Research map / Image generation / Autoregressive and latent generation
ResDiT: Evoking the Intrinsic Resolution Scalability in Diffusion Transformers
Yiyang Ma, Feng Zhou, Xuedan Yin, Pu Cao, Yonghao Dang, Jianqin Yin
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 PE scaling (ablation) | ResDiT · 3072x3072 | CLIP score: 30.72 -> 32.85 | Improved |
| add PE scaling (ablation) | ResDiT · 4096x4096 | KID: 0.1334 -> 0.0217 | Improved |
| add patch-wise independent PE (ablation) | ResDiT · 3072x3072 | KID: 0.2218 -> 0.0189 | Improved |
| add patch-wise independent PE (ablation) | ResDiT · 4096x4096 | CLIP score: 24.69 -> 32.71 | Improved |
| add patch-wise spectral fusion (ablation) | ResDiT · 3072x3072 | KID: 0.0227 -> 0.0189 | Improved |
| add patch-wise spectral fusion (ablation) | ResDiT · 4096x4096 | KID: 0.0488 -> 0.0217 | Improved |
| add all three components over direct generation (ablation) | ResDiT · 3072x3072 | KID_p: 0.1958 -> 0.0199 | Improved |
| add PE scaling, PIPE, PSF over direct generation (ablation) | ResDiT · 4096x4096 | CLIP score: 21.28 -> 32.71 | Improved |
| add Minimum-Overlap Partitioning and Gaussian Weighting Splicing (ablation) | ResDiT · qualitative boundary artifacts | visual quality: clear boundary artifacts -> eliminated | Improved |
| use single-stage ResDiT instead of two-stage HiFlow | FLUX.1-dev · 4096x4096 | KID: 0.0203 -> 0.0217 | Hurt |
| use single-stage ResDiT instead of two-stage HiFlow | FLUX.1-dev · 3072x3072 | KID: 0.0190 -> 0.0189 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 3 on most shared results.
- HiFlow — HiFlow: Training-free High-Resolution Image Generation with Flow-Aligned Guidance (NeurIPS 2025) · better on 5 of 12 shared results
- DiffuseHigh — DiffuseHigh: Training-Free Progressive High-Resolution Image Synthesis Through Structure Guidance (AAAI 2025) · better on 1 of 2 shared results
- Demofusion — DemoFusion: Democratising High-Resolution Image Generation With No $$$ (CVPR 2024) · better on 2 of 2 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
- ResMaster: Mastering High-Resolution Image Generation via Structural and Fine-Grained Guidance (AAAI 2025 · similarity 0.91)
- SEGA: Spectral-Energy Guided Attention for Resolution Extrapolation in Diffusion Transformers (NeurIPS 2026 · similarity 0.89)
- Native-Resolution Image Synthesis (NeurIPS 2025 · similarity 0.89)
- DiffuseHigh: Training-Free Progressive High-Resolution Image Synthesis Through Structure Guidance (AAAI 2025 · similarity 0.89)
- When Latents Forget Pixels: Restoring Fidelity in Diffusion Transformer Super-Resolution (NeurIPS 2026 · similarity 0.89)
- ScaleCrafter: Tuning-free Higher-Resolution Visual Generation with Diffusion Models (ICLR 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.
