AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

StreamDiffusion: A Pipeline-level Solution for Real-Time Interactive Generation

Akio Kodaira, Chenfeng Xu, Toshiki Hazama, Takanori Yoshimoto, Kohei Ohno, Shogo Mitsuhori, Soichi Sugano, Hanying Cho, Zhijian Liu, Masayoshi Tomizuka, Kurt Keutzer

ICCV 2025 · Image generation › Fast diffusion sampling · approach: Training-free diffusion acceleration · 74 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 Stream Batch denoising (ablation)StreamDiffusion · Img2imginference time: reduction by half compared to a conventional U-Net loopImproved
add Self-Negative R-CFG instead of conventional CFG (ablation)StreamDiffusion · CFG comparison, 5 stepsinference time (ms): 64.64 -> 31.47 (2.05x)Improved
add Onetime-Negative R-CFG instead of conventional CFG (ablation)StreamDiffusion · CFG comparison, 5 stepsinference time (ms): 64.64 -> 36.04 (1.79x)Improved
use Onetime-Negative R-CFG at 1 step vs CFG (ablation)StreamDiffusion · CFG comparison, 1 stepinference time (ms): 16.74 -> 16.55 (1.01x)No clear effect
add TensorRT (ablation)StreamDiffusion · Img2img, 1 stepinference time (ms): 21.34 -> 10.65Improved
add TensorRT (ablation)StreamDiffusion · Img2img, 10 stepsinference time (ms): 96.94 -> 62.00Improved
add stochastic similarity filter (ablation)StreamDiffusion · Static scene, 20 frames, RTX 3060average power (W): 85.96w to 35.91wImproved
add stochastic similarity filter (ablation)StreamDiffusion · Static scene, RTX 4090average power (W): 238.68w to 119.77wImproved
add stochastic similarity filter (ablation)StreamDiffusion · Dynamic scene, 1000 frames, RTX 4090average power (W): 236.13w to 199.38wImproved
add future-frame attention via Stream Batch (ablation)StreamDiffusion · User study, temporal consistencypreference (%): 9.67 -> 90.33Improved
use probabilistic SSF instead of hard threshold (ablation)StreamDiffusion · User study, streaming qualitypreference (%): 24.44 -> 75.56Improved
add SSF (ablation)StreamDiffusion · User study, SSF imperceptibilitypreference (%): 48.61 -> 51.39No clear effect

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