alexgenovese/seesr-upscale-and-stable-diffusion-turbo

Image super-resolution model based on SEESR, designed to reconstruct low-resolution images with sharper details and accurate colors.

Public
49 runs
GitHub
Weights
Paper
Playground API Examples README Versions

SEESR Super-Resolution + Stable Diffusion Turbo

About

This is a Cog implementation of SEESR, an advanced AI image super-resolution model.It is designed to upscale low-resolution or compressed images into high-definition versions with enhanced sharpness, fine textures, and accurate color reproduction.

I added Stable Diffusion Turbo for a faster inference and better output in few steps.

SEESR is ideal for:

  • AI photo restoration

  • Image upscaling for prints or digital media

  • Enhancing old or compressed images

  • Improving visual quality for machine learning pipelines

Optimized for Replicate AI deployment, this model delivers fast inference times and production-ready results.

Examples

Check the Examples tab for before-and-after comparisons, including:

  • Portrait restoration with fine facial details

  • Landscape upscaling with enhanced textures

  • Text clarity improvement for scanned documents

How to run

Run on ReplicateUpload or assign your input image under the image parameter to get a high-resolution output.

The model will output an enhanced image with improved sharpness, detail, and color accuracy.

Config options (via cog.yaml) include:

  • Output resolution scale factor

  • Image quality and compression settings

  • Inference performance tweaks

Licensing & Commercial Use

This project is released under the MIT License.You are free to use, modify, and distribute it for both personal and commercial purposes.