LUA Adapter Skips Pixel Upscaling for Diffusion, Speeding Up by 3x
rsasaki0109 · x · 2026-08-23
LUA (Latent Upscale Adapter) is a fast latent space upsampling adapter for diffusion models, accepted to ECCV 2026. It performs ×2/×4 super-resolution directly in the latent space before VAE decoding, requiring no additional diffusion steps. The approach reduces decoding and upscaling time by nearly 3× and transfers to SDXL, SD3, and FLUX with brief fine-tuning, achieving image quality close to native high-resolution generation at 2K and 4K.
Related event: LUA Latent Upscale Adapter Open-Sourced, Speeds Up Diffusion Nearly 3x(2 posts)→
More from Multimodal
- Midjourney Film Aesthetics: Leaks and Kodak Ektar Prompts — michaelrabone · 2026-08-23
- VideoCoCo: Using Executable Code to Fix Video Physics — jiqizhixin · 2026-08-23
- Midjourney V8.2 new style with --sref 1246983633 — azed_ai · 2026-08-23
- MiniMax H3's high prompt adherence creates stiff, frozen videos lacking subtle motion — DifficultAd5938 · 2026-08-23
- Seeking method for seamless stitching of MiniMax-H3 video segments — Disastrous-Agency675 · 2026-08-23
- 30-Second MiniMax H3 Workflow on 12GB GPUs: Seamless Image-to-Video Pipeline — vortis23 · 2026-08-23