SJD-PAC: Accelerating Speculative Jacobi Decoding via Proactive Drafting and Adaptive Continuation
Not on the map. Why?
Abstract (arXiv)
Speculative Jacobi Decoding (SJD) offers a draft-model-free approach to accelerate autoregressive text-to-image synthesis. However, the high-entropy nature of visual generation yields low draft-token acceptance rates in complex regions, creating a bottleneck that severely limits overall throughput. To overcome this, we introduce SJD-PAC, an enhanced SJD framework. First, SJD-PAC employs a proactive drafting strategy to improve local acceptance rates in these challenging high-entropy regions. Second, we introduce an adaptive continuation mechanism that sustains sequence validation after an initial rejection, bypassing the need for full resampling. Working in tandem, these optimizations significantly increase the average acceptance length per step, boosting inference speed while strictly preserving the target distribution. Experiments on standard text-to-image benchmarks demonstrate that SJD-PAC achieves a 3.8× speedup with lossless image quality. Code is available at https://github.com/KangJialiang/SJD-PAC.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 26 of its 30 references that are in the map. First posted on arXiv 19 Mar 2026.
Builds on
26 of 30 references.
References · 26
- 2017Attention Is All You Needcited by 59,932
- 2014Microsoft COCO: Common Objects in Contextcited by 17,498
- 2017Neural Discrete Representation Learningcited by 4,650
- 2021Zero-Shot Text-to-Image Generationcited by 4,276
- 2020Taming Transformers for High-Resolution Image Synthesiscited by 3,229
- 2021CLIPScore: A Reference-free Evaluation Metric for Image Captioningcited by 2,388
- 2022Fast Inference from Transformers via Speculative Decodingcited by 1,615
- 2022Scaling Autoregressive Models for Content-Rich Text-to-Image Generationcited by 1,287
- 2024Chameleon: Mixed-Modal Early-Fusion Foundation Modelscited by 904
- 2025Emerging Properties in Unified Multimodal Pretrainingcited by 855
- 2024Emu3: Next-Token Prediction is All You Needcited by 761
- 2024Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generationcited by 447
14 more
- 2024EAGLE-2: Faster Inference of Language Models with Dynamic Draft Treescited by 362
- 2023SpecInfer: Accelerating Generative Large Language Model Serving with Tree-based Speculative Inference and Verificationcited by 359
- 2024Lumina-mGPT: Illuminate Flexible Photorealistic Text-to-Image Generation with Multimodal Generative Pretrainingcited by 148
- 2023SpecTr: Fast Speculative Decoding via Optimal Transportcited by 129
- 2024ANOLE: An Open, Autoregressive, Native Large Multimodal Models for Interleaved Image-Text Generationcited by 112
- 2024Accelerating Auto-regressive Text-to-Image Generation with Training-free Speculative Jacobi Decodingcited by 65
- 2024LANTERN: Accelerating Visual Autoregressive Models with Relaxed Speculative Decodingcited by 46
- 2025Lumina-mGPT 2.0: Stand-Alone AutoRegressive Image Modelingcited by 35
- 2025ViSpec: Accelerating Vision-Language Models with Vision-Aware Speculative Decodingcited by 21
- 2025Grouped Speculative Decoding for Autoregressive Image Generationcited by 17
- 2025Group Critical-token Policy Optimization for Autoregressive Image Generationcited by 16
- 2025LANTERN++: Enhancing Relaxed Speculative Decoding with Static Tree Drafting for Visual Auto-regressive Modelscited by 11
- 2024Temperature-Centric Investigation of Speculative Decoding with Knowledge Distillationcited by 7
- 2025Speculative Jacobi-Denoising Decoding for Accelerating Autoregressive Text-to-image Generationcited by 7
Built on by
Citing papers per year: 2026: 1.
1 paper.