Research map / Image generation / Fast diffusion sampling
Sortblock: Similarity-Aware Feature Reuse for Diffusion Model
Hanqi Chen, Xu Zhang, Xiaoliu Guan, Lielin Jiang, Guanzhong Wang, Zeyu Chen, Yi Liu
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 |
|---|---|---|---|
| apply SortBlock caching (K=5) | Flux.1-dev · COCO text-to-image | FID: 70.59 -> 70.47; latency 23.23 -> 11.60 s | No clear effect |
| increase refresh interval K from 5 to 9 (ablation) | SortBlock on Flux.1-dev · COCO text-to-image | PSNR: 31.49 -> 27.27; speedup 2.00x -> 2.39x | Mixed |
| use SortBlock instead of TeaCache | Flux.1-dev · COCO text-to-image | PSNR: 22.11 -> 31.49 | Improved |
| use SortBlock instead of TaylorSeer | HunyuanVideo · VBench text-to-video | VBench: 79.14 -> 78.29; PSNR 20.98 -> 28.29 | Mixed |
| use SortBlock instead of TeaCache | Wan2.1 · VBench text-to-video | PSNR: 28.09 -> 29.79 | Improved |
| apply SortBlock only in early denoising stage (0-30%) (ablation) | SortBlock · COCO text-to-image | qualitative structure | Hurt |
| lower recomputation scale beta from 1 to 0.1 (ablation) | SortBlock · COCO text-to-image | qualitative detail | Hurt |
Closest papers by meaning
- BlockDance: Reuse Structurally Similar Spatio-Temporal Features to Accelerate Diffusion Transformers (CVPR 2025 · similarity 0.92)
- ALTER: All-in-One Layer Pruning and Temporal Expert Routing for Efficient Diffusion Generation (NeurIPS 2025 · similarity 0.90)
- Faster Diffusion: Rethinking the Role of the Encoder for Diffusion Model Inference (NeurIPS 2024 · similarity 0.89)
- LazyDiT: Lazy Learning for the Acceleration of Diffusion Transformers (AAAI 2025 · similarity 0.89)
- Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think (ICLR 2025 · similarity 0.89)
- Attention-Driven Training-Free Efficiency Enhancement of Diffusion Models (CVPR 2024 · similarity 0.89)
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.
