#68 · Primary category： Image Generation
SUPIR
SUPIR aims at developing Practical Algorithms for Photo-Realistic Image Restoration In the Wild. Our new online demo is also released at suppixel.ai.
Project last updated：05/12/25
GitHub Stars
5.7K
Forks
476
Contributors
4
License
Other
Why we included this project
SUPIR is a diffusion-based restoration model that turns heavily degraded photographs into clean, high-detail images. It layers SDXL with a LLaVA captioning stage, which helps it recover detail that looks photo-realistic instead of just smoothed over. That makes it a good fit for old scans, heavily compressed web images, and archival or surveillance footage. The repo ships pretrained weights, adjustable restoration strength, and an online demo for quick quality checks, though it is research code from a CVPR 2024 paper, so expect some manual setup around CLIP checkpoints and VRAM. For teams comparing restoration quality, it is a useful reference point against traditional super-resolution tools.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
stable-diffusion-webui
Stable Diffusion web UI
ComfyUI
The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.
diffusers
🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.
upscayl
🆙 Upscayl - #1 Free and Open Source AI Image Upscaler for Linux, MacOS and Windows.
InvokeAI
InvokeAI is a leading open-source creative engine for Stable Diffusion, offering an industry-leading web UI for generating and editing visual media.