d$^2$Cache: Accelerating Diffusion-Based LLMs via Dual Adaptive Caching
Abstract
Diffusion-based large language models (dLLMs), despite their promising performance, still suffer from inferior inference efficiency. This is because dLLMs rely on bidirectional attention and cannot directly benefit from the standard key-value (KV) cache as autoregressive models (ARMs) do. To tackle this issue, we introduce \textit{Dual aDaptive Cache} (d$^2$Cache), which is a training-free approximate KV cache framework for accelerating dLLM inference. d$^2$Cache features a two-stage fine-grained selection strategy to identify tokens and adaptively update their KV states at each decoding step, while caching the KV states of the remaining tokens for reuse. Furthermore, d$^2$Cache naturally offers a more reliable decoding alternative, which can enable quasi left-to-right generation and mitigate premature overconfidence in tokens at the end of the sequence. Extensive experimental results on two representative dLLMs (\ie, LLaDA and Dream) demonstrate that d$^2$Cache not only achieves substantial inference speedups, but also yields consistent improvements in generation quality. The anonymous evaluation codes are available at \url{https://anonymous.4open.science/r/d2Cache-5538}.
Cite
Text
Jiang et al. "d$^2$Cache: Accelerating Diffusion-Based LLMs via Dual Adaptive Caching." International Conference on Learning Representations, 2026.Markdown
[Jiang et al. "d$^2$Cache: Accelerating Diffusion-Based LLMs via Dual Adaptive Caching." International Conference on Learning Representations, 2026.](https://mlanthology.org/iclr/2026/jiang2026iclr-2cache/)BibTeX
@inproceedings{jiang2026iclr-2cache,
title = {{d$^2$Cache: Accelerating Diffusion-Based LLMs via Dual Adaptive Caching}},
author = {Jiang, Yuchu and Cai, Yue and Luo, Xiangzhong and Fu, Jiale and Wang, Jiarui and Liu, Chonghan and Yang, Xu},
booktitle = {International Conference on Learning Representations},
year = {2026},
url = {https://mlanthology.org/iclr/2026/jiang2026iclr-2cache/}
}