Research map / Image generation / Editing and style transfer
CacheEdit: Efficient Multi-round Image Editing via Adaptive Token-wise Reuse.
Jinxin Yu, Xueqing Chen, Yudong Pan, Lian Liu, shengwen Liang, Huawei Li, Xiaowei Li, Ying Wang
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 CacheEdit token-wise activation reuse | FLUX.1 Kontext · Multi-round IIE benchmark | speedup: 1.000 -> 1.915 | Improved |
| add CacheEdit token-wise activation reuse | Qwen-Image-Edit · Multi-round IIE benchmark | speedup: 1.000 -> 1.613 | Improved |
| CacheEdit instead of RegionE | FLUX.1 Kontext · Multi-round IIE benchmark | PSNR vs vanilla: 29.103 -> 39.112 | Improved |
| CacheEdit instead of RegionE | Qwen-Image-Edit · Multi-round IIE benchmark | LPIPS vs vanilla: 0.098 -> 0.063 | Improved |
| add CacheEdit reuse (vs full recomputation) | FLUX.1 Kontext · Multi-round IIE benchmark | GPT-4o overall score (G-O): 6.336 -> 6.814 | Improved |
| add CacheEdit reuse (vs full recomputation) | FLUX.1 Kontext · Multi-round IIE benchmark | GPT-4o semantic consistency (G-SC): 7.566 -> 7.816 | Improved |
| add CacheEdit reuse (vs full recomputation) | Qwen-Image-Edit · Multi-round IIE benchmark | GPT-4o overall score (G-O): 7.384 -> 7.889 | Improved |
| add CacheEdit reuse (vs full recomputation) | Qwen-Image-Edit · Multi-round IIE benchmark | GPT-4o semantic consistency (G-SC): 8.077 -> 8.358 | Improved |
| raise similarity threshold from 0.80 toward 1.00 (ablation) | CacheEdit · FLUX.1 Kontext | reconstruction fidelity (SSIM/LPIPS/PSNR): 0.80 -> 1.00 | Improved |
| lower similarity threshold (more editable tokens) (ablation) | CacheEdit · FLUX.1 Kontext | speedup | Hurt |
| broader edit type (style transfer vs subject edit) (ablation) | CacheEdit · FLUX.1 Kontext | speedup: 2.793 -> 1.156 | Hurt |
| broader edit type (background change vs subject edit) (ablation) | CacheEdit · Qwen-Image-Edit | speedup: 2.376 -> 1.157 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- RegionE — RegionE: Adaptive Region-Aware Generation for Efficient Image Editing (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
- AnyEdit: Mastering Unified High-Quality Image Editing for Any Idea (CVPR 2025 · similarity 0.89)
- Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer (NeurIPS 2025 · similarity 0.89)
- Accelerating Text-to-Image Editing via Cache-Enabled Sparse Diffusion Inference (AAAI 2024 · similarity 0.88)
- UltraEdit: Instruction-based Fine-Grained Image Editing at Scale (NeurIPS 2024 · similarity 0.88)
- InsightEdit: Towards Better Instruction Following for Image Editing (CVPR 2025 · similarity 0.88)
- Towards Efficient Diffusion-Based Image Editing with Instant Attention Masks (AAAI 2024 · similarity 0.88)
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.
