Research map / Image generation / Fast diffusion sampling
Weak-to-Strong Diffusion
Lichen Bai, Masashi Sugiyama, Zeke Xie
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 W2SD weak-to-strong reflection with weight gap | Juggernaut-XL · Pick-a-Pic | HPS v2: 31.6412 -> 32.0992 | Improved |
| apply W2SD with weight gap | DreamShaper · Pick-a-Pic | HPS v2: 30.1477 -> 30.4924 | Improved |
| apply W2SD to personalized LoRA model | SD1.5 + LoRA · Personalized LoRA | CLIP-I: 64.37 -> 68.04 | Improved |
| apply W2SD with expert-selection gap in MoE | DiT-MoE-S · ImageNet 50K | FID: 15.1032 -> 9.1001 | Improved |
| apply W2SD with guidance gap | SDXL · Pick-a-Pic | HPS v2: 29.8701 -> 31.2020 | Improved |
| apply W2SD with prompt-refinement semantic gap | Refined Prompt SDXL · GenEval | HPS v2: 28.5698 -> 29.4023 | Improved |
| add condition gap on top of weight gap (ablation) | W2SD weight gap · Pick-a-Pic | HPS v2: 32.0992 -> 32.9623 | Improved |
| add weight gap on top of condition gap (ablation) | W2SD condition gap · Pick-a-Pic | HPS v2: 32.8217 -> 32.9623 | Improved |
| apply W2SD with weight+guidance gap to video model | AnimateDiff Juggernaut-XL · VBench | Motion Smoothness: 96.8834% -> 96.8706% | Hurt |
| apply W2SD with weight+guidance gap to video model | AnimateDiff Juggernaut-XL · VBench | Dynamic Degree: 13.0653% -> 8.0000% | Hurt |
| make weak model stronger than strong model (negative gap) (ablation) | W2SD · Pick-a-Pic | HPS v2 | Hurt |
| match inference time (T_w2s = T_std/2) with W2SD (ablation) | W2SD · Pick-a-Pic | HPS v2 | Improved |
Closest papers by meaning
- Improving Compositional Generation with Diffusion Models Using Lift Scores (ICML 2025 · similarity 0.89)
- SADA: Stability-guided Adaptive Diffusion Acceleration (ICML 2025 · similarity 0.88)
- One-Step Diffusion Distillation through Score Implicit Matching (NeurIPS 2024 · similarity 0.88)
- Denoising Task Difficulty-based Curriculum for Training Diffusion Models (ICLR 2025 · similarity 0.88)
- Structure-Guided Adversarial Training of Diffusion Models (CVPR 2024 · similarity 0.88)
- Ambient Diffusion Omni: Training Good Models with Bad Data (NeurIPS 2025 · 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.
