AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Caching

Zhiyuan Liu, Yicun Yang, Yaojie Zhang, Junjie Chen, Chang Zou, Qingyan Wei, Shaobo Wang, Yichen Zhu, Linfeng Zhang

ICML 2026 · Image generation › Fast diffusion sampling · approach: Masked diffusion language model decoding · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add dLLM-CacheLLaDA Base · GSM8Kscore: 69.06 -> 70.66 with 5.02x FLOPs speedupImproved
add dLLM-CacheDream Instruct · HumanEvalscore: 57.92 -> 56.09Hurt
add dLLM-CacheLLaDA Instruct · GPQAFLOPs: 22.07T -> 2.73T, 8.08xImproved
add dLLM-CacheLLaDA and Dream Instruct · LongBenchaverage score: 39.14 -> 38.61 (LLaDA); 33.04 -> 32.70 (Dream)Hurt
use dLLM-Cache instead of dKV-Cache or Fast-dLLMDream Base · GPQAscore: 31.31 -> 34.15 vs Fast-dLLM; 32.83 vs dKV-CacheImproved
combine dLLM-Cache with SlowFast SamplingLLaDA Base · GSM8KTPS: 7.37 -> 49.86 TPS, 67.17 accuracyImproved
reduce denoising steps to 32 instead of cachingLLaDA Base · GSM8Kaccuracy: 69.06 -> 22.25Hurt
overwrite full V cache instead of selected tokens only (ablation)V-verify · GSM8Kaccuracy: 78.85 -> 79.83Improved
use cosine similarity instead of L2 distance (ablation)V-verify · GSM8Kaccuracy: 55.95 -> 78.54Improved
use fixed global rho 0.25 instead of lower-layer-heavy allocation (ablation)dLLM-Cache · GSM8Kaccuracy: 78.70 -> 79.83Improved
use fixed global rho 0.25 instead of higher-layer-heavy allocation (ablation)dLLM-Cache · GSM8Kaccuracy: 79.23 -> 79.83Improved
increase response refresh interval Kr from 2 to 6 (ablation)dLLM-Cache · GPQA (Dream 7B)accuracy: 36.16 -> 33.92Hurt

Later papers that compared with this method

2 later accepted papers report + dLLM-Cache as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy