AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

FreeScale: Unleashing the Resolution of Diffusion Models via Tuning-Free Scale Fusion

Haonan Qiu, Shiwei Zhang, Yujie Wei, Ruihang Chu, Hangjie Yuan, Xiang Wang, Yingya Zhang, Ziwei Liu

ICCV 2025 · Image generation › Text-to-image diffusion · approach: Text-driven synthesis and semantic guidance · 21 citations (OpenAlex)

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 changedWhereMeasuredEffect
add Scale Fusion (ablation)FreeScale · SDXL text-to-image 2048x2048FID: 75.717 -> 44.723Improved
add Scale Fusion (ablation)FreeScale · SDXL text-to-image 4096x4096FID: 68.115 -> 49.796Improved
restrict dilated convolution to down and mid blocks instead of also up-blocks (ablation)FreeScale · SDXL text-to-image 4096x4096FID: 67.447 -> 49.796Improved
upsample in RGB space instead of latent space for images (ablation)FreeScale · SDXL text-to-image 2048x2048FID: 72.454 -> 44.723Improved
upsample in latent space instead of RGB space for video (ablation)FreeScale · VideoCrafter2 text-to-videoFVD: 422.245 -> 484.711Mixed
restrict dilated convolution to down and mid blocks for video (ablation)FreeScale · VideoCrafter2 text-to-videoFVD: 523.323 -> 484.711Improved
FreeScale instead of DemoFusion for videoVideoCrafter2 · VideoCrafter2 text-to-videoFVD: 537.613 -> 484.711Improved
FreeScale instead of DemoFusion for 4096x4096 imagesSDXL · SDXL text-to-image 4096x4096FID: 72.378 -> 49.796Improved
FreeScale instead of direct inference at 4096x4096SDXL · SDXL text-to-image 4096x4096FID_c: 42.383 -> 71.369Hurt
FreeScale instead of Real-ESRGAN super-resolutionSDXL · SDXL text-to-image 4096x4096FID: 43.476 -> 49.796Hurt
FreeScale instead of DemoFusion in video inference timeVideoCrafter2 · VideoCrafter2 text-to-videoTime (min): 9.302 -> 3.787Improved

Later papers that compared with this method

1 later accepted paper reports FreeScale as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most 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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy