AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

d$^2$Cache: Accelerating Diffusion-Based LLMs via Dual Adaptive Caching

Yuchu Jiang, Yue Cai, Xiangzhong Luo, Jiale Fu, Jiarui Wang, Chonghan Liu, Xu Yang

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Masked diffusion language model decoding · 32 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 d2CacheLLaDA-Inst · Six-benchmark averagethroughput (tokens/s): 3.54 -> 12.24Improved
add d2CacheDream-Inst · GSM8Kthroughput (tokens/s): 2.62 -> 12.25Improved
add d2CacheLLaDA-Inst · Six-benchmark averagescore (accuracy %): 43.6 -> 44.4Improved
add d2CacheDream-Inst · Six-benchmark averagescore (accuracy %): 51.4 -> 53.4Improved
use d2Cache instead of Fast-dLLMDream-Inst · Six-benchmark averagescore (accuracy %): 51.7 -> 53.4Improved
use d2Cache instead of Fast-dLLMLLaDA-Inst · Six-benchmark averagethroughput (tokens/s): 9.38 -> 12.24Improved
use d2Cache instead of dLLM-CacheLLaDA-Inst · Six-benchmark averagethroughput (tokens/s): 8.16 -> 12.24Improved
replace confidence-based decoding with certainty prior-guided decoding (NAR) (ablation)LLaDA-Inst · 4-benchmark averagescore (accuracy %): 36.0 -> 50.3Improved
replace confidence-based decoding with certainty prior-guided decoding (NAR) (ablation)Dream-Inst · 4-benchmark averagescore (accuracy %): 29.0 -> 58.8Improved
add d2Cache to certainty prior NAR decoding (ablation)LLaDA-Inst · 4-benchmark averagescore (accuracy %): 50.3 -> 51.2Improved
add d2Cache to certainty prior NAR decoding (ablation)Dream-Inst · 4-benchmark averagescore (accuracy %): 58.8 -> 60.6Improved
add d2Cache to semi-AR decoding (ablation)Dream-Inst · 4-benchmark averagescore (accuracy %): 57.6 -> 57.1Hurt

Later papers that compared with this method

1 later accepted paper reports d2Cache as a baseline in their results tables; 1 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 2 of 2 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