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
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add d2Cache | LLaDA-Inst · Six-benchmark average | throughput (tokens/s): 3.54 -> 12.24 | Improved |
| add d2Cache | Dream-Inst · GSM8K | throughput (tokens/s): 2.62 -> 12.25 | Improved |
| add d2Cache | LLaDA-Inst · Six-benchmark average | score (accuracy %): 43.6 -> 44.4 | Improved |
| add d2Cache | Dream-Inst · Six-benchmark average | score (accuracy %): 51.4 -> 53.4 | Improved |
| use d2Cache instead of Fast-dLLM | Dream-Inst · Six-benchmark average | score (accuracy %): 51.7 -> 53.4 | Improved |
| use d2Cache instead of Fast-dLLM | LLaDA-Inst · Six-benchmark average | throughput (tokens/s): 9.38 -> 12.24 | Improved |
| use d2Cache instead of dLLM-Cache | LLaDA-Inst · Six-benchmark average | throughput (tokens/s): 8.16 -> 12.24 | Improved |
| replace confidence-based decoding with certainty prior-guided decoding (NAR) (ablation) | LLaDA-Inst · 4-benchmark average | score (accuracy %): 36.0 -> 50.3 | Improved |
| replace confidence-based decoding with certainty prior-guided decoding (NAR) (ablation) | Dream-Inst · 4-benchmark average | score (accuracy %): 29.0 -> 58.8 | Improved |
| add d2Cache to certainty prior NAR decoding (ablation) | LLaDA-Inst · 4-benchmark average | score (accuracy %): 50.3 -> 51.2 | Improved |
| add d2Cache to certainty prior NAR decoding (ablation) | Dream-Inst · 4-benchmark average | score (accuracy %): 58.8 -> 60.6 | Improved |
| add d2Cache to semi-AR decoding (ablation) | Dream-Inst · 4-benchmark average | score (accuracy %): 57.6 -> 57.1 | Hurt |
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.
- EntropyCache: Decoded Token Entropy Guided KV Caching for Diffusion Language Models (NeurIPS 2026 · better on 6 of 6 shared results)
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.
- + dLLM-Cache — dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Caching (ICML 2026) · better on 7 of 7 shared results
- + Fast-dLLM — Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding (ICLR 2026) · better on 7 of 7 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
- dKV-Cache: The Cache for Diffusion Language Models (NeurIPS 2025 · similarity 0.95)
- Polestar: Drift-Aware Cache Calibration and Token Commitment for Efficient Inference of Diffusion LLMs (NeurIPS 2026 · similarity 0.93)
- EntropyCache: Decoded Token Entropy Guided KV Caching for Diffusion Language Models (NeurIPS 2026 · similarity 0.92)
- Accelerating Diffusion LLMs via Adaptive Parallel Decoding (NeurIPS 2025 · similarity 0.91)
- MiniCache: KV Cache Compression in Depth Dimension for Large Language Models (NeurIPS 2024 · similarity 0.91)
- Get More with LESS: Synthesizing Recurrence with KV Cache Compression for Efficient LLM Inference (ICML 2024 · similarity 0.91)
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.
