking Token-wise Feature Caching: Accelerating Diffusion Transformers with Dual Feature Caching Authors: Chang Zou , Shikang Zheng , Evelyn Zhang , Runlin Guo , Haohang Xu , Zhengyi Shi , Conghui He , Xuming Hu , Linfeng Zhang View a PDF of the paper titled Rethinking Token-wise Feature Caching: Accelerating Diffusion Transformers with Dual Feature Caching, by Chang Zou and 8 other authors View PDF HTML (experimental) Abstract: Diffusion Transformers (DiT) have become the dominant methods in image and video generation yet still suffer substantial computational costs. As an effective approach for DiT acceleration, feature caching methods are designed to cache the features of DiT in previous timesteps and reuse them in the next timesteps, allowing us to skip the computation in the next timesteps. Among them, token-wise feature caching has been introduced to perform different caching ratios for different tokens in DiTs, aiming to skip the computation for unimportant tokens while still computing the important ones. In this paper, we propose to carefully check the effectiveness in token-wise feature caching with the following two questions: (1) Is it really necessary to compute the so-called "important" tokens in each step? (2) Are so-called important tokens really important? Surprisingly, this paper gives some counter-intuition answers, demonstrating that consistently computing the selected ``important tokens'' in all steps is not necessary. The selection of the so-called ``important tokens'' is often ineffective, and even sometimes shows inferior performance than random selection. Based on these observations, this paper introduces dual feature caching referred to as DuCa, which performs aggressive caching strategy and conservative caching strategy iteratively and selects the tokens for computing randomly. Extensive experimental results demonstrate the effectiveness of our method in DiT, PixArt, FLUX, and OpenSora, demonstrating significant improvements than the previous token-wise feature caching. Subjects: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2412.18911 [cs.LG] (or arXiv:2412.18911v3 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2412.18911 Focus to learn more arXiv-issued DOI via DataCite Journal reference: IEEE Transactions on Image Processing, vol. 35, pp. 6211-6220, 2026 Related DOI : https://doi.org/10.1109/TIP.2026.3698363 Focus to learn more DOI(s) linking to related resources Submission history From: Chang Zou [ view email ] [v1] Wed, 25 Dec 2024 14:00:14 UTC (3,488 KB) [v2] Tue, 18 Nov 2025 06:21:13 UTC (1,820 KB) [v3] Sat, 15 Aug 2026 13:04:51 UTC (1,820 KB) Full-text links: Access Paper: View a PDF of the paper titled Rethinking Token-wise Feature Caching: Accelerating Diffusion Transformers with Dual Feature Caching, by Chang Zou and 8 other authors View PDF HTML (experimental) TeX Source view license Current browse context: cs.LG < prev | next > new | recent | 2024-12 Change to browse by: cs cs.AI cs.CV References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) IArxiv recommender toggle IArxiv Recommender ( What is IArxiv? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from

Related documents

• ChunkFlow: Communication-Aware Chunked Prefetching for Layerwise Offloading in Distributed Diffusion Transformer Inference#178852
• Decomposing Runtime, Kernel, and Quantization Speedups via a Matched FP16 Intermediate: A Hardware-Conditioned Case Study on Four NVIDIA RTX A5000 GPUs#363218
• Compact Partial Symmetry Breaking for Graph Search Problems#1035952
• Optimizing Multiple Feature Types for Image Inpainting in the Linear and Nonlinear Setting#1105083
• Adaptive Routing for Efficient Diffusion Transformer-Based PNI Prediction#363274