Research map / Image generation / Text-to-image diffusion
SJD-PAC: Accelerating Speculative Jacobi Decoding via Proactive Drafting and Adaptive Continuation
Jialiang Kang, Han Shu, Wenshuo Li, Yingjie Zhai, Xinghao Chen
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 Proactive Drafting (K-ary tree after rejection) (ablation) | SJD (L=32) · MS-COCO 2017 | step compression: 2.31x -> 2.71x | Improved |
| add Adaptive Continuation on top of PD (ablation) | SJD + PD (L=32) · MS-COCO 2017 | step compression: 2.71x -> 3.52x | Improved |
| expand window length from 32 to 64 (ablation) | SJD-PAC · MS-COCO 2017 | step compression: 3.52x -> 4.51x | Improved |
| increase PD branching K from 2 to 4 (D=3) (ablation) | SJD-PAC · MS-COCO 2017 | step compression: 4.23x -> 4.51x | Improved |
| increase K from 4 to 5 (D=3) (ablation) | SJD-PAC · MS-COCO 2017 | step compression: 4.51x -> 4.47x | Hurt |
| increase PD depth D from 3 to 4 (K=4) (ablation) | SJD-PAC · MS-COCO 2017 | step compression: 4.51x -> 4.49x | Hurt |
| use SJD-PAC instead of SJD | Lumina-mGPT · MS-COCO 2017 | latency speedup: 2.05x -> 3.80x | Improved |
| use SJD-PAC instead of SJD | Lumina-mGPT · MS-COCO 2017 | FID: 31.13 -> 30.69 | No clear effect |
| use SJD-PAC instead of lossy GSD | Lumina-mGPT · PartiPrompts | latency speedup: 4.65x -> 3.97x | Hurt |
| use SJD-PAC instead of SJD2 on Emu3 | Emu3 · MS-COCO 2017 | step compression: 5.62x -> 4.31x | Hurt |
Closest papers by meaning
- Accelerating Auto-regressive Text-to-Image Generation with Training-free Speculative Jacobi Decoding (ICLR 2025 · similarity 0.91)
- Speculative Jacobi-Denoising Decoding for Accelerating Autoregressive Text-to-image Generation (NeurIPS 2025 · similarity 0.91)
- Towards Better & Faster Autoregressive Image Generation: From the Perspective of Entropy (NeurIPS 2025 · similarity 0.86)
- LANTERN: Accelerating Visual Autoregressive Models with Relaxed Speculative Decoding (ICLR 2025 · similarity 0.86)
- Hawk: Leveraging Spatial Context for Faster Autoregressive Text-to-Image Generation (NeurIPS 2025 · similarity 0.85)
- PARD: Accelerating LLM Inference with Low‑Cost PARallel Draft Model Adaptation (ICLR 2026 · similarity 0.85)
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.
