SJD++: Improved Speculative Jacobi Decoding for Training-free Acceleration of Discrete Auto-regressive Text-to-Image Generation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909949530996736 |
|---|---|
| author | Teng, Yao Jiang, Zhihuan Shi, Han Liu, Xian Ning, Xuefei Dai, Guohao Wang, Yu Li, Zhenguo Liu, Xihui |
| author_facet | Teng, Yao Jiang, Zhihuan Shi, Han Liu, Xian Ning, Xuefei Dai, Guohao Wang, Yu Li, Zhenguo Liu, Xihui |
| contents | Large autoregressive models can generate high-quality, high-resolution images but suffer from slow generation speed, because these models require hundreds to thousands of sequential forward passes for next-token prediction during inference. To accelerate autoregressive text-to-image generation, we propose Speculative Jacobi Decoding++ (SJD++), a training-free probabilistic parallel decoding algorithm. Unlike traditional next-token prediction, SJD++ performs multi-token prediction in each forward pass, drastically reducing generation steps. Specifically, it integrates the iterative multi-token prediction mechanism from Jacobi decoding, with the probabilistic drafting-and-verification mechanism from speculative sampling. More importantly, for further acceleration, SJD++ reuses high-confidence draft tokens after each verification phase instead of resampling them all. We conduct extensive experiments on several representative autoregressive text-to-image generation models and demonstrate that SJD++ achieves $2\times$ to $3\times$ inference latency reduction and $2\times$ to $7\times$ step compression, while preserving visual quality with no observable degradation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_07503 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SJD++: Improved Speculative Jacobi Decoding for Training-free Acceleration of Discrete Auto-regressive Text-to-Image Generation Teng, Yao Jiang, Zhihuan Shi, Han Liu, Xian Ning, Xuefei Dai, Guohao Wang, Yu Li, Zhenguo Liu, Xihui Computer Vision and Pattern Recognition Large autoregressive models can generate high-quality, high-resolution images but suffer from slow generation speed, because these models require hundreds to thousands of sequential forward passes for next-token prediction during inference. To accelerate autoregressive text-to-image generation, we propose Speculative Jacobi Decoding++ (SJD++), a training-free probabilistic parallel decoding algorithm. Unlike traditional next-token prediction, SJD++ performs multi-token prediction in each forward pass, drastically reducing generation steps. Specifically, it integrates the iterative multi-token prediction mechanism from Jacobi decoding, with the probabilistic drafting-and-verification mechanism from speculative sampling. More importantly, for further acceleration, SJD++ reuses high-confidence draft tokens after each verification phase instead of resampling them all. We conduct extensive experiments on several representative autoregressive text-to-image generation models and demonstrate that SJD++ achieves $2\times$ to $3\times$ inference latency reduction and $2\times$ to $7\times$ step compression, while preserving visual quality with no observable degradation. |
| title | SJD++: Improved Speculative Jacobi Decoding for Training-free Acceleration of Discrete Auto-regressive Text-to-Image Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.07503 |