Research map / Image generation / Fast diffusion sampling
Few-Step Diffusion Sampling Through Instance-Aware Discretizations
Liangyu Yuan, Ruoyu Wang, Tong Zhao, Dingwen Fu, Mingkun Lei, Beier Zhu, Chi Zhang
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 instance-level conditioning of timesteps (ablation) | INDIS · CIFAR10 NFE=3 | FID: 12.05 -> 9.26 | Improved |
| add instance-level conditioning of timesteps (ablation) | INDIS · CIFAR10 NFE=4 | FID: 14.44 -> 4.14 | Improved |
| add shift and scale factors (ablation) | INDIS · CIFAR10 NFE=3 | FID: 11.12 -> 9.26 | Improved |
| add shift and scale factors (ablation) | INDIS · CIFAR10 NFE=8 | FID: 3.24 -> 2.34 | Mixed |
| add SVD feature rearrangement of noise (ablation) | INDIS · CIFAR10 NFE=3 | FID: 9.60 -> 9.26 | Improved |
| add SVD feature rearrangement of noise (ablation) | INDIS · CIFAR10 NFE=5 | FID: 3.27 -> 3.31 | No clear effect |
| apply INDIS to Uni_PC solver (ablation) | Uni_PC · CIFAR10 NFE=3 | FID: 68.25 -> 10.88 | Improved |
| apply INDIS to DPM-Solver++ (ablation) | DPM-Solver++ · CIFAR10 NFE=3 | FID: 68.43 -> 9.92 | Improved |
| apply INDIS to iPNDM (ablation) | iPNDM · CIFAR10 NFE=3 | FID: 57.39 -> 9.26 | Improved |
| increase teacher solver steps from 10 to 30 (ablation) | INDIS · CIFAR10 NFE=3 | FID: 11.57 -> 9.26 | Improved |
| increase teacher solver steps from 20 to 30 (ablation) | INDIS · CIFAR10 NFE=4 | FID: 6.59 -> 4.14 | Improved |
| use instance-specific timesteps instead of global optimized on toy data | Euler 3-step sampler · 2D synthetic tree data | MSE: 0.0245 -> 0.0158 | Improved |
Closest papers by meaning
- Shortcutting Pre-trained Flow Matching Diffusion Models is Almost Free Lunch (NeurIPS 2025 · similarity 0.91)
- Align Your Steps: Optimizing Sampling Schedules in Diffusion Models (ICML 2024 · similarity 0.90)
- Adaptive Non-Uniform Timestep Sampling for Accelerating Diffusion Model Training (CVPR 2025 · similarity 0.90)
- Test-time Scaling of Diffusions with Flow Maps (NeurIPS 2026 · similarity 0.90)
- PFDiff: Training-Free Acceleration of Diffusion Models Combining Past and Future Scores (ICLR 2025 · similarity 0.90)
- Adaptive Stochastic Coefficients for Accelerating Diffusion Sampling (NeurIPS 2025 · similarity 0.90)
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.
