CURE-Demo / README.md
ses7720's picture
Add two CCDD-11 example scenes
325caf2 verified
|
Raw
History Blame Contribute Delete
1.07 kB

A newer version of the Gradio SDK is available: 6.25.0

Upgrade
metadata
title: CURE
emoji: 🎛️
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 5.38.0
app_file: app.py
python_version: '3.10'
models:
  - ses7720/CURE
tags:
  - image-restoration
  - controllable-image-restoration
  - low-light-enhancement
  - dehazing
  - deraining
  - desnowing
preload_from_hub:
  - ses7720/CURE CURE_restorer.tar,OneRestore_embedder.tar

CURE: Controllable Unified Image Restoration

This Gradio Space exposes the five CURE inference modes:

  • full one-step restoration;
  • ratio-controlled restoration;
  • selective degradation removal;
  • identity/no-restoration inference; and
  • ordered two-stage restoration.

The model weights are loaded from ses7720/CURE. The source code and command-line inference tools are available on GitHub.

Each tab includes representative CCDD-11 sample images. The samples cover three different scenes, each with all 11 supported degradation combinations, making it easy to compare controls without uploading an image first.