Block-wise Adaptive Caching for Accelerating Diffusion Policy

ICLR 2026 Conference Submission332 Authors

Keywords: Efficient AI, Diffusion Policy, Visuomotor Policy, Robotics, Action Generation, Model Caching.

TL;DR: We introduce Block-wise Adaptive Caching, an efficient training-free caching plugin to accelerate Diffusion Policy for triple times.

Abstract:

Diffusion Policy has demonstrated strong visuomotor modeling capabilities, but its high computational cost renders it impractical for real-time robotic control.
Despite huge redundancy across repetitive denoising steps, existing diffusion acceleration techniques fail to generalize to Diffusion Policy due to fundamental architectural and data divergences.
In this paper, we propose **B**lock-wise **A**daptive **C**aching (**BAC**), a method to accelerate Diffusion Policy by caching intermediate action features. BAC achieves lossless action generation acceleration by adaptively updating and reusing cached features at the block level, based on a key observation that feature similarities vary non-uniformly across timesteps and blocks.
To operationalize this insight, we first propose the Adaptive Caching Scheduler, designed to identify optimal update timesteps by maximizing the global feature similarities between cached and skipped features. However, applying this scheduler for each block leads to significant error surges due to the inter-block propagation of caching errors, particularly within Feed-Forward Network (FFN) blocks. To mitigate this issue, we develop the Bubbling Union Algorithm, which truncates these errors by updating the upstream blocks with significant caching errors before downstream FFNs.
As a training-free plugin, BAC is readily integrable with existing transformer-based Diffusion Policy and vision-language-action models. Extensive experiments on multiple robotic benchmarks demonstrate that BAC achieves up to $3 \times$ inference speedup for free.

Primary Area: applications to robotics, autonomy, planning

Submission Number: 332